pL CE % NCHS 5 RA CASTS a» ON Cs yo? The Status of Hospital Discharge Data in Denmark, Scotland, West Germany, and the United States VERSITY OF CALIF. BERKELE U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics Library of Congress Cataloging in Publication Data Kozak, Lola Jean. The status of hospital discharge data in Denmark, Scotland, West Germany, and the United States. (Vital and health statistics: Series 2, Data evaluation and methods research; no. 88) (DHHS publication; no. (PHS) 81-1362) Includes bibliographical references. Supt. of Docs. no.: HE 20 6209: 2/88 1. Hospital utilization—Denmark—Statistical services. 2. Hospital utilization—Scotland— Statistical services. 3. Hospital utilization—Germany, West—Statistical services. 4. Hospital utilization—United States—Statistical services. I. Andersen, Ronald, joint author. II. Ander- son, Odin Waldemar, 1914- joint author. III. Title. IV. Series: United States. National Center for Health Statistics. Vital and health statistics: Series 2, Data evaluation and methods research; no. 88. V. Series: United States. Dept. of Health and Human Services. DHHS publication; no. (PHS) 81-1362. [DNLM: 1. Cross-Cultural comparison—Statistics. 2. Patient discharge—Statistics. W2 A N148vb no. 88] RA409.U45 no. 88 [RA971.6] 312'.07'23s 80-607865 ISBN 0-8406-0211-1 [362.1'1'0684] ACKNOWLEDGMENTS This report was financed in part by Purchase Order PLD-08633-78 from the National Center for Health Statistics and by the Center for Health Administration Studies of the University of Chicago. The report was made possible by a number of persons in several countries who provided the authors with the benefit of their considerable knowledge and experi- ence and supplied published and unpublished materials. These people are listed in appendix I. A particular debt is owed to coordinators in Denmark, Scotland, and West Germany, who bore the brunt of continuing requests for information and critiqued drafts of the report: Karen Dreyer in Denmark, M. A. Heasman and J. M. G. Wilson in Scotland, and Elisabeth Schach and H. U. Senftleben in West Germany. Manfred Pflanz of West Germany, recently deceased, also helped to coordinate the study in its early stages. To the extent that this report faithfully reflects the situations in the countries studied, the authors gratefully acknowledge the informants and coordinators. Of course, for any errors of fact or interpreta- tion we must bear sole responsibility. Special thanks also go to National Center for Health Statistics staff Robert Hartford, Office of International Statistics, and Hugo Koch, Division of Health Care Statistics, who took time from their regular duties to translate materials for the study. x Fis n HA oo ups 3 = aa po . wi "=" r RT Ig = : Je ap =F Ti - ad I = Ea CONTENTS Acknowledgments Introduction Comparison of Characteristics of the Countries. The Countries The Hospital Systems Denmark General Hospital Discharge Reporting Systems Methods of Data Collection Coverage Items Collected Definitions and Procedures Information Published or Available Other Discharge Reporting Systems Data Collection Items Available Aggregate Hospital Reports Household Surveys Scotland General Hospital Discharge Reporting System Methods of Data Collection Coverage Items Collected Definitions and Procedures Information Published or Available. Other Discharge Reporting Systems Aggregate Hospital Reports Household Survey Data Collection Items Available West Germany General Hospital Discharge Reporting Systems Sickness Insurance Funds Schleswig-Holstein Other Discharge Reporting Systems, Psychiatric Reporting System Army Reporting System Aggregate Hospital Reports Household Survey U.S. National Hospital Discharge Survey Methods of Data Collection Coverage Items Collected Definitions and Procedures Information Published or Available Summary vi Health Services Systems Long-Term-Care Facilities. Ambulatory and Home Care Services Costs of Receiving Health Care. Summary Summary and Conclusions. References. Appendixes I. Contributors to Study II. Sources of National Hospital Utilization Statistics in the United States in Addition to the National Hospital Discharge Survey LIST OF TEXT TABLES A. Spatial distribution of populations by country B. Percent distribution of populations by age and sex, acCOrding tO COUNIIY...cccceereressarasesssssreasessssnes C. Selected mortality indexes by country D. Number of hospitals, beds, and beds per 1,000 population by ownership, type of hospital, and country E. Percent distribution of hospital beds by type and size of hospital, according to country......cceeeeeene F. Selected short-term hospital utilization statistics by country G. General hospital discharge reporting systems, by country and reporting Syst€Mu....ceeeessessanessessnceses H. Comparability of coverage of general hospital discharge reporting systems, by country and reporting system J. Comparison of items collected in the United States with those collected in Denmark, Scotland, and West Germany, by reporting system K. Coding of diagnoses in general hospital discharge reporting systems, by country and reporting system SYMBOLS Data not available--------=-smsesmeseeeeeeeieeieeeen ooo Category not applicable----------es-seememeeeeaaaaanen wh Quantity zero - Quantity more than 0 but less than 0.05--—-- 0.0 Figure does not meet standards of reliability or precision--------------eeeeeeseceeeeee- * 55 56 31 33 35 36 DATA EVALUATION AND METHODS RESEARCH Series 2 Number 88 The Status of Hospital Discharge Data in Denmark, Scotland, West Germany, and the United States Study of comparability of cross-national hospital discharge data. Descriptions of discharge reporting systems with emphasis on coverage, types of data collected, procedures and definitions used in data collection and analysis, and statistics routinely available. Discussion of health services system characteristics likely to affect rates of hospital use. DHHS Publication No. (PHS) 81-1362 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics Hyattsville, Md. June 1981 NATIONAL CENTER FOR HEALTH STATISTICS DOROTHY P. RICE, Director ROBERT A. ISRAEL, Deputy Director JACOB J. FELDMAN, Ph.D., Associate Director for Analysis and Epidemiology GAIL F. FISHER, Ph.D., Associate Director for the Cooperative Health Statistics System GARRIE J. LOSEE, Associate Director for Data Processing and Services ALVAN O. ZARATE, Ph.D., Assistant Director for International Statistics E. EARL BRYANT, Associate Director for Interview and Examination Statistics ROBERT C. HUBER, Associate Director for Management MONROE G. SIRKEN, Ph.D., Associate Director for Research and Methodology PETER L. HURLEY, Associate Director for Vital and Health Care Statistics ALICE HAYWOOD, Information Officer OFFICE OF INTERNATIONAL STATISTICS ALVAN O. ZARATE, Ph.D., Director Vital and Health Statistics-Series 2-No. 88 DHHS Publication No. (PHS) 81-1362 Library of Congress Catalog Card Number 80-607865 THE STATUS OF HOSPITAL DISCHARGE DATA IN DENMARK, SCOTLAND, WEST GERMANY, AND THE UNITED STATES Lola Jean Kozak, Office of International Statistics, National Center for Health Statistics, and Ronald Andersen and Odin W. Anderson, Center for Health Administration Studies, Graduate School of Business, University of Chicago INTRODUCTION Interest in the health services systems of foreign countries has increased in recent years. These systems are valuable sources of informa- tion on alternative approaches to health care delivery problems in the United States. Numer- ous studies of other health services systems have been undertaken, and much research that could explore and evaluate the differences in the sys- tems has been suggested. A major requirement for further research is the availability of com- parable and accurate data. Government agencies in most developed countries routinely collect a wide array of statistics on health services that might supply the necessary data base, but many questions about the comparability of these sta- tistics remain unanswered. This report examines the status of one type of routinely collected health data: hospital utili- zation statistics. These data were chosen for analysis for several reasons. Hospitals play a cen- tral role in most health services systems, they deal with the most serious health disorders, and they generally absorb the largest share of total health expenditures. Also, countries usually col- lect a considerable amount of hospital data, and cross-national comparisons of hospital statistics are common. In most developed countries there are three types of data systems that collect hospital utili- zation statistics: discharge reporting systems, aggregate hospital reports, and household sur- veys. Discharge reporting systems collect ab- stracts of information about the characteristics of individual discharges, usually including age, sex, and diagnosis. The discharge reporting sys- tems that cover general hospitals are examined in the most detail in this report, but some infor- mation is given on discharge reporting systems that cover special types of hospitals or patients. Aggregate hospital reports and household sur- veys are also briefly reviewed here. Aggregate hospital reports contain information about a hospital’s total number of discharges and bed days but not about individual patients. House- hold surveys usually focus on levels of health or ambulatory care use, but they often collect some information about hospital use. This report is the second in a series of three on the status of cross-national hospital discharge data. The first report describes discharge report- ing systems in Australia, Canada, England and Wales, Finland, France, and Sweden.! The re- porting systems in the six countries were similar enough to be promising data sources for cross- national research, but important differences were also discovered. For instance, variations in the systems’ coverage and in definitions and pro- cedures used for data analysis would have to be taken into account before data comparisons could be made. This report expands on the first report in two ways. First, to learn more about the range and frequency of the similarities and differences in discharge reporting systems, the reporting sys- tems in three additional countries—Denmark, Scotland, and West Germany—are examined. Second, to facilitate comparisons of the U.S. hospital statistics with those of other countries, characteristics of the U.S. National Hospital Dis- charge Survey are discussed. A detailed descrip- tion of the National Hospital Discharge Survey has been published? and is not repeated here, but the comparability of the U.S. survey with reporting systems in other countries is explored. In the third report of the series hospital utili- zation data from all nine countries and the United States will be compared. Estimates will be made of the effects of differences in dis- charge reporting systems on the data produced. Each country’s hospital utilization statistics will be adjusted to take these effects into account. The discharge reporting systems of Den- mark, Scotland, and West Germany were chosen primarily because these countries display various health services system characteristics that may affect the development and operation of dis- charge reporting systems. For instance, a coun- try with a centralized public health services sys- tem is likely to have less difficulty organizing a nationwide discharge reporting system than is a country with a more diverse health service. Eng- land and Wales, with its centralized National Health Service, was one of the first countries to begin operating a reporting system and has suc- ceeded in collecting comparable countrywide data. Scotland takes part in the National Health Service with England and Wales, but the Scottish Health Service is administered separately, and Scotland has created its own discharge reporting system. Whether this discharge reporting system differs from the one in England and Wales was thought to merit investigation. Denmark has a largely public but decen- tralized health services system. Most hospitals are owned and operated by individual counties rather than by a single national health service. This form of organization could make it more difficult to create and maintain a nationwide dis- charge reporting system. In Sweden decentraliza- tion has led to problems; in Finland it has not. As in Denmark, most Swedish and Finnish hos- pitals are owned by local government units. In Sweden the counties own the hospitals and de- termine the form of the discharge reporting sys- tems. Some counties have established a compre- hensive reporting system for all their hospitals; others collect data only from some of their hos- pitals. As a result, uniform national discharge data are not available from these reporting sys- tems. In contrast, Finland has created a national discharge reporting system to which all hospitals report similar data. This difference has aroused interest about Denmark and its ability to obtain uniform national discharge data. West Germany’s health services system is also decentralized; the States rather than the Federal Government are largely responsible for providing health services. Moreover West Ger- many contains a substantial number of private hospitals, which, in other countries, are often excluded from or fail to participate in govern- ment discharge reporting systems. The result can be seriously biased data, as is the case in France. However West Germany has a health insurance system. Discharge reporting systems have been linked to health insurance systems in Australia and Canada, and in spite of health services de- centralization and the large number of private hospitals in both countries, their reporting sys- tems collect comprehensive discharge data. This report investigates whether West Germany uses its health insurance system in this way or adopts other approaches to data collection. All three countries were also selected for study because they contain research colleagues with whom the authors had previously estab- lished working relationships. The modest fund- ing available for the study precluded travel to the countries for data-gathering purposes and limited the authors to published and unpub- lished materials that could be acquired from various agencies and to correspondence with research colleagues in the three countries. Many of them recommended others knowledgeable in hospital statistics, who provided further assist- ance with the study. The major contributors to the study are listed in appendix I. They supplied compre- hensive replies to the original questions, pa- tiently answered followup queries, and provided valuable comments on drafts of the report. They also helped to obtain copies of recent hospital statistics publications from each country. The statistical publications and personal communica- tions were the main sources of information for the report, but a review of other relevant litera- ture concerning the availability, comparability, and quality of the hospital data was also con- ducted. Information about the U.S. National Hospital Discharge Survey was obtained from the Division of Health Care Statistics of the U.S. National Center for Health Statistics as well as from published descriptions of the survey. The report is divided into several sections. The first section provides demographic and hos- pital system data on the three countries and in- cludes a comparison of these data with similar U.S. data. In subsequent sections the hospital statistical systems in Denmark, Scotland, and West Germany are described. These sections con- tain detailed descriptions of discharge reporting systems that cover general hospitals and brief reviews of discharge reporting systems for spe- cial types of hospitals or patients, annual hospi- tal reports of aggregated hospital utilization data, and national household surveys that in- clude data about hospital use. In later sections the U.S. National Hospital Discharge Survey is compared with the discharge reporting systems of the three countries, and selected aspects of health service systems that can be expected to affect hospital utilization are discussed. The closing section provides a summary and conclusions. COMPARISON OF CHARACTERISTICS OF THE COUNTRIES ® THE COUNTRIES Table A shows that Denmark’s population and area are the smallest of the three countries. Its area is about the size of the New Hampshire and Massachusetts combined; Massachusetts alone contains a larger population. Scotland’s area is similar to that of South Carolina; the country also contains fewer inhabitants than does Massachusetts. Although West Germany’s area is about the size of Oregon, it has almost three times the population of California. It is by far the most densely populated of the countries, but Denmark and Scotland also average consid- erably more persons per square kilometer than does the United States. The percent of popula- tion in urban areas is apparently highest in the United States and lowest in West Germany, but the countries define urban areas differently. Table B shows the age and sex distributions of the countries’ populations. While they are generally similar, some differences exist. Com- pared with the United States, each of the other three countries has a smaller percent of persons 2Similar information about Australia, Canada, Eng- land and Wales, Finland, France, and Sweden can be found in reference 1, pages 4-8. age 15-44 years and a higher percent ages 45-64 years and 65 years and over. West Germany has the highest percent of persons age 65 years and over and the lowest percent under age 5 years. The proportion of females in highest in West Germany and lowest in Denmark. ~ Table C shows selected mortality statistics for each country. The infant mortality statistics vary markedly, with Denmark’s infant mortality rate significantly lower than the other countries’ rates. Denmark also reports the longest life ex- pectancy at birth for males. Danish males can expect to live almost 2 years longer than males in the United States. However U.S. females can expect to live slightly longer than Danish fe- males, and over 2 years longer than females in Scotland and West Germany. The United States reports the lowest death rates, but this would be expected because the percent of the population under age 45 is high- est in the United States. Scotland has the highest death rates, even though it has a larger percent of population under age 45 than does West Germany. Heart disease accounts for a greater proportion of all deaths in the United States than it does in the other three countries, but malignant neoplasms account for a greater pro- portion of all deaths in Denmark than in the other countries. Table A. Spatial distribution of populations by country Country Population distribution and year fi Denmark Scotland West United Germany States Population—1977........... ; cerry = 5,088,000 | 5,196,000 | 61,714,000 | 216,817,000 Area 11 STUDIES KIOIMBLEIS ...coccesrirvsrsrsvssississsssrstsssssinrmsssassmsessessssssssmnsvsmssasres 43,096 78,772 248,577 9,363,123 Persons per square kilometer—1977 or 118 66 247 23 Percent of population in urban areas—1970 ..........ccceeeeerierrrrrrirnnneeseneeeessesssnnns 67 n 162 74 11969. SOURCES: References 3 and 4. Table B. Percent distribution of populations by age and sex, according to country Country Age and sex Denmark | Scotland West United 1976 1976 Germany | States 1976 1977 Percent distribution AU BOBS civ cnniisssseissvatssassinestionsmssvavssnssiovesnmsnsnommroniosesnsess Ein ingagarsansns Pa rhe sone 100.0 100.0 100.0 100.0 = I 49.5 48.1 47.6 48.6 anssaciraiin 50.5 51.9 52.4 51.4 3.6 3.5 2.6 3.6 3.5 3.3 25 34 7.9 8.8 8.1 8.6 7.5 84 7.7 8.2 21.4 20.0 21.8 22.4 20.5 20.0 20.6 22.7 10.8 10.6 9.6 9.7 11.3 11.9 12.3 10.5 5.8 5.1 6.5 44 7.8 8.3 9.3 6.4 SOURCE: Reference 3. Table C. Selected mortality indexes by country Mortality index Country West United Scotland Germany States Denmark IAEA MOFNIY Yo ins csi imi ipsa striae ARR Average life expectancy: 2 Selected cause of death:3 Rate per 1,000 live births 89 | 16.1 | 17.4 | 14.1 Years at birth a 68.3 68.3 69.3 76.8 74.6 74.8 773 Rate per 1,000 population AVL CBUIBEI. ov iiss iussvnesnessisnasninssiotsssanss ssnehass sss sRasssss aR EASA AE REA SARSF FIA AREA A ERTS RARER RE 10.6 125 124 8.8 Malignant neoplasms... word 25 2.6 25 1.8 BFL CISBSE civcresrasrnnsrrsssrsrisssssssattssnssemasansirsnsssss 08 ARIE IAIR TASS 0410449800 AAR ARIA ISAT I 3 53000220 3.8 4.0 35 3.3 11977 data for Denmark, Scotland, and the United States, 1976 data for West Germany. 21977 data for Scotland and the United States, 1975-76 data for Denmark, 1974-76 data for West Germany. 31977 data for the United States, 1976 data for Denmark and Scotland, 1975 for West Germany. SOURCES: References 3, 5, and 6. THE HOSPITAL SYSTEMS Table D presents information about the hos- pital systems of each country. Public ownership refers to hospitals owned by the local, State, or national government, and private ownership refers both to nonprofit and profitmaking hospi- tals. Private ownership of hospital facilities is most marked in the United States. However the majority of these hospitals are community hos- pitals, owned and operated by nonprofit groups. Most U.S. public hospitals are owned by State or local authorities. West Germany has the same proportions of public and private hospitals as does the United States, but West German public hospitals contain more beds than do the private hospitals. As in the United States, most West German public hospitals are owned by State or local authorities, and most private hospitals are not profitmaking. Almost all hospitals in Scot- land and Denmark are public institutions. No statistics were available on the private hospitals in Scotland, but it is known that they account for less than 5 percent of all Scottish hospital beds. In Denmark about 7 percent of all hospital beds are in private hospitals. The public hospi- tals in Denmark are almost all owned and oper- ated by the counties, while in Scotland the pub- lic hospitals are part of the National Health Service. The statistics in table D on short-term and long-term hospitals are not strictly comparable. In the United States short-term hospitals are de- fined as hospitals that have an average length of stay of less than 30 days, and long-term hospitals are those with an average length of stay of 30 days or more. In Denmark hospitals are not cate- gorized by length of stay, but since the length of stay of each individual hospital is reported, it is possible to regroup the hospitals into short-term and long-term categories based on the U.S. defi- nitions. In West Germany hospitals are referred to as either acute care or special care hospitals depending on the type of services provided in them. In general the acute care hospitals have average lengths of stay of less than 30 days, and the special hospitals have average lengths of stay of over 30 days, but this is not always the case. Nevertheless acute care hospitals are reported as short-term hospitals and special care hospitals are reported as long-term hospitals in table D. In Table D. Number of hospitals, beds, and beds per 1,000 population by ownership, type of hospital, and country Country Ownership and type of hospital D en mark Ss cota i d 1 Piso Wiis : 1977 1978 Number of hospitals 1 ER A 135 349 3,416 7,159 Ownership ii PU BNI sre mensions i ncisesson crstansanassirss no Eames Sr eb al eA EEE TREE Er vrs ea Serres eds Bela Hers 115 349 1,258 2,607 PrIVELO.. con sssnssrisissrssmssossinssns srs shs ron EsRIRI EFI RT EASA RICA RIL EAH RP SHA EI RPO RA RR HSA SARS HE 20 --- 2,158 4,552 Type SNOT I-BIIIN vs. irs crs sms shh STE Or HATE EE ORR RB A SES REPS EH ATI TR ATTA 105 181 2,185 6,595 LONGABIM oo. civosmiissvnmuemensnmsearrmurses apr EER 30 168 1,231 564 Number of hospital beds WOE rs cas anions iris cin eis sicui sh Tima AS FES RAR EE raps tna brea CF CHAT Eh ns pS 43,436 53,987 | 722,953 | 1,350,097 Ownership Public.. : va ion ux wee 40,373 53,987 | 380,083 540,253 PINAL. civ nirssirsssnirinessmansirmisrsasnrsssnemsssanssn uss sonss anes ssavers sR RIER ERISA A EHS SATIS RFT RISR ESI ARS 3,063 --- 342,870 809,844 Type Short-term 31,431 26,073 | 487,566 | 1,100,368 Long-term 12,005 27,914 | 235,387 249,729 Beds per 1,000 population TORB sss cisensevsansmrsnss iosinsivieemieniovnns isnnsa nse sis ipiss ars SHIEH SERRA RI RET RMRTS SSE CRIA STIL 8.5 10.4 11.7 6.2 7.9 10.4 6.2 25 0.6 --- 5.5 3.7 SUOFLTOUIN oc covri curs senirssanisiisnonssoeti EIR AR I RHETT AA EERE PAT LAT FPO aA FARR SAE AAAS 6.2 5.0 7.9 5.0 LONGI ci neninsinussvmssssissassriimissniatissinsarsssinssssssessssssssssnsinrmnnsssssss obsess se EIR VEER R AS SSR S ETERS 24 5.4 3.8 14 1Data are for National Health Service hospitals, excluding hospitals for the mentally deficient. Private hospitals account for less than 5 percent of all hospital beds. SOURCES: References 6-9. Scotland hospitals are not grouped by length of stay, and length of stay data are not available for hospitals; they are only available for specialties. The Scottish hospitals with half or more of their beds in specialties that have average lengths of stay of less than 30 days are called short-term hospitals in table D, and the hospitals with more than half their beds in specialties that have aver- age lengths of stay of 30 days or more are called long-term hospitals. Over three-fourths of all U.S. hospital beds are in short-term hospitals. Denmark and West Germany are not too different; at least two- thirds of their hospital beds are in short-term hospitals. In contrast, over one-half of the Scot- tish beds are in long-term hospitals. Although the U.S. total bed-to-population ratio is lower than the other countries’, Scotland reports the same number of short-term hospital beds-per- population as does the United States. West Ger- many has the highest total bed-to-population ratio and the highest number of short-term beds per population. Table E shows the distribution of hospital beds in hospitals of different sizes and types. The definitions of short-term and long-term hos- pitals are the same as for table D. The long-term hospitals, many of which are psychiatric hospi- tals, tend to be larger than the short-term hospi- tals. This is particularly true in the United States, where hospitals containing 1,000 beds or more account for less than 5 percent of the short-term hospital beds but over 40 percent of the long- term hospital beds. More short-term hospital beds in the United States are in hospitals with 200-499 beds than in any other size category. This size category also accounts for a larger per- cent of short-term hospital beds than any other category in Scotland and West Germany. In Den- mark, though, over half of the short-term hospi- tal beds are in hospitals with 500 beds or more. Another contrast is that Scotland has a larger percent of beds in hospitals with less than 100 beds than do the other countries. Table F presents utilization statistics for short-term hospitals, which are defined the same as for table D with the exception of the Scottish hospitals. Since no utilization statistics were available by hospital, the Scottish statistics were obtained by subtracting the statistics on special- ties with average lengths of stay of 30 days or more from the totals for all Scottish Health Service hospitals. This procedure probably ex- cludes some patients and beds that are included in the other countries’ statistics since some short-term hospitals include departments for Table E. Percent distribution of hospital beds by type and size of hospital, according to country Type of hospital and bed size Short-term hospitals 100-199 beds... 200-499 beds... 500-999 beds.............. 3 000 BOOS OF FTI0FB ss urrrsissnsssrmssnbrissi st uesss sass ass sass RAITT ARRAS A Country Denmark | Scotland?! es Sis 1977 1977-78 1977 1978 Percent distribution 100.0 100.0 100.0 100.0 0.2 2.8 0.6 0.6 0.7 5.5 1.9 4.2 3.3 7.2 45 10.4 16.9 12.9 15.9 19.2 24.9 36.7 43.9 41.0 28.0 30.4 18.8 20.1 26.0 4.5 14.4 4.4 tenarEs area Ree 100.0 100.0 100.0 100.0 RARER 0.2 1.1 0.6 0.1 0.0 5.5 29 0.6 3.5 11.9 9.1 3.0 13.9 11.9 20.8 6.6 13.1 22.4 25.4 15.0 ase 32.9 33.6 15.4 32.7 sania 36.4 13.7 27.5 42.0 1Data are for National Health Service hospitals, excluding hospitals for the mentally deficient. SOURCES: References 7-10. Table F. Selected short-term hospital utilization statistics by country Country Short-term hospital utilization statistic Denmark | Seotiana? aha ried 1977 1977-78 1977 1978 Discharges per 1,000 population... 176 125 150 170 Bed days per 1,000 population...... 1,749 1,205 2,382 1,351 Mean length of stay in days........... 10 10 16 8 Bed occupancy rate in percent... 78 67 83 73 DiScHardes Pf DOC PBI YBAE cru eirerissssssrsisssecssrsssnssrrmmssssrssssssssnssesssssinsstasrinssass ss rinssnsnasess 29 25 19 34 1Data are for all National Health Service hospitals with the following specialties excluded: mental illness, mental deficiency, geriatrics, young chronic sick, tuberculosis, rehabilitation, convalescence, and general practice long-stay. SOURCES: References 6, 8, 9, 11. long-term care.b On the whole, though, the sta- tistics refer to similar patients and hospital facilities in each country. The United States stands out on two meas- ures: It has the lowest mean length of stay of the countries and the highest average number of discharges per bed. West Germany presents the greatest contrast to the United States. It has the lowest number of discharges per bed and a mean length of stay twice that in the United States. In addition, the bed-day-per-population ratio and the bed occupancy rate are higher in West Ger- many than in the other countries. Denmark has the highest discharge-per-population ratio, and Scotland has the lowest. Scotland also has the lowest bed-day-per-population ratio and bed occupancy rate of the countries. DENMARK Many separate hospital discharge reporting systems have begun operation in Denmark during the last 10 years. However all collect similar data and use uniform procedures and definitions. The reporting systems supply data to national inpatient registers. One register re- ceives information on somatic inpatients, that is, inpatients with physical illnesses and injuries. A second register obtains data on all psychiatric inpatients. Additional hospital utilization data are available from questionnaires collected annu- ally from all Danish hospitals. Household sur- veys are not a source of hospital data. No na- tional health survey has been undertaken in Denmark for several decades. bIn a subsequent report, the statistics from each country will be adjusted to take such differences into account. GENERAL HOSPITAL DISCHARGE REPORTING SYSTEMS In the 1960’s a small number of Danish hos- pitals started pilot projects to collect hospital discharge data. The next step was taken by the Danish National Health Service, an advisory body composed primarily of physicians, nurses, and pharmacists that has a major influence on health policymaking in Denmark. In 1968 the National Health Service set up a working group to explore the possibility of creating a discharge reporting system for use in all the country’s hos- pitals.!2 The working group developed an inpa- tient registration system called M 70, which five hospitals adopted in 1970. In 1971 two other inpatient registration systems were established: the Arhus system in Arhus Municipal Hospital, and the Funen system in Odense Hospital. The three systems grew rapidly. By 1973 registration systems covered all the hospitals in 10 of the country’s 16 hospital areas and some of the hospitals in 2 other areas. Two-thirds of all gen- eral hospital admissions were reported at that time.!13 By 1979 all but five of the country’s hospitals were involved in a registration system, and over 98 percent of all admissions were cov- ered by one.9:14 The National Health Service and the Asso- ciation of County Councils established a coordi- nating group in 1972 to ensure that comparable data were collected by all the registration sys- tems. The group included representatives from the three general hospital registration systems, the psychiatric inpatient register, the Associa- tion of County Councils, and the Rigshospitalet, which is a university hospital run by the State. The group agreed on a uniform set of items to be collected by each registration system and de- veloped official definitions and classifications for use in all systems. The registration systems continue to differ in some ways, but the differ- ences are now mostly technical in nature.l2 Since 1976 the National Health Service has re- quired hospitals to send it magnetic tapes con- taining certain basic information about every inpatient once a year. The tapes are used to form the National Patient Register of somatic hospital discharges. The national and local registers were begun to facilitate clinical and epidemiological research by increasing access to the information in hospi- tal records, and to collect useful data for medi- cal and administrative decisionmaking within hospitals and for hospital planning on local, re- gional, and national levels. It was also hoped that cost data could be linked to registration data so that more meaningful analyses of hospi- tal costs could be made. The National Patient Register is just beginning full operations, but data from it are already being used for hospital planning and medical research.14 Methods of Data Collection Two systems of data collection are used in Danish hospitals: online and batch.14 Hospitals with online systems add admission data to com- CThe hospital areas are the 14 Danish counties and the 2 municipalities of Copenhagen and Frederiksberg. puterized patient records at the time of admis- sion; they add information concerning surgical procedures when the procedures are performed; and they add discharge data, including diag- noses, when the patient is discharged. Hospitals with batch systems use a specially designed re- porting form as the front page of the patient’s medical record. They transfer information from the form to magnetic tape at the end of each month for all patients discharged during that month. In both systems information is ab- stracted from the medical record and items are coded by medical secretaries in the hospitals, who approximately correspond to medical record administrators in the United States. Am- biguous cases are referred to hospital doctors for clarification. Additional data processing is done several ways in the 14 Danish counties. Rigshospitalet and the counties of Vejle and Funen each have a computer. Frederiksborg and Roskilde counties process data together, as do S¢nderjylland and Ribe counties. The counties of Northern Jut- land, Arhus, and Viborg have joined a further development of the Arhus system, now run by a public computer center, Kommunedata. Some counties that use the batch system also have data processed by Kommunedata. Datacentralen, a computer center established in 1959 by the Danish State, counties, and cities, processes batch system data from the municipality of Copenhagen and from some other hospitals.14 Each hospital prepares computer tapes for the National Patient Register and uses the same set format. The tapes, which contain informa- tion on each hospital discharge in a calendar year, are sent to the National Health Service around April 1 of the next year. Coverage At the beginning of 1979 the inpatient registers for somatic patients covered the dis- charges in all publicly owned somatic hospitals and in all but 5 of the 16 private somatic hospi- tals. Together the 5 hospitals contained only 702 beds, or 2 percent of all somatic beds in the country. One with 320 beds, St. Joseph’s Hospi- tal in Copenhagen, closed September 1, 1979. Another with 105 beds, Fysiurgisk Hospital in Hornbaek, plans to transfer its functions to Rigshospitalet and take part in its registration system. The three remaining hospitals treat long- staying patients; two of them specialize in the treatment of rheumatism. Psychiatric hospitals send patient data to a special register and are not covered by the other registration systems. However hospitals with long-staying patients are covered by the systems. Six long-term hospitals and four specialized hos- pitals in which the average length of stay ex- ceeds 30 days are included in the systems. To- gether these hospitals contain 1,553 beds—4.8 percent of all somatic beds—and they account for 0.5 percent of the discharges and 5.3 percent of the bed days in somatic hospitals.? All patients discharged from the hospitals included in the systems are reported, excepting those treated in psychiatric departments, who are reported to the Psychiatric Register. Mater- nity patients and newborns are reported in the same way as other discharges, except that some obstetric departments collect additional infor- mation. While women with normal pregnancies may choose to deliver their babies at home, in 1977 only 0.6 percent of births took place out- side hospitals and clinics.15 Items Collected The National Patient Register contains vari- ous items about each inpatient discharged from a somatic hospital.!2-14 Identification items in- clude hospital and hospital department codes taken from the Danish National Health Service Hospital Classification list. The patient’s 10-digit Central Persons Register (CPR) number is also reported. All residents of Denmark have a CPR number, which indicates the person’s birth date, sex, and serial number. The health insurance system uses CPR numbers to register consump- tion of all medical care services.16 The potential therefore exists for connecting the records of separate admissions and other uses of medical care. Some work has been done to link a pa- tient’s total contacts with a hospital service for a single disease, but these efforts are limited at present.17 National Patient Register information about the use of hospital services includes the hour and 10 date of admission, and an optional item, the date of referral. The place from which the pa- tient was admitted is recorded (home, other hos- pital department, other hospital, nursing home, old age home, emergency department, other, born in the hospital, or not known). The kind of admission is also reported (acute, through out- patient department for preliminary examination, other cases through outpatient department, other cases called from waiting list, called ac- cording to decision on previous discharge, or born in the hospital). The type of patient is given, that is, inpatient (patient who receives 24- hour treatment), day patient (part-time patient who does not normally spend nights in the hospi- tal), or night patient (part-time patient who does not normally spend days in the hospital). The date of discharge is recorded, as is the place to which the patient was discharged (home, other department, other hospital, nursing home, old age home, convalescent home, other, death, or not known). Whether the patient was discharged alive or dead, and if dead whether an autopsy was performed, are additional items. If the pa- tient was discharged to another hospital or de- partment, the number from the National Health Service Hospital Classification list is recorded. Kinds of aftercare (followed in the same hospital outpatient department, in another hospital’s outpatient department, by patient’s own doctor, other, none, death, and not known) are reported, and there is an optional item concerning any posthospitalization treatment that has been arranged in an institution (same department, other department, other hospital, nursing home, old age home, convalescent home, other institu- tion, no other institutional care, death, or not known). Inpatient characteristics reported to the Na- tional Patient Register, besides birth date and sex (which are part of the CPR number), include marital status, municipality of residence, and a considerable amount of medical data. A number “of diagnoses can be reported for each patient, and several items of information are supplied concerning each diagnosis. First, any modifica- tion of the diagnosis is recorded (none, observa- tion case-diagnosis not proven, observation case- diagnosis disproven, late effects, earlier, recur- rent, treated or under treatment, and operated). An almost unlimited number of operations can be recorded for each diagnosis, provided the entries do not fill space allotted for other diag- noses, and it is possible to indicate whether two or more operations were part of a complex surgi- cal procedure. If a patient underwent an opera- tion in a department other than the one to which he or she was admitted, the number of the operating department may also be given. Finally, there is an item on whether the patient was involved in an accident, and if so, the type of accident (no, traffic accident, occupational accident, sports accident, home accident, other accident including attempted homicide, and optional categories of suicide attempt, and un- certain whether accident or suicide attempt). Diagnoses are coded using an adaptation of the eighth revision of the International Classifi- cation of Diseases, in which codes have been expanded to five digits. A sixth digit may be added for greater detail, and this is sometimes done by specialized departments. In each case the doctor decides which diagnosis should be listed as the most important or primary one. The doctor also decides the order of importance of operations for each diagnosis. An adaptation of , a Swedish four-digit classification system is used to code operations. Fifth and sixth digits can be added for greater detail and to indicate opera- tion complications. Included in the classification system are codes for certain extensive examina- tions and nonsurgical forms of treatment, such as complicated X-ray examinations, biopsies, and cystoscopies. Definitions and Procedures Utilization statistics from the National Pa- tient Register include data for long-term and short-term patients. In addition to long-term hospital patients, long-staying patients within general hospitals, such as those in physiotherapy, convalescent, and long-term care departments, are reported to the register. The bed-day statistics obtained from the register concern the number of days of hospitali- zation of patients discharged during the year. The day of admission is counted as a bed day, but the day of discharge is not. If the hospital stay lasts less than 24 hours, it is counted as 1 bed day in length. Deaths are counted as discharges in calcu- lating utilization statistics, but they can be sepa- rately identified. Transfers between emergency departments or intensive care units and other hospital departments are counted as part of a single admission, but all other transfers between hospital departments are considered discharges and new admissions. However, transfers can be linked through the use of the CPR number. The average length of stay is obtained by dividing the total number of days in the hospital for a given group of discharges by the number of hospital stays in the group. The occupancy rate at a point in time is computed by dividing the number of occupied beds by the official number of beds registered in the hospital and multiply- ing by 100. When the occupancy rate for a cer- tain period of time is desired, the number of bed days used during the period is divided by the number of registered beds multiplied by the number of days in the period, and the resulting figure is multiplied by 100. In some cases all the registered beds are not in use, so the number of available beds is used instead of the number of registered beds. Patient turnover is calculated by dividing the number of patients admitted during a certain period by the average number of beds available in the hospital during that period. Information Published or Available Some data from the various local inpatient registration systems have been available since 1966. Until 1978 the National Health Service published summary statistics from these systems in its annual Medical Report II: Report on Hos- pital and Other Institutions for the Treatment of the Sick in Denmark.!® One example of the tables included in the report is the number of discharges by main diagnosis (100 diagnostic categories), age (14 or under, 15-69, 70 and over), sex, and county; another is the number of operations in each of 16 categories of operations by hospital and hospital department. The National Health Service began a new series of publications in 1972, called Medical Statistics Reports. By 1976 most of the hospital data published in Medical Report II were also published in the Medical Statistics Reports series, so the decision was made to terminate Medical Report II. National data from the registration 1 % systems have not yet been published in Medical Statistics Report. One 1978 publication did contain statistics from the registration systems in three local hospital areas: Copenhagen, Storstrgms, and Ringk¢bing.19 The National Health Service sends unpub- lished statistical tables from the National Patient Register to hospital authorities around the coun- try, but national statistics have not yet become available. Individual registration systems also produce unpublished statistical tables. Each sys- tem provides its users with diagnostic and surgi- cal files prepared cumulatively for each trimester and for each year.12 OTHER DISCHARGE REPORTING SYSTEMS Denmark’s Psychiatric Register began opera- tion before the general hospitals’ registration systems did. Nationwide registration of persons with mental retardation and certain neurological and psychiatric diseases began in the 1920’. Additional psychiatric disorders began to be reported in 1938, and since 1953 all admissions to and discharges from state mental hospitals have been registered. In 1969 the registration system was computerized, and by 1970 all psychiatric institutions were reporting to it.20 The Institute of Psychiatric Demography in Arhus is responsible for operating the Psychi- atric Register. It maintains a computer file on all patients that have received psychiatric treat- ment, and when a patient is admitted to a psy- chiatric facility, the institute sends the facility copies of all information from the patient’s pre- vious psychiatric admissions. The institute regu- larly prepares diagnostic files for participating hospitals and compiles annual reports for the hospitals and the National Health Service.12 The institute also undertakes special research proj- ects concerning the distribution of psychiatric disorders in certain areas and the use of psychi- atric services.21-23 Data Collection The Psychiatric Register receives informa- tion several times during the course of a patient’s hospitalization. An initial report that includes 12 the patient’s name, address, marital status, and other personal data is sent to the institute when a patient is admitted. If the patient remains in the hospital for 3 months, a second report, con- taining the diagnosis, is sent to the institute. The institute also receives a copy of the hospital’s discharge letter to the patient’s general practi- tioner that contains a summary of the case history. A discharge form is also completed, gen- erally by a hospital medical secretary, and it is sent to the institute.20 ~The register covers all inpatients in public or private psychiatric hospital facilities. In addi- tion, psychiatric units in general hospitals— including the psychiatric departments for chil- dren and adolescents—report their inpatients. The register also covers neurosis sanatoria and institutions for alcoholics. Items Available The discharge forms sent to the Psychiatric Register20 include an item that identifies the hospital in which the patient was treated. The patient is identified by his or her CPR number. The date and type of admission are recorded (civil commitment, voluntary commitment, judicial observation, or sentence for custody or treatment), as is the place from which admitted (residence, general hospital psychiatric ward, psychiatric hospital, sanatorium, child psychi- atric ward, somatic ward, or none of the above). Whether the patient was referred from one of the hospital’s outpatient clinics is also noted. The date and type of discharge are reported (alive, died in the hospital, or died during tem- porary absence), as well as the place to which the patient was discharged (residence, general hospital psychiatric ward, psychiatric hospital, sanatorium, somatic ward, child psychiatric ward, or none of the above). Referral to the hospital’s outpatient clinic is recorded sepa- rately. Whether the patient has ever received treatment in a general hospital psychiatric ward, a psychiatric hospital, a sanatorium, a child psy- chiatric unit, or other psychiatric unit is noted, as is whether the patient is a twin. A main and three auxiliary diagnoses can be reported; the patient’s doctor determines the main diagnosis. The same Danish adaption of the International Classification of Diseases that is used in somatic hospitals is used to code the diagnoses. Several tables, routinely produced from the Psychiatric Register,20 consist of information about a detailed list of diagnoses. These lists are available for each hospital and for groups of hos- pitals (mental hospitals, general hospital psychi- atric wards, and sanatoria). The tables give the number of first admissions and all admissions for each diagnosis by sex. The number of admissions with each diagnosis listed as the main diagnosis is reported by sex, as is the number of admis- sions with each diagnosis listed as either the main or a secondary diagnosis. Other tables show 15 main diagnostic categories. The numbers of admissions by sex and age are given for each diagnostic category in one table; in another the numbers of discharges by sex and length of stay are reported for each diagnostic category. The tables are sent annually to the National Health Service. They were published in Medical Report II until it ceased publication in 1978. The last edition of the report contains data per- taining to psychiatric patients hospitalized in fiscal year 1974-75. In 1979 similar tables con- taining data about psychiatric patients hospi- talized in 1975-76 were published separately in the Medical Statistics Reports series.2* A 1978 publication in the Medical Statistics Reports series also covers Psychiatric Register data.25 It reports numbers and rates of inpa- tients in psychiatric treatment facilities on April 1, 1976 (the date when psychiatric state hospi- tals were transferred to the local county admin- istrations). The data were presented by residence in local hospital areas; by type of institution and residence; by age (age 0-14 years, 15-24, 25-44, 45-64, and age 65 years and over) and residence; and by diagnosis, length of stay, and residence. Unpublished data from the Psychiatric Register are sent to individual hospitals quar- terly and annually. The hospitals receive a list of diagnoses that includes the name, CPR number, and other information on each patient.20 AGGREGATE HOSPITAL REPORTS In addition to reporting data to the National Patient Register, Danish hospital staffs complete annual questionnaires for the National Health Service. The questionnaires request information about hospital facilities, costs, personnel, and utilization. General hospitals began using the questionnaires in 1960, and now all Danish hos- pitals use them. Data for a calendar year are reported, and the completed questionnaires are forwarded to the National Health Service by March 1 of the following year. The hospital utilization questionnaire? re- quires the name of the hospital and department, the National Health Service hospital, depart- ment, and specialty code numbers, and the total number of patients treated in the hospital. Sepa- rate data are reported for two groups of pa- tients: 24-hour patients and part-time patients (part-time patients are those who receive only day or night care). The numbers of patients in each group who were in the hospital on the first day of the year and on the last day of the year are recorded. The numbers of admissions, dis- charges, and deaths in each group are also given. A separate count is made of the number of pa- tients in each group who had been in the hospi- tal for the entire calendar year or longer. The total number of bed days used by each group of patients during the year is reported, along with the number of registered beds. Outpatient visits are also counted: The form requires the total number, the number of acute visits, and the number of nonacute visits in the year. The National Health Service uses the same formulas and definitions to calculate utilization statistics from the completed questionnaires as it does for the analysis of National Patient Register data, except that bed days used during the year rather than bed days of discharges are com- puted. The statistics were published in Medical Report II until it was discontinued. Now they are available in publications titled “Output Sta- tistics for the Hospital System,”? which are a regular part of the Medical Statistics Reports series. The publications list all Danish hospitals and departments within hospitals. The following are reported for each department: beds, admissions, discharges, deaths, patients present at the end of the year, bed days, average daily number of pa- tients, occupancy rate, and average length of stay. The numbers of acute, nonacute, and total outpatient visits to each department are also given. A second list presents similar data by 13 hospital region for 15 somatic specialties and for psychiatry and child psychiatry. In addition, the numbers of admissions, discharges, deaths, and patients present at the end of the year are re- ported separately for part-time patients by hos- pital and department. Summary tables show utilization statistics for each hospital region and type of hospital. HOUSEHOLD SURVEYS No major household health survey has been undertaken in Denmark in recent years. The need for periodic surveys has been recognized, especially for those that would gather data about illness and patterns of contact with pri- mary health services.26 However no plans to ini- tiate such surveys have been made.l The only nationwide health survey that has been done in Denmark is the two-part Morbidity Survey of the 1950s. The first part, the 1951-54 Sickness Survey, obtained information on a sample of 87,000 adults living at home. Inter- viewers were asked detailed questions about the diseases they had experienced and about such social characteristics as age, sex, marital status, housing situation, income, and occupation. They also reported any hospitalizations that had occurred during a monthlong period, as well as the hospital’s name and address, the admission date, and the illness or injury for which they were admitted. In 1960 the survey results were published in The Sickness Survey of Denmark.27 The second part of the Morbidity Survey, the 1952-53 Hospital Survey, gathered informa- tion on a sample of 33,000 adults who had been admitted to public medical and surgical hospi- tals. Patients in the State-run university hospital or specialized departments of other hospitals were excluded. Each admission in the sample was reported in part by a doctor, nurse, or secre- tary, and in part by the research staff. Utiliza- tion data included the time spent in the hospital, waiting time before admission, previous admis- sions, and followup care. The reported social characteristics of the patients included age, sex, marital status, occupation, income group, and payment status. Medical data consisted of diag- noses, operations, nursing treatment during hos- pitalization, laboratory tests, X-ray examina- tions, and condition on discharge. In 1959 the results of the survey were published in The Hos- pital Survey of Denmark.28 SCOTLAND Discharge reporting systems in Scotland cover all inpatients treated by the Scottish Health Service. Three reporting systems are used: one for maternity inpatients, one for psy- chiatric inpatients, and one for other hospital in- patients. Also, summary statistics on all types of Scottish Health Service inpatients are collected twice a year. In addition, Scotland is included with England and Wales in the General House- hold Survey, which obtains hospital utilization information. GENERAL HOSPITAL DISCHARGE REPORTING SYSTEM Hospital statistics were first collected in Scotland in the 1940’s. A nationwide Cancer 14 Registration that received data on hospital pa- tients was established in 1945, and certain coun- ties conducted studies of hospital-treated illness under the sponsorship of the Nuffield Provincial Hospitals Trust.2? Not until the Scottish Health Service began, however, were plans made for a comprehensive reporting system. The National Health Service (Scotland) Act of 1947 divided Scotland into five regions; in each a Regional Hospital Board administered hospital and specialist services. The Department of Health, which became the Scottish Home and Health Department in 1962, provided central ad- ministration. In 1950 the regional boards and the Department of Health established a plan to collect national hospital discharge statistics. In 1951 the Northern Region began to collect dis- charge data from its general and maternity hos- pitals, but various constraints prevented the other regions from introducing the system. Over the next few years, individual hospitals in the other regions adopted the reporting system, but not until 1961 were the Scottish Hospital In- Patients Statistics (SHIPS) collected throughout the country. At that point maternity hospitals were excluded from the reporting system, but a separate reporting system for obstetrical inpa- tients was introduced in 1969.29 In 1974 the Scottish Health Service was reorganized. Fifteen health areas were created, each administered by a health board. The newly formed Common Services Agency took over many of the staff functions of the Home and Health Department, and the Information Serv- ices Division of the new agency became responsi- ble for the collection and analysis of SHIPS. Initially the main purpose of SHIPS was to provide hospital use information for the health service’s regional and central administrators. Ad- ministrators were especially concerned with hos- pital usc since hospitals were the most costly part of the health service. It was expected that the collection of individual patient records would allow more flexibility in the analysis of hospital use than was possible with aggregate hospital reports, and that information obtained from the analysis would be very valuable to ad- ministrators who must decide how to best use health resources and plan for the future.29,30 In addition, SHIPS data were expected to provide important epidemiological information. Data on the conditions for which hospital pa- tients were treated and on patients’ demographic characteristics helped to outline some of the more serious morbidity patterns in the commu- nity. The data had to be interpreted with care because only a portion of total morbidity was reported, and cases of treatment rather than in- dividual persons were reported. However the data were still considered useful additions to other sources of epidemiological information.3! After a few years, the potential value of SHIPS to individual hospital managers and clinicians became apparent. In 1968-69 the Scot- tish Consultant Review of In-Patient Statistics (SCRIPS) were introduced. SCRIPS are annual returns that report each physician’s inpatient workload and supply comparative data on other workloads in the country as a whole. The in- tended purpose of SCRIPS was to provide physi- cians with a basis for self-assessment of their patient care practices. This was expected to re- sult in more effective and efficient treatment practices. In addition, receiving individual re- ports could stimulate clinicians’ interest in hos- pital data collection and lead them to improve the data’s accuracy and timeliness.32 The quality of the data has been an ongoing concern. Detailed studies have been done of the extent and type of errors in the data,33:34 and much attention has been given to possible ways of improving the data collection process, includ- ing upgrading recordkeeping within the hospi- tals.35 At present, validity checks in the com- puter system can remove reports of impossible or highly unlikely events, such as hysterectomies in males or senile dementia in children. Much use is made of the data. The Scottish Home and Health Department and health board administrators utilize the data to plan and moni- tor the health service. Many physicians consult the SCRIPS and request additional ad hoc anal- yses from the Information Services Division, and university researchers have increasingly requested special types of data. Methods of Data Collection When an inpatient is discharged from a Scot- tish Health Service hospital or hospital depart- ment, discharge form SMR 1 is completed. While the physician in charge of the case is officially responsible for seeing that the form is accurately completed, medical records and clerical staff ab- stract the patient data, usually with little guid- ance from the physician.30 The records staffs are employed by the health boards; the Information Services Division has no direct managerial con- trol. Often, completing the discharge forms is only one of the records staff’s many responsibili- ties, and delays and backlogs develop. The hos- pital staff codes all the information required on the forms except for the inpatient’s occupation, which is coded centrally.3* In most cases the hospitals send the dis- charge forms to the area health boards, and the boards forward them to the Information Services Division for processing at the national computer center. Two health boards process their own 15 data. Two other computer centers, one in Edin- burgh and one in Glasgow, analyze data for the large hospitals in each of those areas. The local data centers perform the same validity check as is used in national processing and must submit the basic set of standardized data to the national computer center. However they are free to col- lect whatever additional data they wish.3% The national center transfers all discharge data onto computer tapes, performs the validity and feasi- bility checks, and produces the routine output. The tapes are retained in order to meet special requests for tabulations. Coverage SHIPS only cover patients treated by the Scottish Health Service, but more than 95 per- cent of all hospital beds in the country are part of the service. Scottish Health Service maternity and psychiatric patients are not covered by SHIPS regardless of whether they are treated in specialized hospitals or specialized wards of gen- eral hospitals. Data concerning these patients are collected but are sent to separate reporting sys- tems and are processed and tabulated separately. All other inpatients should be reported whether they are in general or specialized hospi- tals or units. The Information Services Division (ISD) has no way to be certain that a form is completed for every discharge, but comparisons between the number of discharge forms received by ISD and the total number of discharges re- ported on the annual aggregate hospital returns have found an unexplained difference of less than 1 percent in number of discharges reported.3! Long-term-care hospitals and units are part of the health service and therefore are included in the reporting system. Since no separate set of institutions similar to U.S. nursing homes exists in Scotland, almost all long-term inpatient care is provided by the health service hospitals.36 The following long-term specialties are covered by the reporting system: rehabilitation and physical medicine units whose patients average 30-day hospital stays, respiratory tuberculosis units with 46-day average stays, geriatric assessment units with 62-day average stays, geriatric long- 16 term units with 345-day average stays, and wards for younger chronic patients with 478-day aver- age stays.37 Items Collected The SMR 1 form37 used to collect hospital statistics contains several identification items, including the hospital code number and the pa- tient’s case reference number. Many hospitals assign a single case number. to a patient for all his or her admissions, but some assign numbers to admissions in sequence or use some other sys- tem, so that patients admitted more than once obtain different case numbers each time they are in the hospital.38 The patient’s last name, first name initial, and last name at birth are also recorded. : Hospital utilization items collected include the date the patient was placed on the waiting list for admission, and the date, source, and type of admission. The date of the principal opera- tion is recorded, as is the date of discharge, whether the patient was discharged alive or dead, the type of bed the patient occupied prior to discharge, and the hospital division or unit from which the patient was discharged or trans- ferred. The physician in charge of the patient’s case is also identified. Social and demographic items collected in- clude age, birth date, marital status, and resi- dence. Residence has been recorded by postal area since” 1974, but much difficulty has been encountered in the use of these codes. The occu- pation of each patient is recorded as well. If the patient is a married woman, the husband’s occu- pation is also recorded. If the patient is a child, the father’s occupation is recorded. Often the husband’s or father’s occupation is not reported. In other cases the information given is impre- cise, but the information is mainly used to assign patients to a social class, and it has been found to be adequate for statistical purposes.30 The medical items include the principal diag- nosis, three other diagnoses, and the external cause of injury (“E”) codes. The principal diag- nosis is defined as the main condition treated or investigated during the admission. If no diag- nosis is made, the main symptom or problem is recorded.37 Since January 1, 1980, the ninth revision of the International Classification of Diseases39 has been used to code diagnoses. A principal operation and one other operation can be reported, and the operations are coded using England’s Office of Population Censuses and Surveys codes. Both operations and diagnoses can be coded to four digits. Definitions and Procedures The statistics routinely produced from the SHIPS data include both long-term and short- term patients. Patients treated in the long-term specialties account for only about 4 percent of the discharges reported to SHIPS, but since they use 37 percent of the total bed days, they have a considerable impact on the statistics. Special analyses are often done of general hospital use by patients age 65 years and over who are the major users of the long-term specialties.4? The bed-day statistics refer to days used by the patients discharged during the yearlong reporting period. Until 1979 admission and dis- charge days were each counted as bed days, but now only the admission day is counted as a bed day.4! The number of discharges includes the num- ber of deaths and transfers. Admissions ended by regular discharges, by deaths, and by trans- fers are often referred to as “spells” in the hospi- tal; data on each type of spell can be obtained separately if desired. Transfers between hospitals and between different specialties within a single hospital are counted as discharges and new ad- missions. In 1976 approximately 8 percent of all spells were interhospital transfers and 3 percent were intrahospital transfers.37 Transfers and re- peat hospitalizations of the same patient can be linked through use of the following identifying data collected on the discharge forms: name, sex, birth date, and sometimes hospital case number.38 However the routine statistics are for spells, not patients. Discharge and bed-day rates are calculated per 1,000,000 population or per 100,000 popu- lation.37,40 Until 1974 discharge rates referred to the population of the patient’s area of treat- ment, but since 1975 they have referred to the \ patient’s area of residence. The average length of stay is obtained by dividing the number of bed days by the number of spells of treatment. Information Published or Available The data collected by SHIPS have been pub- lished annually since 1961 in Scottish Hospital In-Patient Statistics.37 Tables in the publication present statistics on the total number of dis- charges and discharge rates, bed days, bed-use rates, mean length of stay, mean waiting time, and number of surgical operations. These statis- tics are cross-tabulated by various items of infor- mation, including sex, diagnosis, age group (age 0-4 years, 5-14, 15-24, 25-44, 45-64, 65-74, and age 75 years and over), health board area, source of admission (emergency, from the waiting list, booked case, or admission from other sources), condition at discharge (dead or alive), place to which discharged (home, convalescent hospital, other hospital, Local Authority or other care, transfer within the hospital, died, or irregular discharge), hospital division or unit at discharge, and number and type of operations. Five-year trend data are also given on number of discharges and mean length of stay by sex and diagnosis. The annual publication Scottish Health Sta- tistics*0 also presents data from the reporting system. The tabulations are less extensive, but the most recent publication contains tables on the discharge rate by diagnosis over a 7-year period; the discharge rate by diagnosis and health board area; the number of discharges and mean stay by sex, diagnosis, and age (0-4 years, 5-14, 15-44, 45-64, and 65 years and over); and the number of discharges and beds used by area of treatment and area of residence. Unpublished tables are routinely sent to each hospital and health board concerning the hospi- tal’s activity and the activity of the hospitals in the board’s area, respectively. While the tables contain information similar to that which is pub- lished, it is analyzed for each hospital and its divisions or units. The number of discharges is tabulated by age, sex, condition at discharge, source of admission, area of residence, and diag- nosis. The mean length of stay is given by sex, age, and diagnosis; regional and national mean 17 stays are provided for comparison. The number and percent of discharges who were admitted from the waiting list and the mean waiting time are reported by sex and diagnosis, along with regional and national waiting times for each sub- category. The health district of residence of dis- charged patients is given for the hospital as a whole and for each hospital department.3031 In addition, each hospital receives a diag- nostic index consisting of information on all its discharges; the patients are listed by diagnosis. For each patient, the birth date, sex, other diag- noses, physician, month admitted, area of resi- dence, marital status, occupation, social class, source of admission, days on waiting list, days in the hospital, unit from which discharged, condi- tion at discharge, and operations are reported.3! An operation index is also provided. The SCRIPS returns, which are sent to each physician and detail his or her activity, are also compiled from the SHIPS data. One SCRIPS table reports the distribution of cases the physi- cian treated during the year, categorized by diag- nosis. The number of cases, percent of all cases in the specialty, median stay, median waiting time, percent of cases admitted as emergencies, percent of cases discharged to home, percent operated upon, and number of fatalities are given by sex for each diagnostic category. Ex- cept for the number of fatalities, statistics are given in each category for all cases treated in the same specialty in Scotland as a whole. A second table, sent to surgeons, reports the distribution of cases by type of surgical procedure and gives the same statistics for cases in each procedure category (except the percent operated). The per- cent of cases with an admission-to-operation interval of less than 3 days and the median operation-to-discharge hospital stay are also in- cluded. Again, all the statistics are given by sex, and statistics for all cases operated upon in the same specialty in Scotland as a whole are re- ported in each category, except fatalities. A diagnostic index is also provided to physicians that lists all the physician’s discharges by diag- nosis and reports the birth date, sex, up to four diagnoses, external cause of injuries, month of admission, health board of residence, marital status, source of admission, type of admission, days on the waiting list, days of stay, unit from 18 which discharged, disposition, up to two opera- tions, and the case number of each discharge. OTHER DISCHARGE REPORTING SYSTEMS Data about Scottish psychiatric and mater- nity patients are collected by virtually independ- ent discharge reporting systems.d The mental health reporting system, begun in 1963, covers patients in mental hospitals, general hospital psychiatric units, and mental deficiency hospi- tals in the Scottish Health Service. The mater- nity reporting system began to cover health service patients in maternity hospitals and ma- ternity wards of general hospitals in 1969. The mental health system has achieved 100-percent coverage of hospitalized psychiatric patients, and the maternity system covered 99.6 percent of inpatient deliveries in 1977.4! Less than 1 percent of all deliveries in Scotland take place at home. Data are collected on home deliveries in two areas of the country, but the data are not included in the published maternity statistics. The two specialized reporting systems oper- ate in much the same fashion as does the general hospital reporting system. Information ab- stracted from the medical record is entered onto a discharge form for each patient discharged from a participating hospital or unit. The psychi- atric services also complete admission forms on each patient. The medical records or clerical staff codes the forms and forwards them either to the area health board or to the Information Services Division (ISD) of the Common Services Agency. Most psychiatric hospitals send their forms directly to ISD, but two health boards, Grampian and Tayside, computer process their own psychiatric data. Maternity data are usually sent first to the health boards and then to ISD. Most items collected on the mental health and maternity discharge reporting forms are the d1t is not strictly correct to refer to the psychiatric reporting system as a discharge system, since it is based on reporting both at the time of admission and dis- charge, and the main emphasis is on the admission reports. same as those collected on the general hospital reporting forms.34 The hospital and specialty are identified on all the forms, and the patient’s case reference number, name, initials, and birth surname are requested by all. The admission and discharge dates are reported, as is the type of discharge. Social and demographic items on all the forms include birth date, area of residence, and occupation. Common medical items include diagnoses and condition at discharge. The maternity forms require additional in- formation including the mother’s marital status and marriage date, religion, number of previous pregnancies and their outcome, operations, and clinician. Detailed data on the pregnancy, labor, and delivery are requested, and the sex and birth weight of the child are reported.#2 Mental health forms require additional information that con- cerns the patient’s diagnoses on admission and legal status at admission and discharge.34 For the most part, the two specialized re- porting systems use the same definitions and formulas as does the general hospital system. Some additional statistics are calculated for psy- chiatric patients, such as the number of admis- sions, admissions per 100,000 population, number of residents, and residents per 100,000 population. The term “resident” refers to pa- tients present in the psychiatric facility on De- cember 31 of a particular year. The psychiatric system also uses the term “transfer-in” for an admission that results from an inpatient moving from one psychiatric hospital to another.40 Since the early 1970’s data from the mental health reporting system have been published an- nually in Scottish Mental Health In-Patient Sta- tistics.43 The publication containing 1974 data and those that have appeared subsequently are composed of 16 types of tables; most are in sets of 4. The set consists of one table that presents information on all patients in psychiatric hospi- tals and units, and three separate tables that pre- sent the same information for mental hospital patients, psychiatric unit patients, and mental deficiency hospital patients. The information in all the tables is reported separately for males and females. In addition to sex, the publication reports admissions by age (0-14 years, 15-24, 25-34, 35-44, ..., 75-84, and 85 years and over), diag- nosis (detailed and short lists), legal category of admission, health board of treatment, and source of referral (psychiatric outpatient clinic; psychi- atric day unit; domiciliary visit; nonpsychiatric clinic or ward; general practitioner; self, rela- tives, or friends; prison or judicial; Local Authority agency; transfer from other psychi- atric inpatient care; or other sources such as ministers, voluntary agencies, and the like). The number and percent of all admissions that are readmissions are also reported. Statistics on resi- dents in the hospital at the end of the year are reported by age, diagnosis, health board of treat- ment, and duration of stay (less than 1 week, 1 week to 1 month, 1-2 months, 2-3 months, 3-6 months, 6-9 months, 9-12 months, 12-18 months, 18 months to 2 years, 2-5 years, 5-10 years, 10-15 years, 15-20 years, and 20 years or more). Discharges are reported by age, diagnosis, duration of stay, and disposal on discharge (died, left on own accord, home, hostel, trans- ferred to mental hospital, transferred to mental deficiency hospital, transferred to psychiatric unit, transferred to geriatric care, transferred to other inpatient care, penal institution, other, and not known). Time-series data covering an 8-year period are presented on admissions, residents, and discharges. Scottish Health Statistics also presents data from the mental health reporting system. Though less extensive than data in Scottish Mental Health In-Patient Statistics, the data in Scottish Health Statistics are usually more up-to- date. Scottish Health Statistics, 1977,4° pub- lished in 1978, contains tables of provisional data from 1977. The data are separated into two sets of tables: one for mental hospitals and psy- chiatric units together, and the other for mental deficiency hospitals. Reported statistics include the number and rate of admissions, percent read- missions, number and rate of residents, number of discharges, and mean length of stay. Variables used to cross-tabulate the statistics include sex, age, diagnosis, health board of residence, type and category of admission, and type of disposal at discharge. Time-series data covering 1965 to 1977 are also presented. No separate publication is produced from the data collected by the maternity reporting system, but some statistics from the system are 19 included in Scottish Health Statistics. In Scot- tish Health Statistics, 197740 1976 statistics are reported for mothers and newborns. One table presents the number and mean stay of maternity admissions by type of admission (antenatal; postnatal; delivery; and other, mainly abortion) and age (less than 20 years, 20-24, 25-29, ..., 40-44, and 45 years and over). Another table gives the number of discharges, mean length of stay, and percent distribution of maternity discharges in length-of-stay cate- gories (less than 3 days, 3-7, 8-14, 15-28, and 29 days and over) by type of admission (abor- tion; antenatal; delivery, in labor; delivery, not in labor; postnatal; transfers; and others, not pregnant) and type of unit (specialist or general practitioner). The number of maternity dis- charges, percent discharged from obstetric con- sultant units versus obstetric general practice units, and mean stay of discharges are reported by area of residence. The number of deliveries is presented, along with the percent of different modes of delivery (spontaneous, forceps, vacuum extraction, breech, cesarean, and other), the number and percent that were induced, and among induced deliveries, the method used (arm, oxytocins, arm and oxytocins, and others), all according to health board of treatment. The outcome of pregnancy, whether a still or live birth, and birth weight in grams (under 500, 500-1,000, 1,000-1,500, ..., 4,000-4,500, and over 4,500) are tabulated by area of residence. Finally, the number and rate per 1,000 live births of perinatal deaths are given by birth weight and area of residence. Like the general hospital reporting systems, the specialized reporting systems produce un- published tabulations, which are sent to the health boards and hospitals in the systems. In addition, psychiatric clinicians receive returns on the activity of their hospitals. AGGREGATE HOSPITAL REPORTS Besides individual discharge reports, all Scot- tish Health Service hospitals complete aggregate statistical reports of their workloads. The hospi- tals complete the so-called ISD(S)1 forms and send them to their health boards. In some areas 20 the forms are submitted to the health boards weekly; in others, monthly. Twice a year the health boards forward magnetic tapes that con- tain most of the data from the forms to the In- formation Services Division (ISD) of the Com- mon Services Agency. The ISD compiles two annual summaries, one for the year ending March 31 and one for the year ending Sep- tember 30.44 The forms contain information about the total number of available bed days, and from this information ISD calculates the average daily number of beds that were staffed and available for the reception of patients. Since 1976, when form ISD(S)1 was introduced, the calculation of the average daily number of beds has included borrowed, temporary, and lent beds. (Previously such beds were not taken into account; there- fore, adjustments must be made when statistics are compared over time.) Also reported is the number of allocated staffed bed days, which is the sum of the daily number of beds allocated to a specialty that were staffed and available for the reception of the patients during the year. The number of total occupied bed days is given, which is the sum of the number of beds in a specialty occupied at the time of the daily bed count on each day during the year. The bed count generally takes place between midnight and 8 a.m. Since 1976 total occupied bed days has also included borrowed, temporary, and lent bed days. The category also includes patients on temporary leave from the hospital at the time of the bed count. It should be noted that total occupied bed days is not equivalent to total bed days of discharges, which is obtained from the discharge reporting systems. Unlike total occu- pied bed days, total bed days of discharges in- cludes bed days used in a previous year or years if a discharge’s hospitalization extended through more than one reporting period, and it excludes the bed days of patients still in the hospital at the end of the reporting period. The number of inpatients discharged is re- ported on ISD(S)1; this form includes deaths, transfers between hospitals, and transfers be- tween specialties within hospitals, as well as routine discharges. Intrahospital transfers from one specialty to another have been counted as discharges only since 1976 and are also reported separately. Patients, such as day patients or night patients, who are not hospitalized for full 24-hour periods are not counted as discharges. Statistics calculated from the collected data include the average duration of stay, percent occupancy, turnover interval, and throughput. The average duration of stay is obtained by dividing the total occupied bed days by the number of discharged inpatients. The inclusion of intrahospital transfers in the number of dis- charges led to a slight reduction in the average stays reported since 1976. Percent occupancy is the total occupied bed days multiplied by 100 and divided by the number of available staffed beds. Turnover interval is the number of avail- able staffed bed days minus the total occupied bed days divided by the number of inpatients discharged. Throughput is the number of dis- charged inpatients multiplied by 365 and divided by the number of available staffed bed days. Data from the ISD(S)1 forms are reported in the following annual publications: Hospital Utilisation Statistics,!! Scottish Health Statis- tics,%0 and Hospital Bed Resources.” One set of tables in Hospital Utilisation Statistics gives sta- tistics for Scotland as a whole and for each health board by specialty. Almost 50 different specialties are listed, including such medical and surgical specialties as cardiology, urology, and thoracic surgery, and obstetric, psychiatric, and long-term specialties. The tables present the number of approved beds; the number of per- sons on the waiting list; the number of allocated, borrowed, lent, temporary, all available, and total occupied bed days; the number of dis- charges and deaths; and the number of intrahos- pital transfers into and out of the specialty. A second set of tables shows the average number of allocated staffed beds, the number on the waiting list per allocated staffed bed, the num- ber of allocated staffed beds per 100,000 popu- lation, throughput, mean stay, turnover interval, and the number of discharges per 100,000 popu- lation for each specialty by health board. Sepa- rate sets of tables with similar formats report on inpatients in joint-user hospitals, which are ‘Local Authority institutions that make beds available to health boards, and on inpatients in contractual hospitals, which are institutions operated by voluntary bodies in which the health boards use beds. Data are also presented on out- patient visits to health service hospitals. Many of the same statistics are included in Scottish Health Statistics, again tabulated by specialty and health board. Hospital Bed Re- sources contains information on available staffed beds and bed complements of all Scottish hospi- tals and the total bed stock in each health dis- trict of the country. HOUSEHOLD SURVEY Scotland has been included in the General Household Survey since the survey began in 1971. The ongoing survey is conducted by the Social Survey Division of the Office of Popula- tion Censuses and Surveys (OPCS), which is located in England. It covers five main subject areas: population, housing, employment, educa- tion, and health. The data from the survey are expected to supply government agencies with in- formation that will assist resource allocation de- cisions. The survey results have routinely been sent to about a dozen government departments, and researchers outside the government have shown increasing interest in them. Data Collection The survey data are collected from a large sample of households in Great Britain. In 1977, 15,315 households were included.4? Interviews are conducted with all adult household mem- bers, who also furnish information about their children under age 16. Institutionalized indi- viduals are excluded from the study. The sample is drawn in a two-stage process, first sampling electoral wards and then selecting addresses within each ward from the Electoral Register. The samples are stratified by type of area (met- ropolitan or nonmetropolitan), by socioeco- nomic group of the head of the household, and by the proportion of householders who are owner-occupiers. Items Available A great deal of information is collected about the social and demographic characteristics of the individuals in households covered by the survey. Age, sex, marital status, length of time at 2 present address, type of housing, skin color, country of birth, family size, type of employ- ment, income, education level, and region are investigated. Information about the utilization of ambulatory and inpatient health services is also collected, as are data on the incidence of acute and chronic sickness. The information on hospitalization includes a question about inpa- tient care during the 3 months prior to the inter- view and the length of stay in the hospital. If a person is currently on a waiting list for hospital admission, the length of the wait is reported. The OPCS has published the survey results in a series titled General Household Survey.*5 In 1979, data collected in 1977 were published as the seventh number in this series. In recent years the information in the publication on hospitali- zations has been limited to tables of the number of medical and surgical inpatient visits per 1,000 persons in a 3-month reference period, and the average number of inpatient nights per visit, both given by age group and sex. The published data are for Great Britain as a whole, though sometimes statistics for England and Wales are reported separately. Unpublished statistics on Scotland are available and can be obtained, sub- ject to certain restrictions, in the form of either tables or magnetic data tapes. WEST GERMANY In West Germany hospital discharge data are available from a variety of sources. Sickness in- surance funds collect information about their members’ hospitalizations. One small West Ger- man State, Schleswig-Holstein, operates a dis- charge reporting system in its acute care hospi- tals. In another State, Northrhine-Westphalia, a psychiatric hospital reporting system has been organized. In addition, army hospitals have a special reporting system. Official hospital statis- tics for the country as a whole are collected by means of aggregate hospital reports by the Fed- eral Statistical Office in cooperation with State statistical offices. The Federal Statistical Office also operates the Microcensus, an ongoing na- tional household survey that collects some infor- mation about hospitalization. GENERAL HOSPITAL DISCHARGE REPORTING SYSTEMS The sickness insurance funds’ data systems and the Schleswig-Holstein reporting system are the main sources of national and regional diag- nostic data for patients treated in general hospi- tals. The sickness insurance funds and their hos- pital data systems are discussed first in this section, followed by a description of the Schleswig-Holstein system. 22 Sickness Insurance Funds Compulsory health insurance was established in Germany in 1883. At that time a number of voluntary sickness benefit societies already ex- isted in the country. While the insurance law required certain workers to join a society or fund and established regulations concerning the operation of the funds, it allowed the funds to remain independent entities. The insurance sys- tem has since expanded, but continues to be administered by autonomous funds. In 1976 there were 1,425 sickness insurance funds in operation.#6 The funds are governed by boards of directors, whose members equally represent employers and workers. The State gov- ernments and the Federal Ministry of Labor and Social Affairs supervise the funds, which cover approximately 96 percent of the population.4? Membership in a fund is compulsory for most workers, including all manual and salaried workers whose income is below a certain level, a level that changes periodically. Other workers can join funds as voluntary members. The fami- lies of workers are not, in a strict sense, mem- bers of the funds, but insurance coverage does extend to them. There are eight types of sickness insurance funds: Local Sickness Funds, Company Sickness Funds, Guild Sickness Funds, Agricultural Sick- ness Funds, Seaman’s Sickness Funds, Miner’s Sickness Funds, Compensation Sickness Funds for Workers, and Compensation Sickness Funds for Employees. Over half of the insured workers in West Germany belong to the first of these, the Local Sickness Funds.#® Workers usually join the Local Sickness Fund in their area, unless they are employed by a large business that has its own fund, belong to a guild with a fund, or are involved in an occupation that operates a fund. Many of the funds are organized into asso- ciations at the State and Federal levels by type of fund; such as State and Federal associations of the Local Sickness Funds. The sickness insurance funds collect reports of the hospital use of members and their fami- lies. Reports are made of all insured hospitaliza- tions, whether the patients are treated in the acute care hospitals or in the special care hospi- tals, which primarily treat long-staying patients. The data are not separated by type of hospital. Additional information, including the age, sex, and diagnoses of insured hospital patients, is col- lected by the funds on a sample basis. The diag- noses are coded according to the International Classification of Diseases.*? The individual sickness insurance funds send the collected data to the State sickness fund associations, which forward the data to the Fed- eral associations. The various Federal associa- tions supply the data to the Ministry of Labor and Social Affairs, which adds it to its Sozial- datenbank (social data bank). Some hospital-use statistics from the data bank are published regu- larly, for instance in Survey of Social Security;50 unpublished data are available for special studies. The different Federal associations of the sickness insurance funds also publish annual re- ports. These reports contain more detailed infor- mation about hospital use than does the Survey of Social Security but are less comprehensive, since each association’s report only covers the hospitalizations of its members and their fami- lies. An example of an association report is the publication of the Federation of Local Sick- ness Funds on types and causes of illness, and deaths.5! The 1979 edition of this publication contains 1977 statistics from 267 Local Sick- ness Funds that represent 90 percent of all the fund members. The participating funds reported data from a 20-percent sample of the hospital cases that they handled. The data are presented separately for various membership categories (compulsory members, voluntary members, pen- sioners, and members’ families). The number of hospital cases and bed days, average number of days per case, and number of cases and bed days per 10,000 members are shown for each mem- bership category, by sex and diagnosis. The same statistics are given for each membership category except families of members and are divided by sex, diagnosis, and age (14 years or under, 15-19, 20-24, ..., 60-64, and 65 years and over). The same statistics are also reported for compulsory members and are divided by sex, age, and cause of illness (work accident excluding road acci- dent; road accident; occupational disease; traffic accident excluding road accident; sports acci- dent; other accident; murder, assault, battery; suicide, suicide attempt, self-inflicted injury; and military service injury). A major advantage of the sickness insurance fund diagnostic statistics is that the size of popu- lation to which they refer is known. The number of persons with compulsory membership in a fund is established annually and is available by age. The number of persons in the members’ families is determined every 4 years.46 Thus a denominator exists for epidemiological statistics. Schleswig-Holstein In addition to sickness insurance fund hospi- tal statistics, discharge data have been collected by the State Statistical Office in Schleswig- Holstein since 1969. Schleswig-Holstein is one of West Germany’s 10 States. It is located in the northern part of the country, bordering Den- mark, and contains 2.5 million people, 4 percent of the country’s population.52 The Schleswig-Holstein reporting system re- ceives data only from acute care hospitals. In the first year of operation, 34 hospitals partici- pated in the system, and they accounted for 45.6 percent of all patients treated in the State’s acute care hospitals that year. Almost every year since, additional hospitals have joined the sys- tem. By 1977, 53 hospitals, accounting for 70.1 percent of the State’s acute care hospital 23 patients, participated. While participation is vol- untary, all the acute care hospitals in the State are expected to take part in the system in the near future.53 The reporting system was established to pro- vide needed data for health planning and hospi- tal management, and as a possible basis for epi- demiological research. It was thought that the collected information would greatly assist the management of public programs concerning health, especially those to design health care delivery systems that fit needs of the popula- tion. Individual hospital’s procedures, coverage, and efficiency were also expected to be better understood by means of the collected data. The system has served as a model for other State and Federal Government officials who are interested in establishing comprehensive hospital data sys- tems elsewhere in West Germany. The data from the system have been found to have certain disadvantages. A current epide- miological study of the relationship between leukemia and a complex of nuclear power plants serves as one example. The study is hampered because persons with leukemia often receive treatment in hospitals, especially university hos- pitals, that have not been participating in the reporting system. Furthermore, since the statis- tics count cases rather than persons, advances in the treatment of leukemia that have reduced the death rate and that have led to repeated brief periods of hospitalization have resulted in hospi- tal statistics that appear to show increased mor- bidity where no real increase exists.>* The re- porting system’s inability to obtain accurate utilization rates of cases per population will be a major problem until complete coverage of the acute care hospitals is reached. Even then the residents of Schleswig-Holstein who are treated in other States’ hospitals will be excluded from the system, and the residents of other States treated in Schleswig-Holstein will be included, which will complicate epidemiological studies. The fact that the statistics refer to cases rather than to persons is a problem common to most discharge reporting systems and will remain a problem unless a system of record linkage is developed. Methods of data collection.—Information is collected continuously in the Schleswig-Holstein 24 reporting system. When a patient is admitted to a participating hospital, his or her data sheet is begun. Further information is added when the patient is discharged. A physician usually codes the sheets; occasionally other hospital personnel provide assistance.54 The data sheets can be sent directly to the State Statistical Office, but many hospitals have data processing equipment and transfer the in- formation from the sheets to punch cards or magnetic tapes. Thus the State office receives data in various forms. The office produces a single data set from all the material sent to it, processes the data by computer, and performs consistency checks (for instance, checking sex- specific diseases to ensure that they are reported for the proper sex). Sets of tables, some of which are published annually, are then produced by computer. Unpublished tables are returned free of charge to the participating hospital. Several proposals have been made to increase the system’s scope. For example, about half of the hospitals that participate in the reporting system, including almost all the public hospitals, are also associated with a data processing system that emphasizes financial management and in- surance claims. This system is maintained at the Schleswig-Holstein Data Center, which is where the State Statistical Office maintains its data bank. If the specialized data processing system were slightly broadened, its data could be added to the State’s diagnostic statistics. Another plan involves a new reporting form, suggested by the State Statistical Office, that would combine diagnostic data with hospital data from other sources and would allow the information to be readily transformed into machine readable form.54 Coverage. —As mentioned earlier, the Schleswig-Holstein reporting system only covers the State’s acute care hospitals. The special hos- pitals, which primarily treat long-staying pa- tients, do not participate. All types of acute care hospitals may join the system, and by 1976 all the public hospitals owned by the State or its cities were participating, along with all denomi- national hospitals and some of the other private hospitals.5? The participating hospitals may not totally represent all acute care hospitals in the State, but patients’ average length of stay in the participating hospitals is the same as in all Schleswig-Holstein acute care hospitals.8:56 As more hospitals join the reporting system, any existing biases will be corrected. The participating hospitals report all their discharges. Therefore patients in psychiatric wards are covered as are other long-staying pa- tients, such as those in tuberculosis units who averaged 93-day stays in 1976.56 Maternity pa- tients are reported in the same way as other in- patients, and separate data sheets are also com- pleted for infants born in the hospital. The data collected by the reporting system cannot be considered representative of hospitali- zation in West Germany as a whole. Schleswig- Holstein has one of the lowest bed-to-population ratios in the country, and Schleswig-Holstein acute care hospitals have a lower average dis- charge rate, bed-day rate, mean length of stay, and occupancy rate than do all other West German acute care hospitals.3 Items collected.—The data sheet used in the discharge reporting system contains several iden- tification items. The hospital and the specialty department in which the patient is treated are identified by numbers. A patient admission number is also given. The patient’s accommo- dation (one-bed room, two-bed room, room with more than two beds) is reported. If the reason for the hospital stay was not the treat- ment of an established diagnosis but rather an opinion about the patient’s condition (an admis- sion by order of a third party, usually the courts or insurance authorities) this is noted. The type of payment for the hospitalization is also re- ported (self-payment; payment by an insurance fund, and if so, what kind of fund; payment by welfare; or other type of payment). Items concerning hospital utilization include the admission and discharge dates. The kind of admission is coded as newborn, transfer from another department, transfer from another hos- pital, or other. The kinds of discharges include discharge to home, transfer to another depart- ment, transfer to another hospital, transfer to a nursing home, died and autopsy performed, and died and autopsy not performed. Social and demographic items are the pa- tient’s age, sex, and residence. The day and month of birth are reported for children under 1 year of age; for others only the year of birth is required. The district of residence should always be reported. The community of residence is an optional item and usually is not reported. Special note is made of patients who are out-of- State residents. Three diagnoses can be reported and are coded to three or four digits using the eighth revision of the International Classification of Diseases. The primary diagnosis, defined as the main illness for which the patient was treated, is recorded first. The doctor in charge of the hos- pital ward decides which is the main diagnosis when there are more than one. Each diagnosis is described as either a final diagnosis, a provisional diagnosis, or a “state after.” “State after’ means that the treatment was not for the illness itself, which no longer existed, but for a condition that occurred after or was due to the illness. Definitions and procedures.—The collected data do not differentiate short-staying and long- staying patients. Most long-staying patients in the State are treated in the special hospitals, however, and data are not collected in these hospitals. Since statistics are divided by type of hospital department, the statistics for depart- ments that provide treatment to long-staying patients, such as the tuberculosis departments, can be isolated. : The term “completed cases” is used to refer to regular discharges, deaths, and transfers.54 Transfer patients are counted as completed cases whether they are interhospital or intrahospital transfers. Death statistics are usually presented separately as well as together with other com- pleted cases. The number of healthy newborns is not included in the number of completed cases, but the number of sick newborns is included. A bed day is counted for each patient present in the hospital at midnight.#7 The day of admission would therefore be counted as a bed day, but the day of discharge would not. A hospital stay of less than 1 day would not be counted as 1 bed day unless the patient was in the hospital at midnight. The average length of stay is computed by dividing the number of bed days of completed cases by the number of completed cases. The standard deviation associated with the average length of stay is also calculated, and the 25 coefficient of variation, obtained by dividing the standard deviation by the mean length of stay, is computed.56 Information published and available. —Statis- tics from the reporting system are published by the State Statistical Office 2 years after the year they were collected. Unpublished statistics are generally available to participating hospitals with less delay. The 10 university-affiliated hospitals in the system usually receive data about their pa- tients 1 year after collection. Some tabulations are also sent to all participating hospitals in May following the end of the year the data were collected.54 Published statistics from the reporting sys- tem are available in Diseases of Inpatients in Schleswig-Holstein 56 Its two major tables pre- sent information on a list of selected diagnoses that cover 90 percent of all diagnoses. Both tables contain the number of completed cases and deaths, the average length of stay, and the standard deviation and coefficient of variation associated with the average length of stay for the selected diagnoses. Each table also reports the number of cases, bed days, and average length of stay by diagnosis and age (less than 1 year, 1-14, 15-44, 45-64, and 65 years and over). One table presents this information separately for the diag- noses in various hospital specialties (surgery; gynecology; obstetrics; internal medicine; infec- tious diseases; pediatrics; urology; orthopedics; ear, nose, and throat; psychiatry and neurology; dermatology and venereology; ophthalmology; teeth and gums; and radiation) and for diagnoses in hospitals without specialties. Another set of tables presents information on 18 diagnostic groups. The number and per- cent of completed cases and deaths, the number of bed days, and the number of bed days per case are given by sex and diagnostic group. The number of completed cases is also reported by age group, sex, and diagnostic group. The per- cent of the cases in each age and diagnostic group is given for all patients and then sepa- rately for males and females. Within each age group the percent distribution of cases in the diagnostic groups is also given, first for all pa- tients in each age group and then separately for males and females. 26 Four tables give statistics that are not sepa- rated by diagnosis. One reports the number of beds covered in the system and the percent of covered beds among all beds by type of hospital and hospital department. The second table pre- sents the number of cases and the number of bed days per case, the latter separated by sex, by type of hospital department. The third table gives the number of discharges, the average stay by sex, the percent of patients, and the percent of the population, by age group. The fourth re- ports the number of deaths that occurred in each type of hospital department. Unpublished tables sent to each hospital in- clude lists of cases treated by the hospital and by separate hospital departments according to the main diagnosis. For each case the admission number, sex, age, and additional diagnoses are reported. Other unpublished tables present ag- gregated data for the hospital as a whole and for the separate departments, by diagnosis, age, and average length of stay. Also reported are the number of admissions and discharges cross- tabulated by health insurance group responsible for payment and type of accommodation.55 OTHER DISCHARGE REPORTING SYSTEMS Psychiatric Reporting System While West Germany does not have a national psychiatric discharge reporting system, psychi- atric patient data are collected in one State, Northrhine-Westphalia. The State, located on the west side of West Germany, borders the Netherlands and Belgium. Its population, 17 million persons, is the largest of the 10 States.52 The psychiatric reporting system covers the ad- ministrative regions of Diisseldorf and Cologne, which together contain about 9 million people. Data collection.—The psychiatric reporting system is run by the Rhineland Provincial Union, a nongovernment association, and covers 10 association-owned psychiatric hospitals. The hospitals had a total of 10,627 beds in 1980, including beds for mentally handicapped pa- tients.?7 The hospitals were collecting statistics, including some diagnostic statistics, even before World War II. Computer data processing began in 1960, and in 1976 the hospitals installed data terminals for direct access to the association’s central computer. Information about each patient is collected at the times of admission and discharge. Hospital personnel add the information to the computer daily. Some data are processed by individual hospitals, but most are processed centrally by the association administrative staff. Transfers to other hospitals and deaths are considered discharges. Intrahospital transfers are reported but do not count as discharges. Admis- sion and discharge days are counted as separate bed days, and hospital stays of less than 24 hours are counted as 1 bed day. Total bed days for a year are computed by summing the num- ber of patients in the hospital each midnight. Patients who are on leave for short periods, such as holidays, are regarded as present for the mid- night count. Mean length of stay is calculated by multiplying the number of bed days by 2 and dividing the result by the total number of admis- sions and discharges. The statistics produced by the reporting system generally include both long- staying and short-staying patients, but short- staying patients can be identified separately.57 Items available.—A large amount of informa- tion is collected for each patient admitted to one of the 10 psychiatric hospitals in the system. The hospital and department to which the pa- tient is admitted are identified, as are the pa- tient’s name and admission number. Also reported are the date and time of admission, the legal basis of the admission, the place from which admitted, the source of admission (such as prac- ticing physician, psychiatric clinic, or police), the number of admissions, the time since last ad- mission, and whether the admission followed a suicide attempt. Patient characteristics recorded include birth date, sex, nationality, marital status, religion, occupation, place of residence, and residential status (such as lives alone, lives with siblings, or lives with spouse and child). Three admission diagnoses can be reported, and since 1972 the International Classification of Diseases has been used to code the diagnoses. Detailed information about the source and type of payment for the hospital stay is also given. If a patient is transferred from one hospital ward to another, the date of the transfer, the ward to which moved, and up to three transfer diagnoses are noted. If the legal basis of the hos- pitalization has changed, the new legal basis is also recorded. When a patient is discharged the collected information includes the date and hour of the discharge, up to three discharge diagnoses, the place to which discharged, and to whom the patient’s care was transferred (general practi- tioner, neurologist, other specialist, or other). Each year the Rhineland Provincial Union publishes a statistical report called Data, Facts and Trends.’® The 1979 edition contains 1978 summary statistics from each psychiatric hospi- tal and various statistical tables of the character- istics of the psychiatric patients for all the hos- pitals. The statistics for each hospital include number of beds, total admissions, first admis- sions, transfers-in, total patients treated, total discharges, transfers-out, deaths, patients in the hospital on December 31, bed days, average number of patients in the hospital per day, mean length of stay, and occupancy rate. Each hospital reports the number and percent of admissions and discharges in seven diagnostic categories, five age categories, and five legal categories. The number of admissions, treated patients, and dis- charges from each postal region the hospital serves are also reported. All information is given for the hospitals together as well as separately, and changes in several of the statistics from 1971 to 1978 are shown. Most of the tables on patient characteristics give the number and percent of admissions in sets of categories by sex. For instance, the num- ber and percent of admissions in marital status categories (single, married, divorced, widowed, and unknown) are given for males, females, and all patients. Tables show the distributions of ad- missions by age, number of admissions, time since last admission, residential status, source of admission, place from which admitted, whether suicide attempt, and occupation. The number and percent of admissions are also given by legal status and diagnosis, by age and diagnosis, and by age and legal status. The number and percent of treated patients, discharges, and deaths are 27 reported for length-of-stay categories, and the number and percent of discharges are given by place to which discharged. One table shows the number of admissions and discharges, and the average number of admissions per day for each month of the year. Other tables show informa- tion about patients with addictions, such as type of addiction, by age group. Army Reporting System A special discharge system that exists in West Germany’s army hospitals was begun in 1960 by a government agency called the Insti- tute of Defense Forces’ Medical Statistics. The reporting system was established to learn more about army hospital utilization and morbidity patterns among soldiers. Since West German army hospitals and patients are quite different from other West German hospitals and patients, the reporting system’s statistics are not at all representative of nationwide patterns of hospital use. The reporting system covers all army hospi- tals. Individual hospitals complete reporting forms for each discharged patient and send the forms to the institute for coding and processing.?9 The reporting forms include items that iden- tify the hospital, department, and ward. The pa- tient’s name, hospital number, army grade, army activity, and first enlistment are reported, as are the patient’s birthplace, occupation, next of kin, religion, marital status, and sex. The date, time, and type of admission (first, readmission with same illness or injury, readmission for other ill- ness or injury, determination of fitness grade, and other) are listed. Intrahospital transfers, the departments transferred from and to, and the number of days of inpatient treatment in each department are reported. The date of discharge, place to which discharged, and total number of days of inpatient care are also given. A primary diagnosis, four additional diag- noses, and up to five operations can be reported on the forms. The diagnoses are coded using a system similar to the International Classification of Diseases. The codes for operations were de- veloped specifically for the army hospitals. If the patient died, the date, time, cause of death, and whether an autopsy was performed are 28 given. If the patient is disabled, the presumed cause of the disability and whether it was in- curred in the line of duty are reported. The type of fitness for duty is also given. An overall sum- mary of the patient’s illness is written, and it is signed by the ward physician and the physician in charge of the patient’s department. The forms are completed for both long- staying and short-staying patients, and the statis- tics produced from the forms concern both types of patient. Discharge statistics include deaths and interhospital transfers but not intra- hospital transfers. Bed-day statistics refer to the number of bed days used during the year, but mean length-of-stay statistics are computed using the number of bed days ot discharges, which is divided by the number of discharges. The admission and discharge days are counted together as 1 bed day, and hospital stays of less than 24 hours are counted as 1 bed day. Statistical reports are published monthly; another publication is prepared annually. The statistics are also supplied to the Federal Statisti- cal Office for inclusion in its annual publication about hospitals. AGGREGATE HOSPITAL REPORTS Official hospital utilization statistics in West Germany are compiled by the Federal Statistical Office from aggregate hospital reports. All hospi- tals report, whether they are public, nonprofit, or profitmaking institutions. The hospitals re- ceive annual questionnaires from the State Sta- tistical Offices. Each State uses a somewhat different questionnaire, and the definitions of terms used to complete the questionnaires vary from State to State, but the Federal Statistical Office has established a committee to develop uniform forms and terminology.*’ Hospital staffs complete the questionnaires with informa- tion about the hospitals’ operation and utiliza- tion during the calendar year. Usually the infor- mation comes from daily summaries of the hospitals’ activities, the so-called “midnight sta- tistics.” The completed forms are sent to area health departments, which aggregate the in- coming data and report the results to the State Statistical Offices. The State offices then forward the data to the Federal Statistical Office, which is responsible for processing and publishing them. The Federal Statistical Office calculates separate utilization statistics for the acute care hospitals and special hospitals, which generally contain long-staying patients. As a group, pa- tients in acute care hospitals had an average stay of 15.8 days in 1977, while patients in special hospitals averaged 58.7-day stays.® However some types of hospitals that were part of the acute care group reported average stays that ex- ceeded 30 days, and some special hospitals reported average stays of less than 30 days. Transfers between hospitals are always con- sidered discharges and new admissions, but transfers between departments within a hospital are counted as part of a single admission. Deaths are combined with other discharges for the pur- pose of computing utilization statistics, but they are also reported separately. A bed day is counted for each patient in the hospital at mid- night. The total number of bed days thus refers to the number used during the year, not bed days used by patients discharged during the year. The average length of stay is calculated by multiplying the total number of bed days by 2 and dividing by the total number of admissions and discharges. The rates of discharges and bed days per unit of population generally are not computed, but State or nationwide rates can be calculated from the published statistics.60 The statistics from the aggregate reports are published annually in ‘“Hospitals,”® which is number 6 in the Health Care Statistics series published by the Federal Statistical Office. The 1979 edition of “Hospitals” presents 1977 data, which primarily concern hospital beds. For in- stance, the number of hospitals and beds are shown by State, type of hospital ownership, hos- pital size, hospital specialty, and hospital regions within States. The number of beds per 10,000 population are given by State and type of hospi- tal ownership. The number and types of hospital personnel are reported in another set of tables. Hospital utilization statistics are presented in three tables: one for males, one for females, and one for both sexes. The number of patients hos- pitalized at the beginning of the year, admis- sions, total patients treated, discharges, deaths, patients hospitalized at the end of the year, bed days, and average length of stay are reported in each table by type of hospital ownership, hospi- tal specialty, and State. The average occupancy rates of hospitals are also reported by hospital specialty, hospital ownership, and State. An additional table presents data on hospital births and shows the number of women who gave birth and the number who experienced complications, the number of newborns and whether the new- borns were live or dead, the bed days and aver- age lengths of stay of all maternity patients and of those with complications, and the number of inpatients who experienced miscarriages. This information is reported by State and hospital ownership. The Federal Statistical Office publishes simi- lar tables in its monthly statistical publication, Economics and Statistics.5! The tables are usually less detailed than those in “Hospitals” but contain more recent statistics. The statistics in both publications are used to document the existence, location, and level of hospital use, which are necessary to know for administrative purposes. However the information is not suffi- cient for detailed studies of hospital operation, including studies of cost effectiveness, that are important for hospital management and planning.55 HOUSEHOLD SURVEY West Germany began experimenting with population health surveys in 1963.53 Additional health surveys were undertaken in 1966, 1970, and 1972. In 1973 a section on illnesses, acci- dents, and handicaps was added to the Micro- census, an ongoing household survey conducted by the Federal Statistical Office. The Micro- census has operated since 1957 and primarily collects household demographic, social, and economic data. The data have been used for eco- nomic studies and to update the regular census. The Microcensus is conducted four times each year, but the health questions are included only once a year. The Microcensus samples cover almost all of the West German population; only military personnel and institutionalized individuals are 29 excluded. The sampling ratio is between 0.25 and 1.0 percent of the population. Study areas are chosen at random using as a sample frame all the regions covered in the regular census. Every household within a chosen study area is part of the sample. One person is interviewed in each household and answers questions about the health of all the household members. Inter- viewers are lay volunteers, recruited and trained by the State Statistical Offices. The items included in the health section of the Microcensus have changed somewhat over time but have always focused more on illness and disability than on the utilization of health services. For example, the 1978 questionnaire included a list of questions for a household member who had suffered an illness, accident, or handicap during the 4-week period prior to the interview. If a person had experienced more than one health problem, the questions addressed the most serious problem. Whether the illness was chronic, the length and type of illness, and whether the person was still sick when the inter- view took place were asked. The length of time the person was unable to work due to the illness, accident, or handicap was recorded. The ques- tionnaire also reported each household mem- ber’s height and weight, whether he or she smoked, and if so, what and how much he or she smoked. Questions about the use of health services also referred only to the most serious health problem experienced in the 4 weeks prior to the interview. Whether the household member saw a physician for the problem was noted, as was the type of physician (general practitioner or spe- cialist), and whether the visit was to a hospital outpatient department. Hospital stays of at least 1 night were also reported. The definition of a hospital excluded institutions supervised by a single doctor that do not offer regular medical treatment, such as homes for the elderly. Microcensus statistics are published in the Federal Statistical Office Health Care Statistics series. They are also available in such publica- tions as Data from the Health Care System, pub- lished by the Federal Ministry of Youth, Family, and Health. While most published statistics do not concern hospitalization, some information is provided. For instance, tables in Data from the Health Care System, 197762 show the per- cent of sick persons who had been hospitalized by sex and age group; by sex and type of illness; by sex, age group, and ability to work; and by sex, type of illness, and ability to work. It was expected that the Microcensus statis- tics would outline the health status of the West German population and thereby provide infor- mation to develop need-based planning of the health care system. Some researchers and policy- makers have not found the statistics to be as useful as was hoped. The use of nonmedically trained interviewers and the procedure of asking questions about only one health disorder have been criticized. Changes in the questions that are asked about health have made time-trend studies difficult. Moreover switches in the time of year the questions are asked (from May to October) could introduce seasonal variation to the find- ings. It has been suggested that a separate health interview survey might produce more useful health data, and discussion of such a separation has been taking place.4? U.S. NATIONAL HOSPITAL DISCHARGE SURVEY In the United States a number of national, State, and local data systems collect hospital utilization statistics. This chapter is concerned with the National Hospital Discharge Survey. This survey is the major source of national esti- mates of short-stay hospital utilization. Other important discharge reporting systems, which are more limited in scope than the National Hos- pital Discharge Survey, are discussed in appen- 30 dix II. Also in appendix II are brief descriptions of the main national hospital and household sur- veys that collect hospital utilization statistics. The National Center for Health Statistics (NCHS) established the U.S. National Hospital Discharge Survey in 1965.2 Previously NCHS had collected some hospital use information through its household Health Interview Survey, begun in 1958.63 Information about hospital Table G. General hospital discharge reporting systems, by country and reporting system General hospital discharge reporting system Denmark Scotland West Germany United States Inpatient registration systems Scottish Hospital In-Patient Statistics Schleswig-Holstein hospital morbidity study Insurance fund statistical systems . National Hospital Discharge Survey Agency or agencies responsible .................. Yor DBgUNY...corrucanisssersy 1 Main publications™.......... National Health Serv- ice, Association of County Councils, local and regional computer centers 1970 Series publications in Medical Statistics Reports [6,19] Information Services Division of the Com- mon Services Agency, Scottish Health Service 1951—one region, 1961—entire country Scottish Hospital In- Patient Statistics [37] State Statistical Office, Schleswig- Holstein 1969 Diseases of Inpatients in Schleswig- Holstein [56] Sickness insurance funds Annual reports of Federal associations of sickness funds, National Center for Health Statistics 1965 Series 13 publications in Vital and Health Statistics [66,70-73] Scottish Health Medical Report 11: Statistics [40] Report on Hospitals and Other Institutions for Treatment of the Sick in Denmark [18], published until 1978 such as //iness-type, llIness-cause, and Detailed Diagnoses Death Statistics [51] and Surgical Proce- dures for Patients Survey of Social Discharged From Security [50] Short-Stay Hospitals [75] 1Numbers in parenthesis are for references, which give full bibliographic information on the publications. births and deaths was also reported on birth and death certificates. In 1962-63 NCHS developed the Master Facility Inventory (MFI), a compre- hensive list of hospitals, nursing homes, and other inpatient health facilities in the United States.54 As well as being a source of national statistics on the number, type, and geographic distribution of inpatient facilities, the MFI was expected to serve as a sampling frame for sur- veys of specific types of facilities and their users. A continuing survey of hospital discharges, which uses the MFI as the sampling frame, was in the planning stages when the MFI was devel- oped. In 1964 a sample of 95 hospitals was drawn from the MFI for a pilot study of hospital discharges.55 In 1965 a master sample of 690 hospitals was drawn from the MFI, of which 315 were inducted into the National Hospital Dis- charge Survey. An additional 150 hospitals from the master sample were inducted into the survey during the first years of its operation; by 1969 a total of 465 hospitals was involved. In 1972 and every 2 or 3 years since, the master sample has been sup- plemented by a “birth sample” drawn from lists of new hospitals added to the MFI since 1965. In 1978 there were 535 hospitals in the survey sample, and of these, 413 participated in the survey. The survey was designed as a continuing general purpose study of hospital utilization pat- terns. It was not established to answer any single question or to provide data to any particu- lar group. The collected statistics are available to the public and are used by government agencies, health policymakers, university researchers, hos- pital supply companies, and a variety of other groups and individuals. The National Hospital Discharge Survey and the discharge reporting systems in Denmark, Scotland, and West Germany to which it is com- pared are listed in table G. The table shows that the U.S. survey is like the Scottish system in one respect: It is the responsibility of a single na- tional agency. No one national agency in either Denmark or West Germany has a similar respon- sibility. National data are compiled in both countries, but each has a number of reporting systems that are operated by separate agencies. The National Patient Register in Denmark re- ceives data from the separate registration sys- tems, and the Sozialdatenbank in West Germany 31 receives data from the various sickness insurance funds. METHODS OF DATA COLLECTION One way that the National Hospital Dis- charge Survey differs from the other three countries’ reporting systems is that it uses a two- stage sampling design. Sampling is not part of the design of the reporting systems in Denmark, Scotland, or Schleswig-Holstein in West Ger- many. All hospitals covered by these systems are expected to supply abstracts of information for all discharges except those that are specifically excluded from the systems, such as psychiatric discharges in Denmark and Scotland. The sick- ness insurance funds in West Germany also col- lect some information on all the discharges cov- ered by their reporting systems, but they obtain more detailed data on samples of discharges. The samples are drawn in a one-stage process from the universes of all discharges covered by par- ticular types of sickness insurance funds. In con- trast, a sample of hospitals is drawn from the universe of hospitals covered by the U.S. survey, and then a sample of discharges is selected from each sample hospital. The original sample of hospitals for the U.S. survey included all the hospitals in the universe with 1,000 beds or more. The hospitals in the universe with less than 1,000 beds were divided into primary strata by bed size (6-49 beds, 50-99 beds, 100-199 beds, 200-299 beds, 300-499 beds, and 500-999 beds) and region (Northeast, North Central, South, and West). Hospitals within the primary strata were further classified by type of ownership and more detailed geo- graphic divisions. In 1965 a controlled selection technique was used to draw the master sample of hospitals from these hospital classes. A sys- tematic random sample method was used to sup- plement the master sample. The sampling proba- bilities vary from certainty for the largest hospitals to 1 in 40 for the smallest hospitals.2:66 The sample hospitals’ discharges are sampled with probabilities that vary inversely with the probability of selection of the hospital. In hospi- tals with 1,000 beds or more, only 1 percent of discharges are sampled; 40 percent of the dis- 32 charges are sampled in some of the smallest hos- pitals.67 The sample of discharges is randomly selected, usually by using the last digit or digits of the patient’s medical record number. If the hospital’s list of discharges does not show medi- cal record numbers, every kth discharge on the list is selected for the sample, beginning with a discharge chosen at random. An abstract form is completed for each sample discharge. The information for the ab- stract is taken from the face sheet of the pa- tient’s hospital record. In about two-thirds of the hospitals the hospital. staff completes the abstracts. In the remaining hospitals the forms are completed by U.S. Bureau of the Census per- sonnel, acting for NCHS. All hospitals send com- pleted abstracts to a census regional office for review. Each form is checked for completeness, and then all are sent to NCHS. The NCHS staff codes the diagnoses and surgical operations or procedures listed on each abstract. When the * coding is done, the data on the forms are trans- ferred to computer tapes, and the tapes are edited and processed. These data collection procedures differ from the procedures used in the other three countries in the areas of central coding of medical infor- mation and central data processing. Discharge forms are completed and coded by hospital per- sonnel in Denmark, Scotland, and Schleswig- Holstein. The one exception is the occupation item, which is centrally coded in Scotland. In- formation about the coding procedures of the West German sickness insurance funds was not obtained. Not all data processing is centralized in Denmark and Scotland. Local and regional computer centers process discharge data in Den- mark, but the centers also provide the National Health Service with data tapes. In Scotland two area health boards and two local computer cen- ters process data, in addition to the national computer center. COVERAGE Not all U.S. hospitals are within the scope of the National Hospital Discharge Survey. Ex- cluded are institutional hospitals, such as prison hospitals and university student health centers, as well as all Federal hospitals, such as military and Veterans’ Administration hospitals. Hospi- tals with less than six beds and those in which the average length of stay of all patients is 30 days or more are also excluded. The sample of discharges drawn from participating hospitals sometimes excludes patients treated in long-term care units if the units keep records separately from the rest of the hospital. All other dis- charges are sampled, but the survey reports usually exclude data for newborns. The coverage of the discharge reporting sys- tems in the other three countries is compared with the coverage of the U.S. survey in table H. It is important to note that none of the report- ing systems in the other countries excludes hos- pitals on the basis of patients’ average length of stay. In Scotland and Denmark psychiatric hos- pitals are excluded from the reporting systems; but almost all other long-term hospitals are in- cluded. The Schleswig-Holstein study covers acute care hospitals, but these hospitals are defined by the type of treatment they provide, not by the patients’ average length of stay. While most of the acute care hospitals in West Ger- many report average patient stays of less than 30 days, not all do, and a few special care hospitals report average stays of less than 30 days. The sickness insurance funds in West Germany obtain data from all hospitals, those classified as acute and special care. Several different patterns of coverage exist in regard to psychiatric patients. The general dis- charge reporting systems in Denmark and Scot- land not only exclude patients in psychiatric hospitals but also patients treated in psychiatric units of other hospitals. Both countries have specialized reporting systems that cover the psychiatric hospitals and units. The Schleswig- Holstein study excludes all patients in psychi- atric hospitals, since the hospitals are classified as special care hospitals, but the study covers patients discharged from psychiatric units in acute care hospitals. The West German sickness insurance funds collect information on any psychiatric patient whose hospitalization was covered by fund insurance, whether the patient is discharged from a psychiatric hospital or a psychiatric unit in another type of hospital. The U.S. survey covers patients discharged from psychiatric hospitals in which the average length of stay is less than 30 days. Patients discharged from hospital psychiatric units are also covered if the units are in hospitals that are within the scope of the survey and if they are not long-term units with separate record systems. Table H. Comparability of coverage of general hospital discharge reporting systems, by country and reporting system re Coverage Denmark Scotland West Germany United States National Patient Scottish Hospital Schleswig-Holstein Insurance fund National Hospital : In-Patient hospital morbidity statistical y Register Statistics study systems Discharge Survey System COVers.........cceueeees Ninety-six percent of Scottish Health Serv- Seventy percent of Hospitalizations cov- Sample of discharges all general and special- | ice hospitals discharges from acute ered by sickness insur- | from sample of short- ized somatic hospitals care hospitals in ance funds stay hospitals System does not cover as of 1979 Five private general and specialized so- matic hospitals as of 1979 Psychiatric hospitals Psychiatric units in nonpsychiatric hospitals Private hospitals and private beds in Scot- tish Health Service hospitals Psychiatric hospitals Psychiatric units in nonpsychiatric hospitals Maternity patients Newborns Schleswig-Holstein as of 1977 Thirty percent of dis- charges from acute care hospitals in Schleswig-Holstein as of 1977 Special care hospitals, including psychiatric hospitals, in Schleswig-Holstein Hospitalizations not covered by sickness insurance funds Long-stay hospitals Federal hospitals Hospitals with less than six beds Institutional hospitals Some long-term units in short-stay hospitals 33 Another important difference in coverage is the exclusion of maternity patients and new- borns from the general reporting system in Scot- land. A special reporting system covers Scottish maternity patients and newborns. Newborns are within the scope of the other reporting systems, but only in Denmark are they routinely counted as discharges in published reports. Private hospitals are not fully represented in some of the systems. Hospitals that have not yet joined the reporting systems in Denmark and Schleswig-Holstein are all privately owned. The small number of private hospitals in Scotland are excluded from the Scottish reporting system, as are patients discharged from private beds in Scottish Health Service hospitals. The U.S. sur- vey and the West German sickness insurance funds’ statistical systems cover private hospitals. The funds’ statistical systems are the only ones that exclude uninsured patients, however, and the U.S. survey is alone in excluding Federal hospitals and hospitals with less than six beds. ITEMS COLLECTED The abstract forms used for the 1977 Na- tional Hospital Discharge Survey contain the items listed in table J. The hospital and patient are identified by numbers on the forms, which are used only for reviewing and processing the data. No information that would permit identi- fication of individual hospitals or patients is released from the survey. In Denmark, Scotland, and Schleswig-Holstein not only are hospitals identified on the forms, statistics for individual hospitals are tabulated. The tabulations are pri- marily for the use of the individual hospitals. Few of the reporting systems’ routine publica- tions present data separately by hospital. Most of the other items collected by the U.S. survey are also collected in the other three countries. The major exception is the item on race, which is only collected in the United States. Other exceptions are the expected source of payment, which is only collected in the United States and Schleswig-Holstein, and the date of surgical procedure(s), which is only collected in ‘the United States and Scotland. In addition, ‘the Schleswig-Holstein study does not obtain in- 34 formation about marital status or surgical procedures. The content of the item on patient disposi- tion differs among reporting systems. For the U.S. survey, the types of dispositions include routine discharge, discharged home; left against medical advice; discharged, transferred to another facility or organization; discharged, re- ferred to organized home care service; died; and not stated. The information about patient dispo- sition in the other reporting systems includes whether or not the patient was transferred to another department within the hospital. Trans- fers to other facilities or organizations are also specified in more detail in the other reporting systems. For instance, a transfer to another hos- pital is reported separately from a transfer to a nursing home in all three countries. The Den- mark and Schleswig-Holstein system also report whether or not a patient who had died was autopsied. The U.S. survey is, however, the only one of the four to collect information on trans- fers to home care services, and only Scotland and the United States note irregular discharges, such as leaving against medical advice. Some items are collected by the reporting systems in Denmark, Scotland, or Schleswig- Holstein, but not by the U.S. survey. For in- stance, the other reporting systems obtain in- stance, the other reporting systems obtain information about the source or type of ad- mission. The content of the item(s) varies. In Scotland the categories of admissions include emergency and from waiting list. In Schleswig- Holstein the categories include transfer from another department, transfer from another hos- pital, and newborn. In Denmark categories simi- lar to those in Scotland and Schleswig-Holstein are used, and others are included, such as from nursing home and through outpatient department. Another item not included in the U.S. sur- vey, but identified in Denmark and Scotland, is the hospital department in which the patient was treated. In Denmark the hour of admission is also included, as is the kind of accident suf- fered by patients admitted for treatment of acci- dental injuries. In Scotland the physician in charge of the patient’s case is identified, the pa- tient’s occupation is given, and the date the Table J. Comparison of items collected in the United States! with those collected in Denmark, Scotland, and West Germany, by reporting system Item PIOSDIAl IYUIMDIBE ..uviierecriviarmmsmsrsresarsssssssssmssorsrnersasssssssasssspsssssnsssnsninss Case number ........... Date of admission Date of discharge Disposition of patient Region of residence............... Expected source of Peyment. Principal diagnosis................ Other diagnoses....... Principal surgical procedure .. Other surgical procedures......... Date of Surgical. PrOCBAUIBIS) .......covseisssssinscssvmsmasssnsrssssnspmsssssnsnssssins Denmark Scotland West Germany? National Patient Scottish Hospital Schieswig-Holstein Register In-Patient hospital 1979 Statistics morbidity study 1977 1979 edna Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No No No Yes Yes No Yes Yes Yes No No Yes Yes Yes Yes Yes Yes Yes Yes Yes No eet Yes Yes No ARE No Yes No 11977 U.S. National Hospital Discharge Survey. 2Detailed information was not obtained about the items collected by the sickness insurance funds in West Germany. patient was placed on the waiting list is noted. The Schleswig-Holstein study includes an item about the patient’s accommodation, that is, a one-bed room, two-bed room, or room with more than two beds. The diagnoses reported in the Schleswig-Holstein study are also qualified as being final, provisional, or ‘‘state after” (which refers to a condition that occurred after or was due to the illness). The principal diagnosis and additional diag- noses are coded somewhat differently by the reporting systems. As seen in table K, the princi- pal diagnosis is defined in various ways. In the U.S. survey the principal diagnosis is defined as the condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital.66 The Scottish reporting system uses the definition recom- mended in the ninth revision of the Interna- tional Classification of Diseases,?® and the definition used in the Schleswig-Holstein study is similar. Both refer to the main condition or illness treated, which is not necessarily the con- dition responsible for admission. In Denmark the physician decides which condition is most important, and it is coded as the principal diag- nosis. The criteria that Danish physicians use to determine the importance of conditions are not specified. Thus the most important condition may or may not be the condition responsible for admission or the main condition treated. All the reporting systems use the Interna- tional Classification of Diseases (ICD) to code diagnoses, but each utilizes a different variation. In Denmark and Schleswig-Holstein, the eighth revision of ICD is used, but the National Patient Register uses a special Danish adaption. The Scottish system and the U.S. survey have begun using the ninth revision of ICD, but in the United States the clinical modification (ICD- 9-CM)b8 is employed. Three- or four-digit codes are utilized in Scotland and Schleswig-Holstein, four- to five-digit codes in the United States, and five- to six-digit codes in Denmark. In Denmark and Scotland medical records staffs code data; specially trained NCHS staff code the U.S. sur- vey data. In Schleswig-Holstein physicians usu- ally code the diagnoses. 35 Table K. Coding of diagnoses in general hospital discharge reporting systems, by country and reporting system Coding of diagnoses Denmark Scotland West Germany 1 United States National Patient Register Scottish Hospital In-Patient Statistics Schleswig-Holstein hospital morbidity study National Hospital Discharge Survey Definition of principal diagnosis COMB «ovrorrississirasssssaresrisrssssninasis COS USB... co caiisnsrssansssnrssnsrrssrnes De1aiL.OF COUR. ivvrarirmsmrisssssssnisbones Coding done bY ......cvmimimrrmsises The condition the physician decides is most important International Classifi- cation of Diseases, eighth revision, Danish adaption Five to six digits Medical secretaries in individual hospitals The main condition treated or investigated during the hospital stay International Classifi- cation of Diseases, ninth revision, as of 1980 Four digits Medical records staff in individual hospitals The main illness for which the patient was under treatment International Classifi- cation of Diseases, eighth revision Three to four digits Usually physicians, sometimes other hos- The condition estab- lished after study to be chiefly responsible for occasioning the patient's admission to the hospital International Classifi- cation of Diseases, ninth revision, clinical modification, as of 1979 Four to five digits National Center for Health Statistics staff pital personnel, in individual hospitals 1petailed information was not obtained about the coding of items by the sickness insurance funds in West Germany. DEFINITIONS AND PROCEDURES The statistics most commonly produced by discharge reporting systems concern discharges, bed days, and average lengths of stay. Some dif- ferences in the way these statistics are computed exist among the reporting systems. The U.S. National Hospital Discharge Survey defines a dis- charge as the termination of a period of hospi- talization by death or by the disposition of the patient to his or her place of residence, a nursing home, or another hospital.66 The reporting sys- tems in Denmark, Scotland, and Schleswig- Holstein also regard deaths and releases of pa- tients to their homes or other institutions as discharges, but in addition these reporting sys- tems label transfers from one department or specialty to another within a hospital as dis- charges. The effect of including intrahospital transfers in the number of discharges probably varies from country to country. Only the Scottish reporting system publishes statistics about intrahospital transfers. In 1976, 3 percent of all the discharges reported by the Scottish 36 system were transfers from one specialty to another within a single hospital.37 In one respect, the discharge reporting sys- tems count bed days the same way: Each system sums up the bed days accumulated by dis- charged patients during a yearlong reporting period. This procedure does not count the bed days of patients still in the hospital at the end of the reporting period, but if a discharged patient’s hospitalization extended through more than the 1-year reporting period, all the bed days used in the previous year or years are counted. It is im- portant to be aware that this is not how bed days are usually computed for the annual ques- tionnaires on utilization that are completed by hospital personnel in most countries. Instead, the total days of care supplied by the hospital during the year is usually reported. The total days of care includes bed days used by patients who are still hospitalized at the end of the reporting period and excludes bed days used in previous reporting periods. The two procedures for computing bed days result in nearly equiv- alent bed-day statistics for a year when most of the patients reported are hospitalized for short periods, but the procedures are likely to result in different bed-day statistics when many long- staying patients are covered.5? The bed days of a discharge are computed in the U.S. survey by counting all the days from the date of admission to the date of discharge. The day of admission is counted as 1 bed day, but the day of discharge is not. Stays of less than 1 day are counted as 1 bed day. The report- ing systems in Denmark and Scotland follow the same procedures in counting bed days, but until 1979 both the admission and discharge days were counted as bed days in Scotland. In Schleswig-Holstein a bed day is counted for each midnight the patient was in the hospital. This procedure usually results in counting the admis- sion day as 1 bed day, the discharge day and hospital stays of less than 1 day as 0 bed days. Average length of stay is calculated in the U.S. survey by dividing the total number of bed days of patients discharged during a year by the number of patients discharged. The same for- mula for average length of stay is used by the reporting systems in Denmark, Scotland, and Schleswig-Holstein, but the resulting statistics are not completely comparable with the U.S. statistics because of the different ways dis- charges and bed days are computed. The most important variation is in the pre-1979 Scottish statistics on the average length of stay. Counting the discharge day as a bed day increases the average stay by a full day. However counting intrahospital transfers as discharges decreases the average length of stay. In Scotland the decrease in the average stay for all patients was only a fraction of a day in 1976, but the de- crease could be greater in Denmark or Schleswig- Holstein if intrahospital transfers were more common. The average length of stay reported for Schleswig-Holstein is also somewhat reduced by the failure to count 1 bed day for hospital stays of less than 1 day. INFORMATION PUBLISHED OR AVAILABLE Statistics from the National Hospital Dis- charge Survey have been published in NCHS’s Vital and Health Statistics series 13 reports since 1966. Three reports are usually published from survey data collected during each calendar year. One type of report presents summary statistics on hospital utilization. The summary reports for 1965-73 contain only nonmedical statistics,’? but subsequent reports include some statistics on diagnoses and surgical operations. More de- tailed information about hospital utilization by diagnosis is presented in a second report,’! and a third report supplies more detailed data on surgical operations.’2 In addition to these annual reports, special analyses of the survey data are sometimes presented in separate publi- cations, such as in a recent report on hospital utilization of persons with alcohol-related diagnoses.”3 A recent annual report, ‘Utilization of Short-Stay Hospitals, Annual Summary for the United States, 1978,”66 contains the following detailed tables: 1. Number, percent distribution, and rate of patients discharged from short-stay hospitals, by sex and age. 2. Number, percent distribution, and rate of days of care, average number of hospital beds occupied daily, and aver- age length of stay for patients dis- charged from short-stay hospitals, by sex and age. 3. Number and percent distribution of patients discharged from short-stay hos- pitals by age and length of stay, accord- ing to sex. 4. Number and percent distribution of patients discharged from short-stay hos- pitals by color and age of patient, ac- cording to sex. 5. Number and percent distribution of days of care for patients discharged from short-stay hospitals by color and age of patient, according to sex. 6. Average length of stay for patients dis- charged from short-stay hospitals, by color, age, and sex. 7. Number of patients discharged from short-stay hospitals and days of care, by sex, age, geographic region, and bed size of hospital. 37 = 132, 38 10. 11, 13. 14. 15. 16. 17. 18. Rates of patients discharged from short- stay hospitals and of days of care, by geographic region, age, and sex. Average length of stay for patients dis- charged from short-stay hospitals, by geographic region, age, and sex. Average length of stay for patients dis- charged from short-stay hospitals, by sex, age, geographic region, and bed size of hospital. Number of patients discharged from short-stay hospitals and days of care, by type of ownership of hospital and age and sex of patient. Average length of stay for patients dis- charged from short-stay hospitals, by type of ownership of hospital, age of patient, and sex. Number of patients discharged from short-stay hospitals, rate of discharges, and average length of stay, by category of first-listed diagnosis and age. Number of discharges and average length of stay for patients discharged from short-stay hospitals, by category of first-listed diagnosis, sex, and color; and rate of discharges by category of first-listed diagnosis and sex. Number of patients discharged from short-stay hospitals, rate of discharges, and average length of stay, by category of first-listed diagnosis and geographic region. : Number of patients discharged from “short-stay hospitals and average length of stay, by category of first-listed diag- nosis and bed size of hospital. Number of all-listed diagnoses for pa- tients discharged from short-stay hospi- . tals, by diagnostic category and age, sex, ‘color, geographic region, and bed size of hospital. Number of all-listed operations for pa- tients discharged from short-stay hospi- tals, by surgical category, age, sex, and color. 19. Rate of all-listed operations for patients discharged from short-stay hospitals, by surgical category, age, and sex. 20. Number of all-listed operations for pa- tients discharged from short-stay hospi- tals, by surgical category and geographic region. 21. Rate of all-listed operations for patients discharged from short-stay hospitals, by surgical category and geographic region. 22. Number of all-listed operations for pa- tients discharged from short-stay hospi- tals, by surgical category and bed size of hospital. Tables 1 and 2 present statistics for the fol- lowing age groups: under 1 year, 1-4 years, 5-14 years, 15-24 years, 25-34 years, 35-44 years, 45-54 years, 55-64 years, 65-74 years, and 75 years and over. In other tables only the age groups of under 15 years, 15-44 years, 45-64 years, and 65 years and over are used. In tables 3-12 statistics are given separately for the cate- gories of females including deliveries and females excluding deliveries, but other tables show only one category for females. Statistics on color are given for two main groups: white and all other. The all-other group includes all categories other than white. In tables 4-6 an additional category is reported—color not stated—and the patients in this category accounted for more discharges and bed days in 1978 than did the patients in the all- other category. In tables 16, 17, and 22 the hospital bed-size categories are 6-99 beds, 100-199 beds, 200-299 beds, 300-499 beds, and 500 beds or more. In tables 7 and 10 the cate- gories are 6-99 beds, 100-499 beds, and 500 beds or more. The geographic regions reported in the tables are Northeast, North Central, South, and West, which correspond to the regions used by the U.S. Bureau of the Census. Tables 13-16 present statistics on first-listed diagnoses. The first-listed diagnosis is the one identified as the principal diagnosis or else the one listed first on the face sheet of the medical record. All-listed diagnoses, reported in table 17, are the first-listed diagnosis and up to four other diagnoses listed on the face sheet of the medical record.66 The statistics on diagnoses are pre- sented for broad groupings of diseases and injuries, which correspond to classes I- XVII of the Eighth Revision International Classification of Diseases, Adapted for Use in the United States’ (ICDA), the classification system used for the survey during 1970-79. Some statistics "are also presented for subcategories. For in- stance, statistics are given for class VI, Diseases of the Nervous System and Sense Organs (ICDA 320-389), and for the subcategories of Diseases of the Central Nervous System (ICDA 320-349), Cataract (ICDA 374), and Diseases of the Ear and Mastoid Process (ICDA 380-389). All-listed operations, reported in tables 18-22, are the first three operations listed on the face sheet of the medical record, excluding cer- tain procedures and treatments not generally considered surgery.66 The statistics on opera- tions are reported for classes 1-17 of the ICDA section, Surgical Operations, Diagnostic and Other Therapeutic Procedures; and statistics are given for some subcategories of the classes. The annual reports that concentrate on hos- pital utilization by diagnosis present statistics for a more extensive list of subcategories within each ICDA class than do the summary reports. For instance, while statistics are presented for 4 subcategories within class IX, Diseases of the Digestive System, in the 1978 summary re- port, 56 statistics are given for 15 subcategories of the class in the 1975 report on hospital utili- zation by diagnosis.’! Similarly, the report on surgical operations presents statistics for detailed lists of subcategories within the classes of operations.”?2 In addition to Vital and Health Statistics series 13 reports, NCHS published data in 1979 from the National Hospital Discharge Survey in a report titled Detailed Diagnoses and Surgical Procedures for Patients Discharged From Short- Stay Hospitals: United States, 1977.75 The re- port presents statistics for each ICDA code. For instance, statistics are given for over 150 codes for class IX, Diseases of the Digestive System. The statistics are shown by age and sex of pa- tients discharged, conditions diagnosed, and sur- gical procedures performed. Other publications of this type are expected to be produced using data from subsequent years of the survey’s operation. Statistics from the survey are also published in such reports as Health, United States’® and The Nation’s Use of Health Resources.”” Un- published data are available in the form of data tapes and special tabulations, which are prepared to meet specific requests for information. Re- strictions placed on the release of the data are only those related to the confidentiality of insti- tutions and individuals. The main sources of published data from the reporting systems in the United States, Denmark, Scotland, and West Germany are listed in table G. Most of these publications present hospital utili- zation statistics by diagnosis. Typically the num- ber and rate of discharges and bed days and the average length of stay are given for categories of diagnoses by age and sex. The diagnoses are grouped somewhat differently in each publica- tion, but all use the International Classification of Diseases as the basis for the groupings. The number of age groups also varies; the statistics on diagnoses may be given for as few as three age groups or as many as seven. Discharge and bed-day rates are computed using different defi- nitions of the population. For instance, tle U.S. rates are calculated using the civilian noriinstitu- tionalized population, in Scotland the total population is used, and the West German sick- ness insurance funds present rates per 10,000 members. Diagnostic statistics are also given by other types of categories. Geographic regions consti- tute a common category, but there is; consider- able variation among other types of categories for which data are presented. The U.S. publica- tions are the only ones to show statistics on diagnoses by color and bed size. In D enmark and Schleswig-Holstein diagnostic statistics are given by type of hospital department. The sickness insurance funds in West Germany’ show diag- nostic statistics by type of membership in the funds, and in Scotland diagnostic statistics are given by source of admission and type of patient disposition. Statistics on different types o'f surgical oper- ations are not available in West Germany but are published elsewhere. Statistical comparisons are hampered not only by different groupings of the operations but also by the use of different clas- sification systems to code operations. The United States is the only one of the countries to use the ICD codes for operations. It should also be pointed out that not all the publications present national data. In Den- mark the last national data published in the 39 discontinued series of medical reports, Medical Report II, were for 1974-75. Only data from local areas of the country have been published thus far in the Medical Statistics Reports series. In West Germany the main sources of data are either the Schleswig-Holstein publications, which only cover one State, or the publications from the several associations of sickness insurance funds, each of which only covers hospitaliza- tions insured by a particular type of fund. The Scottish publications, like the U.S. publications, present national data. SUMMARY The U.S. National Hospital Discharge Survey and the reporting systems in Denmark, Scotland, and West Germany are similar in some ways and different in others. Abstracts of information for discharged patients are completed in each re- porting system, and the abstracts contain many similar items. Information about patients’ diag- noses is collected in each system, and each uses a variation of the International Classification of Diseases to code the diagnoses. The reports pub- lished by the systems usually provide statistics on diagnoses by age and sex. However the re- porting systems cover different sets of hospitals and patients. The procedures used to collect data vary, and the procedures for computing sta- tistics are dissimilar. Although the reporting sys- tems’ similarities allow cross-national compari- sons of hospital utilization by diagnosis, the systems’ differences require that the statistics be carefully adjusted before the comparisons are made. HEALTH SERVICES SYSTEMS In addition to understanding different dis- charge reporting systems, those who intend to compare hospital utilization data among coun- tries should have some knowledge of the coun- tries’ health services system characteristics. Certain characteristics are likely to significantly affect hospital use rates. For instance, the num- bers and types of available long-term-care facili- ties can aiffect hospital use. Patterns in the provi- sion of ambulatory and home care can be important, and the costs of receiving hospital and other health care services should be consid- ered. These aspects of the health services systems in Denmark:, Scotland, West Germany, and the United States are discussed briefly in this sec- tion. Further information about the health services systems in each country can be found in various publi cations.416,48,78-84 LONG-TERM-CARE FACILITIES Cross-national comparisons of hospital utili- zation data need to take into account differ- ences in the types of facilities that provide long-term care. When long-term care is primarily the responsibility of the hospital system, higher 40 bed-day rates and higher average lengths of stay are likely to be reported than when other types of long-term-care facilities exist. The number of beds per population in long- term hospitals is one indicator of the extent to which the hospital system supplies long-term care. Table D shows that the United States has fewer beds per 1,000 population in long-term hospitals than does Denmark, Scotland, or West Germany. However, as seen in table B, the United States also has a smaller percent of the population age 65 years and over (the major users of long-term-care facilities) than do the other three countries. If long-term hospital beds per 1,000 population age 65 years and over are compared, the United States remains low with about 13 beds per 1,000 persons age 65 years and over; while Denmark has about 17; West Germany, 26; and Scotland, 40. These statistics do not include beds in long- term units of short-term hospitals. All four countries have some beds in such units, and the number has been increasing in Denmark and Scotland. In Denmark the units usually have been established for aged and chronically ill in- dividuals who require more intensive medical supervision than that available in nursing homes.85 Some units in Scottish hospitals treat elderly patients who require extensive medical and nursing care, while other units provide a level of care similar to that available in nursing homes in other countries.36 The statistics on long-term hospital beds in Scotland also do not include beds in facilities for the mentally deficient, but those facilities are considered part of the hospital system. In 1977- 78 there were 19 hospitals for the mentally deficient in Scotland, and they contained 6,635 beds.” The statistics do include beds in joint-user hospitals, which are hospitals where local social services authorities and the health service share facilities, and beds in contractual hospitals, which are private institutions that make agree- ments with the health service to provide care. In 1977-78 there were 22 joint-user and contractual hospitals in Scotland, and they contained 1,531 beds, almost all of which were for long-term care. Bed statistics for West Germany’s long-term hospitals include beds in institutions called the Kurkrankenhiuser, which do not have close counterparts in the other three countries. The Kurkrankenhauser may be equivalent to rehabili- tation centers, extended care facilities, or rest homes. They are usually owned by pension funds and are primarily located in resort areas.46:82 In 1977 the Kurkrankenhauser accounted for 57 percent of the discharges from West German special care hospitals and had an average length of stay of 30.1 days.8 Each country has some long-term care facili- ties outside the hospital system. The United States has a large number of beds in such facili- ties. In 1976, 20,468 nursing homes operated in the United States, and they had a total of 1,414,865 beds, which was 6.5 beds per 1,000 population, or 61.7 beds per 1,000 population age 65 years and over.? There were also 6,280 other inpatient institutions with 375,805 beds in the United States in 1976, including facilities for the mentally retarded, the emotionally disturbed, dependent children, unwed mothers, alcoholics, and others. Comparable data on long-term-care facilities in the other countries are not available, but Den- mark is said to have a good supply of beds in institutions outside the hospital system.87 The institutions include nursing homes; convalescent homes; rehabilitation centers; facilities for the mentally retarded, the blind, the deaf, and others with handicaps; and old age homes. Most facilities for the handicapped are the responsi- bility of the national government, but other in- stitutions are either provided or supervised by municipal social welfare authorities. The rela- tionships between the hospitals and these insti- tutions have been problematic because the latter are administered by different committees at the local level and different ministries at the na- tional level. However there is growing recogni- tion of the need to reduce the burden on hospi- tals of long-staying patients who do not require hospital treatment, and each county is now expected to set up a joint hospital-welfare com- mittee to facilitate collaboration between the two systems of care. Fewer types of long-term care facilities exist outside the hospital system in West Germany. A nursing home system exists but is considered inadequate.88 Approximately 1 percent of per- sons age 65 years and over are in nursing homes, but it is estimated that 2 percent need nursing home care. Old age homes are more prevalent than nursing homes; approximately 4 percent of the elderly reside in them. These homes do not routinely supply medical care, though, and plans call for them to be replaced by sheltered housing in the community. As mentioned earlier, Scotland does not have a nursing home system. Some long-term pa- tients receive care in homes run by local social services authorities or by voluntary organiza- tions. In 1972, 390 of these homes contained over 13,000 beds,®9 and the number of beds has been steadily increasing. Most are residential facilities and provide only minimal medical services. Thus the four countries differ widely in the extent to which their hospital systems are responsible for long-term care. The Scottish hos- pital system includes the most long-term-care facilities, and the West German hospital system includes a larger proportion of such facilities than do the hospital systems in Denmark or the United States. The U.S. hospital system proba- bly includes the smallest proportion of long- term-care facilities, but further data are needed to study this point. 4 AMBULATORY AND HOME CARE SERVICES Hospital use may be affected by several aspects of ambulatory and home care services. The supply of physicians and the way they are organized are likely to have an effect. The avail- ability of nurses for ambulatory care and home nursing programs is expected to be important, as is the degree of development of home support services. The number of physicians per 10,000 popu- lation does not vary greatly among the four countries. In 1977 Scotland had the lowest num- ber, 16.7 physicians per 10,000 population;?° West Germany had the highest, 20.3 per 10,000 population.46 The United States reported 17.9 physicians per 10,000 population in 1977;76 Denmark had 19.5 physicians per 10,000 popu- lation in 1976.90 In the United States 15 percent of profes- sionally active physicians were in general or family practice in 1977, and the rest reported a specialty practice. Most physicians, 64 percent, were office based; that is, they spent the greatest amount of their time in practices based in pri- vate offices. Another 27 percent were hospital- based; that is, salaried hospital physicians; and 9 percent were primarily involved in teaching, administration, research, and other non-patient- care activities. About 24 percent of physicians not employed by the Federal government were in group practices in 1975, an increase from 18 percent in 1969.76 In the United States ambulatory and hospi- tal care are generally provided by the same physician. Physicians in office-based practices usually can admit their patients to hospitals and attend them in the hospitals. Physicians in hospital-based practices also provide a certain amount of ambulatory care. Some see ambula- tory patients in part-time private practices, and some provide ambulatory care in hospital out- patient departments and emergency rooms. In Denmark approximately 30 percent of physicians were in general practice in 1975, and the rest were employed by hospitals and pro- vided specialized care.8’7 The number and per- cent of specialists has been increasing over the last 25 years, but the number of general practi- 42 tioners has remained relatively constant. General practitioners are remarkably evenly distributed geographically. Every area with a population of 2,330 plus or minus 10 percent has a general practitioner in its midst.16 Traditionally, hospital and ambulatory care have been quite separate in Denmark. Until recently hospitals maintained very few outpatient departments, and hospital physicians still have only limited possibilities for private practice. Most ambulatory care is provided by general practitioners, who cannot follow patients that have been referred to a hospital. A referral from a general practitioner is required before a hospi- tal will admit a patient, except in an emergency. During the last 15 years most general practi- tioners in Denmark have joined group practices. The groups usually establish themselves in pri- vate health centers or clinics that contain shared laboratories and other diagnostic facilities so that most diagnostic and therapeutic problems can be handled without referring a patient to a hospital. All surgical cases except those involving very simple procedures are referred to hospitals. In 1977, 57 percent of physicians in Scot- land were hospital based, 37 percent were in general practice, and 6 percent were in public health or held administrative positions.#? As in Denmark, almost all specialists are employed by hospitals. General practitioners engage in private practice under contracts with the Scottish Health Service. In most cases general practitioners pro- vide only ambulatory care services, but there are some exceptions. General practitioners in rural areas can admit patients to cottage hospitals and look after them during their hospital stays. In urban areas some general practitioners work part-time in hospitals, but usually they serve as part of a specialized unit staff and do not care for their own patients.?! As a rule, specialists see only hospitalized patients, but again there are some exceptions. Specialists see ambulatory patients referred to hospital outpatient depart- ments, and some specialists maintain part-time private practices in which they treat ambulatory patients. A small number of physicians not affiliated with the health service provide ambula- tory care and follow their patients who are ad- mitted to a private hospital. In the majority of cases, though, patients receive ambulatory care and hospital care from different physicians.48 Recently general practitioners in Scotland have been joining group practices and moving into health centers.?! Primary care teams of physicians, nurses, and midwives have developed in these settings. The health centers are also equipped with more diagnostic and treatment facilities than solo practitioners have been able to support in the past, and the availability of the facilities may decrease the number of hospital referrals made by general practitioners. In 1977 approximately 47 percent of all physicians in West Germany were in private practice, 45 percent were employed by hospi- tals, and 8 percent worked in other institutions, such as public health departments, industries, and research organizations.#6 About 48 percent of full-time hospital physicians were specialists, and 47 percent of all physicians were specialists. A large number of the hospital staff who were not specialists were in training, frequently spe- cialist training. The number of specialists has in- creased in recent years, while the number of general practitioners has remained constant. Ambulatory and hospital care are usually provided by different physicians in West Ger- many. Very few hospitals have outpatient de- partments, and only 7 percent of all physicians, the “Belegirzte,” follow patients both in and out of the hospital.82 Most hospital care is sup- plied by physicians employed by the hospitals; most ambulatory care, by physicians in private practice. Physicians in private practice usually do not belong to group practices.8! However many share laboratory facilities and other equip- ment or maintain their own laboratories and equipment; thus it is unnecessary for them to refer patients to hospitals for many tests and minor procedures. Differences in the supply, specialization, and organization of physicians probably affect hos- pital utilization among the four countries. While physician services may substitute for hospital services, a larger supply of physicians also could lead to increased hospital use since more physi- cians can evaluate more patients and find more health disorders that need hospital treatment. The larger the proportion of specialists, hospital based or not, the more hospitalizations there may be, since the complex medical equipment and facilities located in hospitals are likely to be used more by specialists than by general practi- tioners. Large numbers of hospital-based physi- cians, whether or not they are specialists, may also increase the emphasis on hospital care. However the growth in group practices could de- crease hospital use since more extensive diag- nostic and treatment facilities can be supported by most group practices than by most solo prac- tices. Further study is needed of these possible relationships. The effects of nurses being involved in ambulatory and home care, and of the availa- bility of other home support services also require further study. When nurses play a significant role in the provision of ambulatory care services, they may substitute in part for physicians and thus increase access to the health services sys- tem. However, whether increased access leads to an increase or decrease in hospital use requires careful investigation. The availability of home nursing and other home support services proba- bly reduces the need for hospital and other in- patient care. These services should allow patients to complete their convalescence at home after hospital treatment for an acute illness, and help maintain the chronically ill and disabled outside institutions. However it is possible that users of home services have different characteristics than do users of inpatient facilities. Most nurses in Denmark, Scotland, West Germany, and the United States work in hospi- tals and other inpatient institutions. However some nurses in each country provide ambulatory or home care services. In the United States nurses have a long history of involvement in ambulatory and home care. Public health nurs- ing, school nursing, private duty nursing, and home visit nursing all existed before 1915.78 At that time private duty nursing, which provided home care as well as hospital care, outnumbered all other nursing categories. The proportion of private duty nurses declined over time, though, and the proportion of nurses employed in hospi- tals increased. In 1974, 75 percent of the registered nurses in the United States worked in hospitals and nursing homes. Another 4 percent were in nursing education, 7 percent worked in public health and schools, and the remainder were in occupational health, private duty, doc- tors’ offices, and other fields.92 Recently there has been renewed interest in the United States in nursing roles outside 43 hospitals. The role of the nurse practitioner has been created to help with the shortage of physi- cians in certain areas of the country. Nurse prac- titioners are trained to perform tasks previously performed by physicians. In 1979 there were an estimated 16,240 nurse practitioners in the United States, and the majority provided pri- mary care.’® There is also increasing interest in home nursing as part of comprehensive home care programs. The home care programs are seen as possible alternatives to costly long-term insti- tutional care.93 In 1975 approximately 74 percent of the nurses in Denmark were employed in hospitals, and the rest were in social welfare; that is, they worked as public health nurses or in nursing homes, old age homes, and the like.87 Public health nursing did not exist in the country until 1937, when a program was started to train nurses as infant home visitors who would provide regu- lar well-baby checkups. Beginning in 1946 infant home visitors also supplied school nursing serv- ices, and in the 1950’s experimental programs combined home nursing for adults and children with infant home visiting and school nursing.16 Today specially trained public health nurses sometimes provide all these services, sometimes only school or infant care. Some registered nurses without special training also supply home and school nursing care. However there is a shortage of nurses in Denmark. There are not enough registered nurses in general or enough specially trained nurses for the public health posts. Therefore the home nursing services are not readily available in all areas. In 1977, 92 percent of Scotland’s nursing personnel worked in hospitals, but almost all other nurses were in community services.40 Community nurses are primarily home nurses and health visitors. The health visitors are public health nurses with special training in social serv- ices, mental health problems, and interviewing skills. The role of the community nurse devel- oped in the nineteenth century when voluntary organizations began hiring nurses to visit people’s homes to help the sick and teach proper child care. Now employed by the health service, the community nurses continue to provide a signifi- cant amount of health care. In 1973 home nurses made over 3 million visits, 2 million of which were to patients age 65 years or over, and health visitors made an additional 1.9 million home visits, 18 percent of which were to the elderly.89 Almost all the nursing personnel in West Germany work in hospitals. There are very few public health nurses in the country who provide ambulatory care services. Physicians in public health departments and private practices supply services, such as well-baby examinations, that public health nurses supply in many other coun- tries. Some home nursing services do exist. Five major voluntary agencies have organized home nursing services, and a few public agencies, profitmaking agencies, and religious organiza- tions also have done 50.88 Home nursing services are primarily available in urban and industrial areas, but even there insufficient personnel exist to supply all the necessary services. Comprehen- sive social services centers are being developed to help meet the needs of the population, especially the aged. The centers are staffed with nurses and social workers and offer a wide range of support services to local areas. ; In addition to home nursing, other inhome support services are available in each country. These include homemaker or home-help services and meals-on-wheels programs. Home help con- sists of assistance with household management, such as cleaning, washing, shopping, as well as personal assistance with dressing, bathing, and the like. Meals-on-wheels programs supply hot meals to persons, usually the elderly, who are unable to shop for or prepare food and who have no one to help them on a regular basis. In Denmark local social welfare authorities are required by law to provide home help to needy elderly and disabled pensioners.8% Local social services authorities in Scotland have established numerous home-help and inhome meal services.89 In West Germany social services centers provide home-help and inhome meal services, as do a variety of other voluntary and government agencies. At present the services are usually available only in urban areas, but further expansion is planned.88 In the United States welfare and voluntary agencies provide home support services, and there is some Federal financing of services covered by Medicare and the amendments to the Social Security Act of 1975. The number of homemaker-home-health- aides is much smaller than needed, though, and the Medicare reimbursement policies are so com- plex and restrictive that they inhibit the growth of services.88,93 To summarize, each of the four countries exhibits different patterns in the availability and organization of ambulatory and home care serv- ices. The United States is unique in that ambula- tory and hospital care are generally provided by the same physician. The United States also has the highest percent of specialist physicians, the lowest percent of hospital-based physicians, and the highest proportion of nurses working outside inpatient care facilities. Denmark has the highest percent of hospital-based physicians, an espe- cially even distribution of general practitioners, and widely available home-help services. Scot- land reports the lowest number of physicians per 10,000 population, but it has well-established community nursing and home support programs. West Germany has the highest number of physi- cians per 10,000 population, the lowest percent of specialist physicians, and very few nurses working outside hospitals. COSTS OF RECEIVING HEALTH CARE In this section the patients’ cost of receiving hospital care is compared with the costs of re- ceiving other forms of health care. The focus is on whether persons in each of the countries have barriers to or incentives for hospital use. Since physicians usually decide when to admit and dis- charge patients, there is also a brief discussion of how physicians are paid and whether the method of payment could provide economic incentives for hospital use. In 1976, 89 percent of the U.S. civilian non- institutionalized population was covered by some type of health insurance, while 11 percent, or about 23 million people, had no insurance.93 Private hospital insurance, which covered 76 per- cent of the population, was the most common. Some persons who carried private insurance were also covered by government Medicare and Medicaid programs. An additional 4 percent of the population was covered only by Medicare, and 6 percent was covered exclusively by Medic- aid. Almost all those enrolled in Medicare are age 65 years or over, while Medicaid covers cer- tain low income individuals. Frequently persons must make copayments for insured health serv- ices, and some services are generally excluded from insurance coverage. In 1977 the proportion of hospital expenses paid directly by U.S. patients was quite small: 6 percent.9% Private insurance covered 37 per- cent of the expenditures for hospital care; Medicare, 24 percent; Medicaid, 9 percent; and other Federal, State, and local government pro- grams, 22 percent. The remaining 2 percent was accounted for by philanthropy and industry. Nevertheless, persons using hospital services could incur significant out-of-pocket expenses. A 1975 survey showed that the average annual out-of-pocket hospital expense to persons with an expense was $264. For 41 percent of persons with a hospital expense, the amount paid directly was less than $50, but for 12 percent it was over $500.95 U.S. physicians are usually paid on a fee-for- service basis. However hospital-based physicians receive salaries from the hospitals, and a small proportion of office-based physicians are in pre- paid group practices and are paid on a capitation basis. Capitation is the payment of a fixed amount each year per person without regard to the amount of medical care the person receives. In 1977, 39 percent of the total cost of physi- cian services was paid directly by patients; 37 percent was paid by private insurance; 24 per- cent was paid by government programs.9¢ The average annual out-of-pocket expense for physi- cian services, among persons with an expense, was $107 in 1975.95 Most U.S. nursing homes are private profit- making institutions. Approximately 41 percent of the total expense of nursing home care was paid directly by patients in 1977, and govern- ment programs covered 57 percent of the costs. Data on the cost of nursing home care to those with an expense is not available from the 1975 survey. Direct payments for drugs amounted to 83 percent of the total cost of drugs in 1977.94 Private health insurance cov- ered 8 percent of the total cost of drugs, and government programs covered 9 percent. In 1975 the average annual expense for prescrip- tion drugs was $59 for persons with an expense, and the average annual expense for optical serv- ices among persons with an expense was $67.95 In 1977 direct payments accounted for 92 per- cent of the total expense for eyeglasses and ap- pliances. Private insurance covered 2 percent of the total cost of eyeglasses and appliances that year, and government programs covered 6 percent.9% These statistics indicate that the out-of- pocket costs of receiving health care in the United States could be a barrier to hospital use for some; an incentive for others. The cost of hospital care is likely to be difficult to pay for persons without insurance or for insured persons who must make substantial copayments. How- ever, among those with comprehensive insurance coverage of hospital expenses, the higher propor- tion of the costs for other health services that must be paid out-of-pocket could encourage hospital use. In Denmark the entire population is required to take part in the health insurance system. Since 1973 the system has been administered by the county councils and has been financed by a graduated income tax. Two types of insurance exist. Under the first plan there are no direct charges for physician services, but persons must agree to see only one general practitioner during the year and to see specialists only if referred by the general practitioner. Under the second plan persons may see any general practitioner or spe- cialist they wish at any time, but they are charged directly by the physician and are reim- bursed only in part by health insurance. Until 1976 all persons with an income below a certain limit, some 80 percent of the population, had the first type of coverage, and those with higher incomes had the second type. Since 1976 there has been free choice between the plans for all.16 The expense of public hospital treatment in Denmark does not require insurance coverage since hospital care is provided without any direct charges to the patient. Hospitals receive their funds from the county governments and patients must attend a hospital in the county where they reside to receive free care. There were small fees for hospital use until 1973, but health insurance covered the fees for the most part.87 Care in convalescent homes is also free of charge. However, to reside in an old age home or another facility operated by the local social wel- fare authorities, an individual must give up his or her pension, except for a small amount for daily necessities. If an individual has income in addi- tion to the pension, most of that income must be used to pay for maintenance expenses in the welfare facility.85 Hospital physicians are salaried by the hospi- tals and do not charge inpatients for their serv- ices. General practitioners are reimbursed by the insurance system using both capitation and fee- for-service principles. About half of the income most general practitioners receive from the in- surance system is based on the number of persons on their lists, and half is based on the services the general practitioners have pro- vided.!6 Drugs are divided into three categories. The county government pays 75 percent of the cost of one category, the necessary drugs, and pays 50 percent of the cost of the second cate- gory, the less important drugs. The cost of other drugs must be paid directly by the patient.82 In sum, there are no cost barriers to receiving hospital care in Denmark, and there is little economic incentive for patients to choose hospi- tal care over ambulatory care. However free long-term care in hospitals might be preferred by patients to the loss of their pensions and the other costs of treatment in a social welfare facility. In Scotland health insurance exists, but is carried only by a small proportion of the popu- lation. Most health care is financed by the health service from general government revenues and is provided to patients without direct charge. Hos- pital care can be obtained without charge in health service hospitals, but some patients choose to pay a small fee for more private accommodations. Other patients prefer to be treated by a private physician and pay a special fee for the privilege.#8 Patients who use private hospitals or private beds in health service hospi- tals are charged for care, but the number of pri- vate hospitals and beds is quite small. The fees for special treatment in the health service hospi- tals and for private hospitalizations are usually covered, at least in part, by private insurance policies. Ambulatory care is provided without charge for the most part. Small charges are made for some medicines, dentures, eyeglasses, and other appliances, but physicians’ services are generally supplied without cost to the patient. General practitioners are paid by the health service on a capitation basis, with various modifications. There are special adjustments to encourage phy- sicians to practice in underserved areas and in- creased payments for treatment of the elderly, night calls, and other practices the health service wishes to encourage.?! Specialists are salaried by the health boards but are often allowed to engage in part-time private practice in which they charge patients directly. In addi- tion a small number of physicians in full-time private practice charge their patients directly for treatment. Private insurance funds cover the charges for private ambulatory care in many cases. Thus there are no significant cost barriers to receiving hospital care in Scotland, nor are there incentives for patients to choose hospital care rather than other kinds of health care. Most ambulatory care is free to the patient, and so is most long-term care, since it takes place in hospitals. In West Germany 96 percent of the popula- tion is covered by insurance from a local or specialized sickness insurance fund. Many per- sons covered by insurance from a sickness insur- ance fund also carry supplementary private insurance, and a small percent of the population only carries insurance from a private insurance company.48:81,82 The sickness insurance funds cover the costs of hospital care necessary for the treatment of an illness or injury. If the patient wishes to have more private hospital accommo- dations or to be treated in the hospital by the physician of his or her choice, there is a direct charge, but often it is covered by the supplemen- tary insurance policies. The sickness insurance funds cover the costs of long-term care in hospi- tals, including the Kurkrankenhduser. The costs of nursing home treatment are included in the standard coverage of the sickness insurance funds in some States, and the coverage can be obtained on an optional basis in several other States. The costs of ambulatory care for the treat- ment of an illness or injury are covered almost entirely by the sickness insurance funds. The physician from whom treatment is received must be registered with the funds, but almost all phy- sicians in private practice are registered. Patients who make visits to nonregistered physicians usually carry private or supplementary insurance to cover the costs of these visits. Patients must make small copayments for prescription drugs and appliances, and not all of the costs of pre- ventive health services are covered by insurance, but the sickness insurance funds are moving toward more complete coverage of preventive care. Hospital physicians are usually salaried, but if they provide ambulatory care outside the hospital, they receive fee-for-service reimburse- ment. Physicians in private practice are reim- bursed on a modified fee-for-service basis. The sickness insurance funds pay prearranged lump sums to local physician associations, which in turn pay the physicians’ shares of the lump sum based on the number and type of services the physicians have provided to insured persons.48 In West Germany, as in Denmark and Scot- land, there is little indication of cost barriers or incentives that influence patients who choose hospital care. Nursing home care could be more costly for West German patients than hospital or ambulatory care, but health services are gener- ally available without significant out-of-pocket costs. In addition to the probable effect of eco- nomic barriers and incentives on patient choices in the United States, U.S. physicians may have more economic incentive to hospitalize patients than do physicians in the other three countries. The use of fee-for-service reimbursement in the United States for most physicians’ services that are provided in hospitals makes it financially rewarding for physicians to admit patients and perform expensive treatments. Salaried hospital physicians have no such incentive. When capita- tion alone is used as a basis of payment for office-based physicians, and the office-based physicians do not follow patients in the hospi- tals, there may be an economic incentive for physicians to refer patients who require time- consuming care to hospitals. The office-based physician’s income is not increased by providing complex services, and it is not decreased by having the hospitals provide the services. Fee- for-service payment for office-based physicians who do not follow patients in hospitals could create economic incentives for the office-based physician to provide as many services as possible outside the hospital. In Denmark, Scotland, and West Germany, office-based physicians generally 47 do not follow hospitalized patients, but none of the countries uses either capitation alone or pure fee-for-service reimbursement. In spite of the modifications, there may be some incentive for office-based physicians in Scotland to refer pa- tients to hospitals and for office-based physi- cians in West Germany to avoid hospital referrals. SUMMARY This section has surveyed characteristics of the health service systems in Denmark, Scotland, West Germany, and the United States. Certain characteristics in each country are likely to in- crease hospital utilization. In the United States the characteristics include economic incentives to physicians and patients for hospital use and the high percent of physicians in specialty prac- tices. In Denmark the high proportion of SUMMARY AND The comparability of hospital utilization statistics in Denmark, Scotland, West Germany, and the United States was discussed in this report. The main focus was on general hospital discharge reporting systems, which collect ab- stracts of information about individual dis- charges. For Denmark, Scotland, and West Ger- many, other types of reporting systems were also discussed, including those that cover special hospitals or groups of patients (such as psychi- atric or maternity patients), annual hospital reports of aggregate utilization data, and house- hold surveys that collect information about hospital use. Although the United States has many impor- tant hospital data systems, only the National Hospital Discharge Survey was described here. Other U.S. sources of national hospital utiliza- tion statistics are briefly reviewed in appendix II. The National Hospital Discharge Survey was chosen for comparison with the general hospital discharge reporting systems in the other three countries because it is the major source of na- tional estimates of short-stay hospital use for the United States. hospital-based physicians may increase hospital use. In Scotland the use of capitation to pay office-based physicians and the absence of a nursing home system probably increase hospital use, and in West Germany the large number of physicians per population and the lack of suffi- cient alternative facilities for long-term care are likely to increase hospital use. On the other hand, each health services system has some characteristics that should de- crease hospital use: for instance, the large num- ber of alternative facilities for long-term care in the United States, the well-established programs for home care in Denmark and Scotland, and the small percent of physicians who are specialists in West Germany. Further research is needed to understand the interactions of these factors and the effects they have on hospital use. Increased understanding of the effects of health services system characteristics should result in more use- ful comparisons of hospital utilization statistics. CONCLUSIONS Several similarities were found in the com- parison of the four countries’ reporting systems; for instance, many of the same items of informa- tion are collected in each system, and each uses a variation of the International Classification of Diseases to code diagnoses. However differences also exist, particularly in the types of hospitals and patients covered in the reporting systems and in the definitions and procedures used by the systems to compute utilization statistics. The report also contained a brief review of the health services systems in the four countries. Information was given about characteristics of the health services systems that are likely to affect hospital utilization rates. The characteris- tics include the availability of separate long- term-care facilities, patterns in the provision of ambulatory and home care services, and the costs to patients of receiving various kinds of health care. Each country’s health services sys- tem was shown to have some characteristics that are expected to increase hospital use and others that are likely to decrease it. The information presented indicates that discharge reporting systems in other countries could be valuable sources of data for compara- tive studies. However the differences in the re- porting systems require further evaluation, and procedures need to be developed to adjust the data to take the differences into account. Work is underway in this area. As mentioned in the introductory section, a 10-country study is being done to estimate the effects of differences in the reporting systems on the data the systems pro- duce. The study covers the four countries dis- cussed in this report and the six countries whose discharge reporting systems are described in an earlier report.! Data from each country is ad- justed to take the differences in the reporting systems into account. Further research is also needed into the effects health services system characteristics have on hospital utilization rates. This type of research is valuable in its own right. Health policymakers are deeply concerned with con- trolling hospital use and associated high costs. Understanding the systemic factors that affect hospital utilization rates should assist policy- making decisions. When the broad effects of health services system characteristics are under- stood, it will also be possible to make more detailed comparisons of hospital utilization sta- tistics. Specific questions—such as whether dis- charge rates for respiratory disease are related to rates of automobile ownership, and what the relationship is between hospital staffing ratios and average length of stay—can be studied cross- nationally once controls are introduced for the system differences that affect discharge rates (such as percent of specialists) or differences that affect length of stay (such as the presence or absence of a nursing home system). The cross-national research already done that uses data from discharge reporting systems gives an indication of the promise of future re- search. R. F. Bridgman uses data from areas of eight countries, including Schleswig-Holstein in West Germany, to study the relationship be- tween the level of socioeconomic development and hospital utilization.96 In other studies hos- pital caseloads are compared cross-nationally,97 the relationship between hospital use and hospi- tal expenditures is examined,?® and the effect of the number of surgeons on surgical rates is explored.?? Further research using data that have been adjusted to take differences in the statistical systems into account, and that control for the broad effects of health services system characteristics, should help answer many addi- tional questions about hospital use that concern the United States and other countries of the world. 0 0 O ~——eerreee 49 REFERENCES 1 National Center for Health Statistics: The status of hospital discharge data in six countries, by L. J. Kozak, R. Andersen, and O. W. Anderson. Vital and Health Sta- tistics. Series 2-No. 80. DHEW Pub. No. (PHS) 80-1354. Public Health Service. Washington. U.S. Government brining Office, Mar. 1980. 2National Center for Health Statistics: Development of the design of the NCHS Hospital Discharge Survey, by W. R. Simmons and G. A. Schnack. Vital and Health Sta- tistics. PHS Pub. No. 1000-Series 2-No. 39. Public Health Service. Washington. U.S. Government Printing Office, Sept. 1970. 3 United Nations: Demographic Yearbook, 1977. New York. United Nations, 1978. 4World Health Organization: Health Services in Europe, 2d ed. Copenhagen. World Health Organization, 1975. 5 Registrar General Scotland: Annual Report, 1977, Part 1—Mortality Statistics. Edinburgh. Her Majesty’s Stationery Office, 1978. 6National Center for Health Statistics: Health, United States, 1980. DHHS Pub. No. (PHS) 81-1232. Public Health Service. Washington. U.S. Government Printing Office, Dec. 1980. "Information Services Division of the Common Serv- ices Agency, Scottish Health Service: ISD(S)1 scheme— year ending March 1978, Hospital Bed Resources. Edin- burgh. Scottish Health Service, 1979. Statistisches Bundesamt: Krankenhauser, 1977. Gesundheitswesen. Fachserie 12-Reihe 6. Stuttgart und Maigz. W. Kohlhammer GMBH, 1979. 9Sundhedsstyrelsen: Ydelsesstatistik for sygehu- svaesenet, 1977. Medicinalstatistiske Meddelelser. Pub. Noy 3/8: 7. Copenhagen. Sundhedsstyrelsen, 1978. ONational Center for Health Statistics: Unpublished a from the Division of Health Care Statistics. Information Services Division of the Common Services Agency, Scottish Health Service: ISD(S)1 scheme—year ending March 1978, Hospital Utilisation Statistics. Edinburgh. Scottish Health Service, 1979. 12Dreyer, K.: Present activities in health information systems in Denmark, in P. Armitage, ed., National Health Survey Systems in the European Foonomic Community. Luxembourg. Commission of the European Communi- ties, 1977. pp. 26-32. 13 Nordic Medico-statistical Committee (NOMESKO): Computer-Based Patient Statistics: Part One—Hospital In-Patients. Stockholm. NOMESKO, 1974. 50 14Dreyer, K., Deputy Director of Division, National Health Service of Denmark: Personal communication. 15Sundhedsstyrelsen: Medicinsk fodselsstatistik, 1977. Medicinalstatistiske Meddelelser. Pub. No. 1979: 1. Copenhagen. Sundhedsstyrelsen, 1979. 16 Grannik, D. et al.: The National Health System in Denmark. DHEW Pub. No. (NIH) 77-673. National In- stitutes of Health. Washington. U.S. Government Print- ing Office, 1976. 17S¢rensen, S.: Denmark, in G. A. Kool, ed., Hospital Statistics and a Minimum Basic Data Set. Report on a European Communities Seminar. Edinburgh. Nov. 2-4, 1976: pp. 13-15. 18 Sundhedsstyrelsen: Medicinalberetning II: Beretning om sygehuse og andre helbredelsesinstitutioner, Finansdret 1974-75. Copenhagen. Sundhedsstyrelsen, 1978. 19 Sundhedsstyrelsen: Sygdomsforlgb og sygehusforb- rug i tre amtskommuner. Medicinalstatistiske Meddelel- ser. Pub. No. 1978: 4. Copenhagen. Sundhedsstyrelsen, 1978. 20 Dupont, A. et al.: A cumulative national psychiatric register: Its structure and application. Acta Psychiatr. Scand. 50: 161-173, 1974. 21K astrup, M. et al.: Psychiatric treatment in a de- limited population—with particular reference to out- pafints. Acta Psychiatr. Scand. 53(1): 35-50, Jan. 1976. 22 Nielsen, J., and Nielsen, J. A.: Eighteen years of community psychiatric service in the island of Sams¢. Br. J. Psychiatry. 131: 41-48, July 1977. 23Weeke, A. et al.: Incidence of depressive syndromes in a Danish county. Acta Psychiatr. Scand. 51(1): 28-41, Jan. 1975. 24 Sundhedsstyrelsen: Beretning for psykiatriske insti- tutioner i Danmark, finans8ret 1975-76. Medicinal- statistiske Meddelelser. Pub. No. 1979: 3. Copenhagen. Sundhedsstyrelsen, 1979. 25Sundhedsstyrelsen: Befolkningens forbrug af psy- kiatriske sengepladser, en analyse i henhold til bopael pr. 1. April 1976, by K. Dreyer and A. Dupont. Medicinalstatistiske Meddelelser. Pub. No. 1978: 2. Copenhagen. Sundhedsstyrelsen, 1978. 26 National Health Service of Denmark: Report Sub- mitted by the Working Group on the Information System Appointed by the Priority Committee on the Heg)th Service. Unpublished document. 27 Committee on the Danish National Morbidity Sur- vey: The Sickness Survey of Denmark. Copenhagen. Committee on the Danish National Morbidity Survey, 1960. 28 Committee on the Danish National Morbidity Sur- vey: The Hospital Survey of Denmark. Copenhagen. Committee on the Danish National Morbidity Survey, 1959. 29Scottish Home and Health Department: Scottish hospital morbidity data, 1961-68, by E. Lockwood. Scottish Health Service Studies. Pub. No. 20. Edinburgh. Soottish Home and Health Department, 1971. OHeasman, M. A.: Scottish Hospital In-Patient Sta- tistics. Med. Care. 8(4) Supp.: 113-120, July-Aug. 1970. 31Heasman, M. A.: Scottish Hospital In-Patient Sta- * tistics—sources and uses. Health Bull. 26(4), Oct. 1968. 2parkin, D. M. et al.: Scottish Consultant Review of In-Patient Statistics (SCRIPS). A survey of users’ opinions. Health Bull. 34(5): 273-278, Sept. 1976. 33Heasman, M. A.: SCRIPS: Success or failure? A critical study of the Scottish Consultant Review of In-Patient Statistics, in G. McLachlan, ed., 4 Question of Quality: Roads to Assurance in Medical Care. Lon- don. Oxford University Press for the Nuffield Provincial Hospitals Trust, 1976. pp. 171-185. 34 Clarke, J. A.: Scotland, in G. A. Kool, ed., Hospital Statistics and a Minimum Basic Data Set. Report on a European Communities Seminar. Edinburgh. Nov. 2-4, 1976. pp- 43-48. 35 Hutcheon, A. W., and Wilson, G. M.: An attempt to improve the standard of diagnostic entries and records in a medical unit. Health Bull. 34(4): 221-227, July 1976. 36 Heasman, M. A., Director, Information Services Division of the Common Services Agency, Scottish Health Service: Personal communication. "Information Services Division of the Common Serv- ices Agency, Scottish Health Service: Scottish Hospital In-Patient Statistics, 1976. Edinburgh. Scottish Health Service, 1978. 8 Last, J. M.: Data collection requirements for patient management. Med. Care. 8(4) Supp.: 159-168, July-Aug. 1970. 39World Health Organization: Manual of the Interna- tional Statistical Classification of Diseases, Injuries, and Causes of Death, Based on the Recommendations of the Ninth Revision Conference, 1975. Geneva. World Health Organization, 1978. OInformation Services Division of the Common Services Agency, Scottish Health Service: Scottish Health Statistics, 1977. Edinburgh. Her Majesty’s Sta- tionery Office, 1978. Wilson, J. M. G., Senior Research Fellow, Informa- tion Services Division of the Common Services Agency, Scottish Health Service: Personal communication. 42 Ferguson, J. B. P., and Murray, A.: Survey of the use of form S.M.R.M. (Scottish Medical Record— Maternity) as a normal delivery discharge letter. Health Bull. 28(3): 30-35, July 1970. Information Services Division of the Common Services Agency, Scottish Health Service: Scottish Men- tal Health In-Patient Statistics, 1976. Edinburgh. Scot- tish Health Service, 1978. 44 Farmer, P. J., Information Services Division of the Common Services Agency, Scottish Health Service: Personal communication. 45 Office of Population Censuses and Surveys: General Household Survey, 1977. London. Her Majesty’s Sta- Hongry Office, 1979. 46genftleben, H. U., Office of Health, District Com- mittee of Main-Kinzig-Kreis: Personal communication. 7Schach, E., Research Associate, Computer Center, University of Dortmund: Personal communication. 48Simanis, J. G.: National Health Systems in Eight Countries. DHEW Pub. No. (SSA) 75-11924. Social Security Administration. Washington. U.S. Government Peijying Office, 1975. 9Pflanz, M., Former Professor and Head, Institute of Epidemiology and Social Medicine, Hanover Medical School: Personal communication. ~ 50Der Bundesminister fiir Arbeit und Sozialordnung: Ubersicht iiber die Soziale Sicherung, 1974. Bonn. Der ‘Bundesminister fur Arbeit und Sozialordnung, 1975. 51 Bundesverband der Ortskrankenkassen: Statistik der Ortskrankenkassen: Krankheitsarten-, Krankheitsura- schen-, und Sterblichkeits-Statistik, 1977. Bonn. Bun- degverband der Ortskrankenkassen, 1979. 52Bundesministerium fiir Jugend, Familie, und Gesundheit: Das Gesundheitswesen der Bundesrepublik Deutschland. Band 5. Ausgabe 1974. Stuttgart und Majpe. W. Kohlhammer GMBH, 1974. 53Huschenbeth, L: Health statistics in the Federal Republic of Germany: Present activities, in P. Armitage, ed., National Health Survey Systems in the European Economic Community. Luxembourg. Commission of the Lh Communities, 1977. pp. 42-47. 54 Heinemann, A., Director of the Hospital Diagnostic Statistics Reporting System, Statistical Office of Schleswig-Holstein: Personal communication. 55Griesser, G.: General statistics and hospital informa- tion, in G. A. Kool, ed., Hospital Statistics and a Mini- mum Basic Data Set. Report on a European Communi- ties Seminar. Edinburgh. Nov. 2-4, 1976. pp. 63-79. 56Statistisches Landesamt Schleswig-Holstein: Die Krankheiten der Krankenhauspatienten in Schleswig- Holstein, 1976. Kiel. Statistisches Landesamt, Schleswig- Holstein, 1978. 57Kukla, R., Rhineland Provincial Union: Personal communication (obtained and translated by H. U. Senfileben), 58 Landschaftsverband Rheinland: Zahlen, Fakten und Tendenzen, 1978. Koln. Landschaftsverband Rheinland, Jugs 1979. 59 Thomas, Mr., Institute of Defense Forces’ Medical Statistics: Personal communication (obtained and trans- lated by H. U. Senftleben). ONaegele, Mrs., Federal Statistical Office: Personal coltaunication (obtained and translated by E. Schach). 61Statistisches Bundesamt: Wirtschaft und Statistisk. . Heft 2,1977. Der Bundesminister fur Jugend, Familie, und Ge- sundheit: Daten des Gesundheitswesens, 1977. Bonn. Der Bundesminister fur Jugend, Familie, und Gesundheit. 51 63 National Center for Health Statistics: Reporting of hospitalization in the Health Interview Survey. Vital and Health Statistics. PHS Pub. No. 1000-Series 2-No. 6. Public Health Service. Washington. U.S. Government Printing Office, July, 1965. 64 National Center for Health Statistics: Development and maintenance of a national inventory of hospitals and institutions. Vital and Health Statistics. PHS Pub. No. 1000-Series 1-No. 3. Public Health Service. Washington. U.S. Government Printing Office, Feb. 1965. 65 National Center for Health Statistics: Participation of hospitals in the pilot study of the Hospital Discharge Survey, by A. M. Brown, I. Altman, and D. J. Thomp- son. Vital and Health Statistics. PHS Pub. No. 1000- Series 2-No. 19. Public Health Service. Washington. U.S. Government Printing Office, Oct. 1966. 66 National Center for Health Statistics: Utilization of short-stay hospitals, annual summary for the United States, 1978, by B. J. Haupt. Vital and Health Statistics. Series 13-No. 46. DHEW Pub. No. (PHS) 80-1797. Public Health Service. Washington. U.S. Government Printing Office, Mar. 1980. 67 National Center for Health Statistics: Quality con- trol in the Hospital Discharge Survey, by K. W. Harris and K. L. Hoffman. Vital and Health Statistics. Series 2- No. 68. DHEW Pub. No. (HRA) 76-1342. Health Re- sources Administration. Washington. U.S. Government Printing Office, Dec. 1975. 68 Commission on Professional and Hospital Activities: International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Ann Arbor. Edwards Brothers, Inc., 1978. 69 American Association of Medical Record Librarians: Glossary of Hospital Terms. Chicago. American Associa- tion of Medical Record Librarians, 1969. 70 National Center for Health Statistics: Utilization of short-stay hospitals, summary of nonmedical statistics, United States, 1973, by W. F. Lewis. Vital and Health Statistics. Series 13-No. 23. DHEW Pub. No. (HRA) 76- 1774. Health Resources Administration. Washington. 2 5 Government Printing Office, July 1976. "I National Center for Health Statistics: Inpatient utilization of short-stay hospitals by diagnosis, United States, 1975, by L. Glickman. Vital and Health Statis- tics. Series 13-No. 35. DHEW Pub. No. (PHS) 78-1786. Public Health Service. Washington. U.S. Government Printing Office, Apr. 1978. 72National Center for Health Statistics: Surgical operations in short-stay hospitals, United States, 1975, by A. L. Ranofsky. Vital and Health Statistics. Series 13- No. 34. DHEW Pub. No. (PHS) 78-1785. Public Health Service. Washington. U.S. Government Printing Office, Apr. 1978. 73 National Center for Health Statistics: Utilization of short-stay hospitals by persons discharged with alcohol- related diagnoses, United States, 1976, by M. Sanchez. Vital and Health Statistics. Series 13-No. 47. DHHS Pub. No. (PHS) 80-1798. Public Health Service. Washington. U.S. Government Printing Office, May 1980. 52 74 National Center for Health Statistics: Eighth Revi- sion International Classification of Diseases, Adapted for Use in the United States. PHS Pub. No. 1693. Public Health Service. Washington. U.S. Government Printing Office, 1967. 75 National Center for Health Statistics: Detailed Diag- noses and Surgical Procedures for Patients Discharged From Short-Stay Hospitals: United States, 1977, by B. J. Haupt. DHEW Pub. No. (PHS) 79-1274. Public Health Service. Washington. U.S. Government Printing Office, Sept. 1979. 76 National Center for Health Statistics: Health, United States, 1979. DHEW Pub. No. (PHS) 80-1232. Public Health Service. Washington. U.S. Government Prjing Office, 1980. 7TNational Center for Health Statistics: The Nation’s Use of Health Resources, 1979, by B. J. Haupt. DHEW Pub. No. (PHS) 80-1240. Public Health Service. Washing- ol U.S. Government Printing Office, 1980. 78 Anderson, O. A.: The Uneasy Equilibrium. New Haven, College and University Press, 1968. 79Sidel, V. W., and Sidel, R.: 4 Healthy State. New York. Pantheon Books, 1977. 80 Levitt, R.: The Reorganized National Health Serv- ice, 2d ed. London. Croom Helm, 1977. 81 Maynard, A.: Health Care in the European Commu- nity. London. Croom Helm, 1975. 82Hu, T., ed.: International Health Costs and Expendi- tures. DHEW Pub. No. (NIH) 76-1067. National Insti- tutes of Health. Washington. U.S. Government Printing Office, 1976. 83 Blanpain, J. et al.: National Health Insurance and Health Resources: The European Experience. Cambridge. Harvard University Press, 1978. 84 eichter, H. M.: 4 Comparative Approach to Policy Analysis: Health Care Policy in Four Nations. Cam- bridge. Cambridge University Press, 1979. 85 Furstnow-S¢rensen, B.: Care of the Old, 3d ed. Copenhagen. Ministries of Labor and Social Affairs, 1970. 86 National Center for Health Statistics: Inpatient health facilities as reported from the 1976 MFI survey, by J. F. Sutton and A. Sirrocco. Vital and Health Sta- tistics. Series 14-No. 23. DHEW Pub. No. (PHS) 80- 1818. Public Health Service. Washington. U.S. Govern- mens Printing Office, Jan. 1980. 87Koch, J. H. et al.: Health Services in Denmark. Copenhagen. Amtsrdsforeningen i Danmark, 1976. 88Kahn, A. J., and Kamerman, S. B.: Social Services in International Perspective. DHEW Pub. No. (SRS) 76- 05704. Social and Rehabilitation Service. Washington. u. 5 Government Printing Office, 1977. 89Kane, R. L., and Kane, R. A.: Long-Term Care in Six Countries. DHEW Pub. No. (NIH) 76-1207. National Institutes of Health. Washington. U.S. Government Pinsug Office, 1976. 90World Health Organization: World Health Statistics Annual, 1978. Vol. III, Health Personnel and Hospital Establishments. Geneva. World Health Organization, 1979. 91Brotherston, J.: Primary care financing: A Scottish perspective, in C. D. Burrell and C. G. Sheps, eds., Pri- mary Health Care in Industrialized Nations. Annals of the New York Academy of Sciences, vol. 310. New York. New York Academy of Sciences, 1978. pp. 252- 256. 92National Center for Health Statistics: Health Re- sources Statistics, 1976-77, by C. M. Croner. DHEW Pub. No. (PHS) 79-1509. Public Health Service. Wash- ington. U.S. Government Printing Office, 1979. 93 National Center for Health Statistics: Health, United States, 1978. DHEW Pub. No. (PHS) 78-1232. Public Health Service. Washington. U.S. Government Printing Office, Dec. 1978. 94 Gibson, R. M,, and Fisher, C. R.: National health expenditures, fiscal year 1977. Soc. Secur. Bull. 41(7): 3-20, July 1978. 95 National Center for Health Statistics: Personal out- of-pocket health expenses, United States, 1975, by C. S. Wilder. Vital and Health Statistics. Series 10-No. 122. DHEW Pub. No. (PHS) 79-1550. Public Health Service. Washington. U.S. Government Printing Office, Nov. 1978. 96 Bridgman, R. F.: Hospital Utilization: An Interna- tional Study. Oxford. Oxford University Press on behalf of the Regional Office for Europe of the World Health Organization, 1979. TPearson, R. J. C. et al.: Hospital caseloads in Liver- pool, New England, and Uppsala: An international com- parison. Lancet. 2: 559-566, Sept. 1968. 98 Andersen, R., and Hull, J. T.: Hospital utilization and cost trends in Canada and the United States. Med. Care. 7(6): 4-22, Nov.-Dec. 1969. Bunker, J. P.: Surgical manpower: A Comparison of operations and surgeons in the United States and in Eng- land and Wales. N. Engl. |. Med. 282(3): 135-144, Jan. 1970. 100 Commission on Professional and Hospital Activities: Length of Stay in PAS Hospitals, by Diagnosis, United States, 1976. Ann Arbor. Commission on Professional and Hospital Activities, Oct. 1977. 101 Health Care Financing Administration: Medicare: Inpatient use of short-stay hospitals, 1977, by C. Hel- bing. Health Care Financing Notes. DHEW Pub. No. (HFCA) 03022. Health Care Financing Administration. Baltimore. Health Care Financing Administration. 102Health Care Financing Administration: HFCA pro- gram statistics. Health Care Financing Review. 2(1): 79- 85, Summer 1980. 103 National Institute of Mental Health: Private psychi- atric hospitals, 1974-75, by M. J. Witkin. Report Series on Mental Health Statistics. Series A-No. 18. DHEW Pub. No. (ADM) 77-380. Alcohol, Drug Abuse, and Men- tal Health Administration. Washington. U.S. Govern- ment Printing Office, 1977. 104 National Institute of Mental Health: Psychiatric services and the changing institutional scene, 1950-1985, by M. Kramer. Report Series on Mental Health Statis- tics. Series B-No. 12. DHEW Pub. No. (ADM) 77-433. Alcohol, Drug Abuse, and Mental Health Administra- tion. Washington. U.S. Government Printing Office, 1977. 105 National Institute of Mental Health: The design of management information systems for mental health organizations: A primer, by R. L. Chapman. Report Series on Mental Health Statistics. Series C-No. 13. DHEW Pub. No. (ADM) 76-333. Alcohol, Drug Abuse and Mental Health Administration. Washington. U.S. Government Printing Office, 1976. 106 National Institute of Mental Health: Deinstitu- tionalization: An analytical review and sociological perspective, by L. L. Bachrach. Report Series on Mental Health Statistics. Series D-No. 4. DHEW Pub. No. (ADM) 76-351. Alcohol, Drug Abuse, and Mental Health Ad- ministration. Washington. U.S. Government Printing Office, 1976. 7 American Hospital Association: Guide to the Health Care Field, 1979 ed. Chicago. American Hospital Associ- ation, 1979. 108 American Hospital Association: Hospital Statistics, 1979 ed. Chicago. American Hospital Association, 1979. 109 National Center for Health Statistics: Current esti- mates from the Health Interview Survey: United States, 1978, by J. D. Givens. Vital and Health Statistics. Series 10-No. 130. DHEW Pub. No. (PHS) 80-1551. Public Health Service. Washington. U.S. Government Printing Office, Nov. 1979. 110 Andersen, R. et al.: Two Decades of Health Serv- ices: Social Survey Trends in Use and Expenditures. Cambridge. Ballinger, 1976. 111 Aday, L. A. et al.: Health Care in the U.S.: Equita- ble for Whom? Beverly Hills. Sage Publications, 1980. 000 53 54 LL IL APPENDIXES CONTENTS Contributors to Study Sources of National Hospital Utilization Statistics in the United States in Addition to the Na- tional Hospital Discharge Survey General Hospital Discharge Reporting Systems. Special Hospital Reporting Systems Aggregate Hospital Reports Household Surveys 55 56 56 56 57 87 APPENDIX | CONTRIBUTORS TO STUDY DENMARK Karen Dreyer Deputy Director of Division National Health Service of Denmark Dr. Finn Kamper-J¢rgensen Danish National Institute of Social Research Dr. Jens Schmidt Institute of Psychiatric Demography hus Psychiatric Hospital SCOTLAND P. J. Farmer Information Services Division Common Services Agency Scottish Health Service Dr. M. A. Heasman Director Information Services Division Common Services Agency Scottish Health Service Dr. J. M. G. Wilson Information Services Division Common Services Agency Scottish Health Service WEST GERMANY A. Heinemann Director of Hospital Diagnostic Statistics Reporting System State Office of Statistics Schleswig-Holstein R. Kukla Rhineland Provincial Union Mrs. Naegele Federal Statistical Office Dr. Manfred Pflanz Institute of Epidemiology and Social Medicine Hanover Medical School ; Elisabeth Schach Computer Center University of Dortmund Dr. H. U. Senftleben Office of Health District Committee of Main-Kinzig-Kreis Mr. Thomas Institute of the Defense Forces’ Medical Statistics Mr. Wortmann Federation of Local Sickness Funds UNITED STATES Mary Moien Division of Health Care Statistics National Center for Health Statistics Iris Shimizu Survey Design and Methods Staff National Center for Health Statistics Al Sirrocco Division of Health Care Statistics National Center for Health Statistics 55 APPENDIX II SOURCES OF NATIONAL HOSPITAL UTILIZATION STATISTICS IN THE UNITED STATES IN ADDITION TO THE NATIONAL HOSPITAL DISCHARGE SURVEY In addition to the National Hospital Dis- charge Survey, many local, State, and national statistical systems in the United States collect hospital statistics. The following are brief de- scriptions of some of the major sources of national hospital statistics. GENERAL HOSPITAL DISCHARGE REPORTING SYSTEMS Abstracting Services Approximately two-thirds of the hospitals in the United States subscribe to an abstracting service. The hospitals supply data about each of their discharges to the services, which, for a fee, process the data on computers and provide the - hospitals with various tabulations. The largest abstracting service is the Professional Activity Study (PAS), begun in 1953 by the Commission on Professional and Hospital Activities (CPHA). It covers 1,460 hospitals, or about 31 percent of all the discharges from nongovernmental short- stay hospitals in the United States. Information abstracted for PAS includes patient characteris- tics, length of stay, diagnoses, tests, treatments, and physician. In addition to the tabulations provided to the subscribing hospitals, PAS data are published in the CPHA series, Length of Stay in PAS Hospitals,100 and data tapes are available to researchers on a contractual basis. Medicare Claims File Health insurance is provided by a variety of public and private agencies and organizations in NOTE: A list of references follows the text. 56 the United States, many of which compile data from hospitalization claims. Medicare is a na- tional public health insurance program that covers most people age 65 years and over, those with chronic renal disease, and those who meet the disability provisions of the Social Security Act. The Medicare Claims File is maintained by the Health Care Financing Administration (HCFA), Department of Health and Human Services (DHHS). Information from 20 million Medicare claims has been collected since the file was begun in 1966. Twenty-percent systematic samples of claims have also been drawn and coded for medical diagnosis. HCFA publishes some hospital statistics from the file in its peri- odicals Health Care Financing Notes10l and Health Care Financing Review.192 Hospital utili- zation statistics by diagnosis are available on tape. SPECIAL HOSPITAL REPORTING SYSTEMS Mental Health National Reporting Programs The National Institute of Mental Health (NIMH), DHHS, conducts several major surveys: the Inventory of Mental Health Facilities, the Annual Census of Patient Characteristics for State and county mental hospitals, and sample surveys of selected mental health facilities. Eight types of facilities are covered by the NIMH sur- veys: psychiatric hospitals, psychiatric services in general hospitals, community mental health centers, other multicomponent mental health facilities, residential treatment centers for emotionally disturbed children, outpatient men- tal health facilities, mental health day-night facili- ties, and transitional mental health facilities. For the inventory data are collected on services pro- vided, caseload, staffing patterns, and expendi- tures of the individual mental health facilities. State and county mental hospitals and commu- nity mental health centers are surveyed annually for the inventory; other mental health facilities are surveyed biennially. For the annual census data are collected on the age, sex, and diagnostic characteristics of resident patients and admis- sions in State and county mental hospitals. In the sample surveys detailed information is col- lected on patient characteristics in certain types of mental health facilities; the type of facilities covered varies. NIMH publishes data in its Re- port Series on Mental Health Statistics103-106 and has data tapes available. Federal Hospital Reporting Programs Several Federal agencies operate statistical systems that receive data about hospitalizations in particular types of Federal hospitals. The Bureau of Medical Services, DHHS, operates an Inpatient Data System that receives data about hospitalizations in the eight general hospitals run by the Public Health Service. The Indian Health Service (IHS), DHHS, has an Inpatient Data Sys- tem that contains information on hospital serv- ices provided to American Indians and Alaskan natives in the 50 IHS hospitals and other hospi- tals that have contracts with the IHS. The Vet- erans’ Administration (VA) maintains the Pa- tient Treatment File, which contains a medical record abstract for each discharge from the 172 VA hospitals and discharges whose hospital care was provided under the auspices of the VA. In the Department of Defense the various Armed Forces operate hospitals for their personnel in the United States and abroad, and they collect statistics on the use of these hospitals. AGGREGATE HOSPITAL REPORTS Annual Survey of Hospitals Beginning in the 1940’s the American Hos- pital Association (AHA) has conducted an Annual Survey of Hospitals. The survey covers almost all hospitals in the United States and the U.S. territories. Each hospital is mailed a ques- tionnaire requesting information about its serv- ices, utilization, personnel, and finances. Over 90 percent of the hospitals usually respond to the questionnaire, and estimates are made of data for nonreporting hospitals. Statistics from the survey are published annually in the AHA publications Guide to the Health Care Field07 and Hospital Statistics! and are available on tape. HOUSEHOLD SURVEYS National Health Interview Survey The National Center for Health Statistics has been operating the National Health Interview Survey since 1957. The survey covers the civilian noninstitutionalized population of the United States. A multistage probability design is used to draw samples of households, and all adult mem- bers of the households are interviewed. The in- terviews are conducted each week throughout the year. Approximately 120,000 persons in 40,000 households are interviewed each year. Information is gathered about the demographic characteristics of the household members, their health status, and their use of health services. Data from the survey are published in series 10 of Vital and Health Statistics,}%9 and are avail- able on tape. National Surveys of Health Services The Center for Health Administration Studies at the University of Chicago conducted national household surveys of health care utilization and expenditures in 1953, 1958, 1963, 1970, and 1976. The surveys covered the noninstitu- tionalized population of the United States. Self- weighting probability samples of households were drawn for each survey, and the last three surveys also oversampled certain persons and families of special concern in health policy for- mation. The 1976 survey oversampled rural southern black persons, persons of Spanish herit- age living in the Southwest, and persons experi- encing episodes of illness. For the 1976 survey 7,787 persons were interviewed in 5,432 57 households. Information was collected about so- Two Decades of Health Services: Social Survey cial and demographic characteristics, health sta- Trends in Use and Expenditures!10 and Health tus, and access to the health care system. Major ~~ Care in the U.S.: Equitable for Whom ?111 Data sources of the results of the 1976 survey are: tapes from each survey are also available. 0 OO rovemmmpmonsn 58 VITAL AND HEALTH STATISTICS Series Series 1. Programs and Collection Procedures.—Reports which describe the general programs of the National Center for Health Statistics and its offices and divisions and data collection methods used and include definitions and other material necessary for understanding the data. Series 2. Data Evaluation and Methods Research.—Studies of new statistical methodology including experi- mental tests of new survey methods, studies of vital statistics collection methods, new analytical techniques, objective evaluations of reliability of collected data, and contributions to statistical theory. Series 3. Analytical Studies.—Reports presenting analytical or interpretive studies based on vital and health statistics, carrying the analysis further than the expository types of reports in the other series. Series 4. Documents and Committee Reports.—Final reports of major committees concerned with vital and health statistics and documents such as recommended model vital registration laws and revised birth and death certificates. Series 10. Data From the Health Interview Survey.—Statistics on illness, accidental injuries, disability, use of hospital, medical, dental, and other services, and other health-related topics, all based on data collected in a continuing national household interview survey. Series 11. Data From the Health Examination Survey and the Health and Nutrition Examination Survey.—Data from direct examination, testing, and measurement of national samples of the civilian noninstitu- tionalized population provide the basis for two types of reports: (I) estimates of the medically defined prevalence of specific diseases in the United States and the distributions of the population with respect to physical, physiological, and psychological characteristics and (2) analysis of relationships among the various measurements without reference to an explicit finite universe of persons. Series 12. Data From the Institutionalized Population Surveys. —Discontinued effective 1975. Future reports from these surveys will be in Series 13. Series 13. Data on Health Resources Utilization. —Statistics on the utilization of health manpower and facilities providing long-term care, ambulatory care, hospital care, and family planning services. Series 14. Data on Health Resources: Manpower and Facilities. —Statistics on the numbers, geographic distri- bution, and characteristics of health resources including physicians, dentists, nurses, other health occupations, hospitals, nursing homes, and outpatient facilities. Series 20. Data on Mortality. —Various statistics on mortality other than as included in regular annual or monthly reports. Special analyses by cause of death, age, and other demographic variables; geographic and time series analyses; and statistics on characteristics of deaths not available from the vital records based on sample surveys of those records. Series 21. Data on Natality, Marriage, and Divorce.—Various statistics on natality, marriage, and divorce other than as included in regular annual or monthly reports. Special analyses by demographic variables; geographic and time series analyses; studies of fertility; and statistics on characteristics of births not available from the vital records based on sample surveys of those records. Series 22. Data From the National Mortality and Natality Surveys.—Discontinued effective 1975. Future reports from these sample surveys based on vital records will be included in Series 20 and 21, respectively. Series 23. Data From the National Survey of Family Growth.—Statistics on fertility, family formation and dis- solution, family planning, and related maternal and infant health topics derived from a biennial survey of a nationwide probability sample of ever-married women 15-44 years of age. For a list of titles of reports published in these series, write to: Scientific and Technical Information Branch National Center for Health Statistics Public Health Service Hyattsville, Md. 20782 DHHS Publication Number (PHS) 81-1362 Series 2-No. 88 NCHS U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES PESTACE AND FEES PAID Public Health Service U.S. DEPARTMENT OF H.H.S. Office of Health Research, Statistics, and Technology HHS 396 National Center for Health Statistics x : : 3700 East-West Highway THIRD CLASS Hyattsville, Maryland 20782 BULK RATE OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 IS U.S.MAIL ————— For publications in the Vital and Health Statistics Series call 301- 436-NCHS. ~~ 40 Je re-X Sree Po nom oo nx ol 4 [TLC ies LO = © | aE FE ar EP, Dg SR a0. o [I 2 =] 5 ERI : A [va BC : Sl HEL £ 4180 3 | fy 12 J — 2 ln ie WAY 4 SZ oi E88 LE aE D 40 LN3INL1HVd3 @]\AVA JB vs JIIAH3S NVIN | SESE IC ERS a lua) |euolleN e 3JIA 2 RII CIS ETRE Saale SNLYs 68 ON ¢ S31i3S e SO1S| | | ] a COPYRIGHT INFORMATION The National Center for Health Statistics has obtained permission from the copyright holders to reproduce certain quoted material in this report. Further reproduction of this material is prohibited without specific permission of the copyright holders. All other material contained in the report is in the public domain and may be used and reprinted without special permission; citation as to source, however, is appreciated. SUGGESTED CITATION National Center for Health Statistics, A. Gittelsohn and P. Royston: Annotated bibliography of cause-of-death validation studies, 1958-80. Vital and Health Statistics, Series 2, No. 89. DHHS Pub. No. (PHS) 82-1363. Public Health Service, Washington. U.S. Government Printing Office, September, 1982. Library of Congress Cataloging in Publication Data Gittelsohn, Alan M. Annotated bibiliography of cause-of death validation studies, 1958-1979. (Vital and health statistics, Series 2, Data evaluation and methods research; no. 89) (DHHS publication; no (PHS) 81-1368) Authors: Alan Gittelsohn, Patricia N. Royston. 1. Death—Causes—Statistics—Bibliography. 2. Death—Causes— Classification—Bibliography. 3. Death—Proof and certification— Bibliography. 5. Errors, Theory of—Bibliography. |. Royston, Patricia N. Il. National Center for Health Statistics (U.S.) Ill. Series. IV Series: DHHS publication; no. (PHS) 81-1368. [DNLM: 1. Death certificates—Bibliograph. 2. Mortality—United States—Bibliography. 3. Diagnostic errors—Bibliography. W2 A N148vk no. 89) RA409.U45 no. 89 [25725] 312.0723s 81-16855 (RA405.A1] ISBN 0-8406-0236-7 For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 VIALE HEALTH oIATHollLo Annotated Bibliography of Cause-of-Death Validation Studies: 1958-1980 An annotated bibliography of studies relating to the errors in diagnosing, recording, coding, and tabulating cause of death. Data Evaluation and Methods Research Series 2, No. 89 DHHS Publication No. (PHS) 82-1363 U.S. Department of Health and Human Services Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics Hyattsville, Md September, 1982 National Center for Health Statistics ROBERT A. ISRAEL, Acting Director JACOB J. FELDMAN, Ph.D., Associate Director for Analysis and Epidemiology GAIL F. FISHER, Ph.D., Associate Director for the Cooperative Health Statistics System GARRIE J. LOSEE, Associate Director for Data Processing and Services ALVAN O. ZARATE, Ph.D., Assistant Director for International Statistics E. EARL BRYANT, Associate Director for Interview and Examination Statistics ROBERT L. QUAVE, Associate Director for Management MONROE G. SIRKEN, Ph.D., Associate Director for Research and Methodology PETER L. HURLEY, Associate Director for Vital and Health Care Statistics ALICE HAYWOOD, Information Officer Office of Research and Methodology MONROE G. SIRKEN, Ph.D., Associate Director JAMES T. MASSEY, Ph.D., Chief, Survey Design Staff ROBERT J. CASADY, Ph.D., Chief, Statistical Methods Staff KENNETH W. HARRIS, Special Assistant for Program Coordination and Statistical Standards Foreword Compilation of national cause-of-death statistics is a basic program responsibility of the National Center for Health Statistics. Consequently, the Center is very interested in the quality of the medical certifi- cations reported on death certificates since it is of fundamental importance in interpreting cause-of- death statistics. The more that is known about the quality of the medical certifications, the better the Center and its data consumers will be able to use national cause-of-death statistics. This bibliography presents an annotated listing of published and unpublished papers on the quality of cause-of-death statistics. The compilation is based on a search of the relevant literature, as well as a survey of producers and users of mortality statistics. It contains 128 references to a wide assortment of studies that were conducted in this country and abroad during the preceding 24 years. Since the deaths covered by the referenced studies vary considerably by cause, place, and year of death, as well as by the methods used to evaluate the certifi- cations, this publication contains several indexes to assist the readers in locating appropriate references. These indexes are described in the Introduction. The bibliography was compiled intially to assist the Center in assessing the state of knowledge with respect to the quality of medical certification, and the state of the art with respect to methods that have been used to evaluate the quality of cause-of-death statistics. It is being published to provide health researchers with a comprehensive reference guide to the literature on this subject. Definitive conclusions about the quality of medical certification on the death certificates have not been reached based on the studies listed in this bibliography. Evidently the quality varies greatly by cause of death and characteristics of the decedent. However, the findings provide no basis for compla- cency. Some studies detected large discrepancies between the diagnoses certified on death certificates and those reported by autopsy and hospital records. The most striking finding is that so little is known about the quality of medical certification and its effect on diagnostic statistics in general and on national cause-of-death statistics in particular. Apparently no country has a well-defined program for systematically assessing the quality of medical certifications being reported on death certifi- cates or for measuring the error effects on the levels and trends of cause-of-death statistics. There are, for example, only three references in this bibliography to U.S. studies that were national in scope. These studies were based on deaths due to lung cancer [44], cardio- vascular - diseases [83], and chronic respiratory diseases [73], that occurred during 1958, 1960, and 1963, respectively. Expanding these national studies to include other diagnoses and updating them period- ically would be tremendously useful in interpreting differentials and trends in national cause-of-death statistics. The studies that are referenced in this bibliogra- phy used a variety of methods to evaluate the quality of medical certifications on death records. The three national studies noted above, for example, were based on death record follow-back surveys of the medical providers who had certified the causes of death or had formerly treated the decedents. It may be of interest to summarize the principal design features of these follow-back surveys: ® The scope of a survey is limited to particular cause(s) of death. ® Disease algorithms are developed that specify the criteria for evaluating the certitude of diagnosis for the causes reported on the death certificate. ® Survey questionnaires are designed to collect the information required by the disease algorithms. ® Samples of deaths certified to particular causes of death are selected from the Current Mortality Sample, which is a 10 percent national sample of death records that is processed on an accelerated schedule. ® Surveys are conducted with the medical certifiers and other medical providers that treated the sample decedents. ® Clinicians with specialties in the particular dis- eases work closely with the support staff that assesses the medical information reported in the survey. The studies referred to above have demonstrated that these procedures are well suited to evaluate the causes of death that are reported on the death certifi- cate. The methodology would have to be expanded, however, to detect the causes which should have been, but were not reported on the death certificates. Estimating the false negatives requires a complimen- tary type of follow-back survey that is based on a representative sample of death records that did not certify the particular causes of death under investi- gation. Both types of follow-back surveys are essen- tial in the evaluation of the quality of medical certi- fications. Many persons contributed to the preparation of this bibliography. The literature search and initial drafts of the annotations were performed primarily by Bruce Thompson, a graduate student at Johns Hopkins University, working under the direction of Dr. Alan Gittelsohn. Dr. Helen Abbey, Johns Hopkins University, also contributed to this work. Dr. Iwao M. Moriyama and Ms. Lillian Guralnick served as peer reviewers of this report, and they provided additional references as well as helpful suggestions. All of these contributions are gratefully acknowledged. Monroe G. Sirken Contents Fo Ee TR Tee =n TE 1 5 . 1 Baty he. ae ARORA a Fou eam 2 en Tans) La } ) Kl ) # oe wi ' } 3 . . * } ’ je x 3 w - a | . mat Tad = sid } v Cicehil L] ta pe L] } a } a = ~ h ) . BN Y - EN “ . - . - In E] } i a n . 2 “ Annotated Bibliography of Cause-of-Death Validation Studies Alan Gittelsohn, Ph.D., Johns Hopkins University, and Patricia N. Royston, National Center for Health Statistics Introduction This bibliography contains annotations for 128 articles pertaining to the quality of cause-of-death statistics. The articles are numbered in alphabetical order by surname of the author. Four requirements were imposed for including reports in this bibliography: 1. Reports deal with the validity of cause-of-death statements on death certificates or clinical records. 2. Reports were written in English, or English trans- lations were available. 3. Reports were written or published after 1957. 4. Reports were based on data (as opposed to opinion). Three sources were used in identifying relevant articles: 1. A library search for pertinent articles was initiated with the use of the Medical Literature Analysis and Retrieval System (MEDLARS) computer sys- tem. This search located articles published be- tween 1966 and 1977 whose titles and abstracts included specified words. The following phrases and keywords were used: cause with death, death with certificate, death certificate, death adjacent statistic, correlation, precision, accuracy, accurate, difference, reliability, reliable, validity (valid). The term death certificates was entered as the main point index heading. The Index Medicus was searched by using the same key words and phrases back to 1957 for coverage of the 23-year period—1958-80. The citation file was updated with the new entries obtained from this source. Letters requesting information about relevant articles were sent to all State health department statistical offices in the Nation, to selected depart- ments of epidemiology in medical schools and schools of public health, and to workers in vital statistics in national and international agencies. This effort added new citations, most of which were unpublished reports or reports on mortality that contain segments on reliability. Classification system Each article was classified and assigned corre- sponding codes according to (1) the data sources used, (2) the causes of death studied, (3) the country in which the deaths were registered, and (4) the latest data year. The four sets of codes assigned to each article appear in parentheses, separated by hyphens, to the right of the authors’ names. The section of this Bibliography entitled “Index” lists the articles by their code categories. The four sets of codes are described below. 1. Data sources Codes Definition A Autopsy reports B Other sources C Clinical records D Death certificates S Statistics This set of codes appears first and can contain up to 4 of the letters listed. The letters indicate the materials that were analyzed in the study. In most studies the cause of death on the death certificate was compared with clinical records, autopsy reports, or both. However, a number of studies are included that do not address the quality of the certified cause of death directly but do address the question of how accurately the cause of death can be determined without an autopsy by comparing autopsy results with clinical records. Other articles are included that are based on only one of the data sources listed. 2. Cause of death Codes Definition 1 Infectious diseases 2 Cancer 3 Cardiovascular disease 4 Stroke, cerebrovascular disease 5 Respiratory disease 7 Suicide 8 Other causes 9 Combination The cause-of-death code is a one-digit numeric code that follows the source-of-data code. Studies carrying codes 1-7 address errors in the assignment of only one disease or class of diseases as the cause of death. Code 8 indicates that the study deals with a single cause or class of causes of death other than those previously listed. Code 9 denotes that more than one cause or class of causes was studied. 3. Country Codes Definition AU Australia NW Norway CN Canada CY Ceylon FN Finland IR Ireland IS Israel JP Japan NW Norway NZ New Zealand SW Sweden UK United Kingdom Us United States 77 Several countries The country code generally refers to the country in which the study deaths were registered. A small number of studies (coded ZZ) investigated differences in diagnostic and classification practices among several countries. 4. Latest data year The latest-data-year code appears last and denotes the latest year for which records or statistics were Codes Definition examined. If the article omits this information, it is coded 99. 58-79 1958 through 1979 99 Years of data not given Bibliography 1. Anonymous (D-9-ZZ-99) The Accuracy and Comparability of Death Statistics WHO CHRONICLE 21(1):11-17, January 1967 To investigate variations in coding practices among countries, the World Health Organization (WHO) conducted an experiment in six European countries (Czechoslovakia, Denmark, England and Wales, Finland, the Netherlands, and Sweden) in which a sample of 1,000 death certificates was circulated to the six countries for coding. Each of the sample death certificates gave more than one cause of death and was chosen so that the underlying cause of death was not obvious. All certificates were coded by the WHO Center for Classification of Diseases, London, according to the Seventh Revision Inter- national Classification of Diseases (ICD), 1955 and this coding was used as the standard for comparison. Variation was appreciable from country to country in the extent of agreement with WHO codes, the agree- ment being lowest for Czechoslovakia and highest for Finland. Preliminary findings indicated that national differences in the interpretation of the selection rules caused most of the discrepancies. This finding led to two further studies. First, 218 disputed certificates were recirculated to the six countries for reconsid- eration of and comments on their original code assignments. It was found that the provisions of the ICD selection rules had been ignored in a number of instances. In a further investigation, the WHO asked each country for 20 death certificates in each of nine specified categories. These certificates were then reproduced and sent to all the other countries for coding. The returned codings varied significantly. The author suggested several ways to improve the quality of mortality statistics, including multiple-cause coding, new restrictions in the Eighth Revision of the ICD, and better training in filling out and coding death certificates. 2. Anonymous (D-9-US-73) Alabama’s Coroner System as it Relates to Vital Statistics: A Report to the State Committee of Public Health Department of Public Health, Division of Vital Statis- tics, Montgomery, Alabama, May 7, 1976. Unpub- lished. In 1973, 14.8 percent of the 35,239 deaths in Alabama were certified by people without formal medical training (i.e., by a coroner who is not a physician). The authors noted that any differences in certifying practices between physicians and coro- ners would bias mortality statistics in Alabama. To determine the extent of such differences, physician- certified deaths and coroner-certified deaths were compared for 1973. The comparison indicated that more than 40.5 percent of the nonexternal causes of death certified by coroners were placed in the cate- gory of Symptoms and Ill-Defined Conditions. Phy- sicians, on the other hand, had only 2.3 percent of their certificates in this category. The authors suggest three possible courses of action: (1) provide medical training for coroners, (2) provide the coroner with funds for medical consultation, or (3) replace the coroner system with a medical examiner system. 3. Acheson, Roy M.; Nefzger, M. Dean; and Heyman, Albert (BCD-4-US-68) Mortality From Stroke Among U.S. Veterans in Georgia and 5 Western States: II. Quality of Death Certification and Clinical Records. JOURNAL OF CHRONIC DISEASES 26:405-414, 1973 A previous report compared statistics on stroke mortality rates for veterans in Georgia with rates for five Western States (Colorado, Idaho, Montana, Utah, and Wyoming). To investigate reasons for the higher stroke mortality rates in Georgia, two groups of death certificates were selected for study. Death certificates of all veterans who died in the six study States be- tween July 1, 1967, and June 30, 1968, with a certificate entry coded for cerebrovascular disease (Seventh Revision ICD, codes 330-334) composed the case group certificates; a 30 percent sample of all other deaths in the six States during the same year constituted the control group. The resulting sample sizes were 604 cases and 1,210 controls for Georgia and 560 cases and 1,644 controls for the five Western States. Information about the death, prior illnesses, and medical care was collected from several sources, including hospitals, physicians, medical examiners, funeral directors, and relatives. In the West, the au- topsy rate was found to be higher, physicians signed death certificates of the case group more often and, before death, some common diagnostic tests were performed much more often. 4. Adelstein, A. M. (S-8-UK-71) Certification of Hypothermia Deaths BRITISH MEDICAL JOURNAL 1:482, 1973 This letter to editor describes how hypothermia deaths may be classified in one of four rubrics and how, as a symptom, it is disregarded when an ac- ceptable cause is mentioned. One table is presented showing deaths with mention of hypothermia by ICD coded cause. 5. Alderson, M. R. and Meade, T. W. (CD-9-UK-62) Accuracy of Diagnosis on Death Certificates Com- pared With That in Hospital Records BRITISH JOURNAL OF PREVENTIVE AND SOCIAL MEDICINE 21:22-29, 1967 The hospital records and death certificates were compared for 1,216 deaths occurring in hospitals in 1962 in the county borough of Oxford, Oxfordshire (except Henley M.B. and R.D.C.), Abingdon Borough, and Abingdon Rural Area. Both the underlying cause of death (UCD) and the principal condition treated in hospital (HD) were grouped using the ICD List B codes (abbreviated list of 50 causes for tabulation of mortality). The principal condition treated was compared to the diagnosis of the underlying cause on the death certificate for the entire 1,216 cases. The UCD and HD fell in different List B groups in 39 percent of the cases studied. Associations with indefinite diagnoses, length of stay, and hospital specialties led the authors to conclude that these discrepancies were not random occurrences. To investigate these discrepancies, hospital records for a random 1-in-12 sample of the 1,216 cases were examined to see whether these data supported the diagnoses of the principal condition treated and the diagnosis on the death certificate. In the authors’ opinion, only 7.2 percent of the cases had justifiable discrepancies. 6. Anderson, Donald O. (CD-5-CN-65) Geographic Variation in Deaths due to Emphysema and Bronchitis in Canada CANADIAN MEDICAL ASSOCIATION JOURNAL 98(5):231-241, 1968 To investigate the variation among Canadian provinces in mortality rates for selected chronic nonspecific respiratory diseases, a study was con- ducted with physicians in three provinces in 1965. The study sample included all death certificates registered in British Columbia, Manitoba, and Ontario with any one of four codes (Seventh Revision ICD) listed as underlying cause, and one-third of the certificates with any of the four codes listed as a contributory cause. The four codes were Unspecified bronchitis (code 501), Other chronic bronchitis (code 502.1), Bronchitis with emphysema (code 502.0), and Emphysema without mention of bronchitis (code 527.1). As soon as possible after each death was registered, the certifying physician was queried by mail about the basis for and the certainty of the diagnosis. World Health Organization criteria for a diagnosis of Chronic bronchitis (ICD codes 501, 502.0, 502.1) were met in 79.4 percent of the cases when bronchitis was the underlying cause and in 76.0 percent of the cases when bronchitis was a contrib- utory cause. Assignment of Pulmonary emphysema as the underlying or contributory cause of death was of acceptable validity in 85 percent of the cases, but only about 40 percent of the diagnoses were well established by pathological or high-grade physiolog- ical measures. 7. Anderson, Donald O. (S-5-CN-60) Observations on the Classification and Distribution of Pulmonary Emphysema in Canada CANADIAN MEDICAL ASSOCIATION JOURNAL 89:709-716, 1963 After a general review of epidemiologic principals relevant to mortality, the author examines chronic nonspecific pulmonary disease mortality rates by regions of the United States and Canada. Although regional differences in mortality may be due to altered etiologic factors and actual disease incidence, the author suggests that diagnostic convention may play a role. Examples of nosological inconsistencies are given, and the use of symptoms complexes for clinical prognosis rather than anatomic nomenclature is discussed. 8. Barclay, T. H. Crawford and Phillips, A. J. (CD-2-CN-56) The Accuracy of Cancer Diagnosis on Death Certificates CANCER 15:5-9, 1962 Seven thousand one hundred and forty-six deaths ascribed to cancer in Saskatchewan between 1950 and 1956 were reviewed in an attempt to con- firm each cancer diagnosis. Overdiagnosis was meas- ured as the proportion of all deaths due to cancer that were not confirmed. Underdiagnosis was meas- 6 ured by reviewing all death certificates with no men- tion of cancer and identifying those deaths that the Cancer Commission of Saskatchewan identified as having cancer. (The Cancer Commission, which operates free cancer diagnosis and treatment centers for residents, handles at least 85 percent of all cancer cases.) Cancer of stomach, lung, pancreas, large intes- tine, and prostate were overdiagnosed most often, while cancer of the skin and buccal cavity (including lips) were underdiagnosed most often. The authors conclude that death certificate diagnoses are insuf- ficiently accurate to permit their use as a reliable indication of the incidence of cancer. 9. Barraclough, Brian M. (S-7-ZZ-68) Differences Between National Suicide Rates BRITISH JOURNAL OF PSYCHIATRY 122:95-96, 1973 The author presents recorded suicide rates (SR) and undetermined rates (UR) for 22 countries for 1968. With the exception of Chile, where the UR is high relative to the SR, the relative rankings of the 22 countries based on SR and SR + UR are highly correlated (rho = 0.95). The implication is that the countries differ in SR, a point the author regards as justifying inquiries to explain the differences that range from 1 per 100,000 in Malta to over 40 in West Berlin. 10. Barraclough, Brian M. (S-7-UK-68) Are the Scottish and English Suicide Rates Really Different? BRITISH JOURNAL OF PSYCHIATRY 120:267- 273,1972 The author contrasts the criteria and procedures used in Scotland with those in England and Wales for deciding what evidence is necessary to write “suicide” on the death certificate. Post mortem examination occurred in 80 percent of the cases of violent or unnatural deaths in England and Wales in comparison with 40-50 percent in Scotland. Examination of the suicide, undetermined, and accident rates in the two countries for all causes and for selected groupings of violent deaths for adults in 1968 reveals a signifi- cantly higher occurrence of the undetermined cate- gory in Scotland. Over one-third of the poisonings by liquids or solids were undetermined as to suicide or accident in Scotland as contrasted with 16 percent in England and Wales. The higher undetermined rate in Scotland, in combination with the lower post mortem rates in Scotland, suggest that differences in suicide rates between the two countries may be an artifact of procedures and criteria. 11. Barraclough, Brian M.; (AB-7-UK-70) Holding, Trevor; and Fayers, Peter Influence of Coroners’ Officers and Pathologists on Suicide Verdicts BRITISH JOURNAL OF PSYCHIATRY 128:471- 474, 1976 In a Western London coroner’s district, 330 cases of suicide, accidents, and open verdicts were studied to determine the effect that coroners’ officers and pathologists have on suicide verdicts. All deaths re- sulting in an open verdict between January 1, 1969 to December 31, 1970, which were coded to the ICD codes E980-989, were included. To each of these deaths a suicide and an accidental death were matched according to age, sex, and cause of death. Analysis of these triples showed that no one pathol- ogist or coroner’s officer was associated with an excessive percentage of suicides, accidental deaths, or open verdicts, thus leading to the conclusion that these verdicts are determined by fact rather than the personal prejudices or feelings of the pathologist or the coroner’s officer. 12. Bauer, Fredrick W. and Robbins, Stanley L. (AC-2-US-65) An Autopsy Study of Cancer Patients: I. Accuracy of the Clinical Diagnoses (1955 to 1965), Boston City Hospital JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 221(13):1471-1474, 1972 In 1922, H. G. Wells found that of 578 autopsied patients with cancer, 36.5 percent had an incorrect diagnosis prior to death. Since then it has been gen- erally accepted that cancer death rates are underesti- mates but the extent of the error has not been known. This study at Boston City Hospital was initiated after a sample of autopsies indicated that substantial errors in cancer death certificates might exist. Of the 10,977 autopsies that were performed between 1955 and 1965, 2,734 (25 percent) of the reports were found to contain a diagnosis of cancer. Of the cancer diagnoses, 60 percent confirmed the clinical diagnosis. The 40 percent of cancer autopsy reports containing errors were divided as follows: 26.2 percent had un- diagnosed cancer, and 13.9 percent had incompletely diagnosed cancer. The authors stated that not all of the cancer diagnoses implied that cancer was the cause of death. Of the 2,734 cancer patients, 24 per- cent had nonmalignant causes of death. The authors conclusions were: (1) the largest sources of errors were in the deep organ malignancies, and (2) errors in diagnoses were not confined to those malignancies that did not influence the patients’ death; in 64 percent of the undiagnosed cancer, the tumors had extended beyond the primary site, and in 45 percent, the cancer was fatal. 13. Beadenkopf, William G.; Abrams, Malcolm; Daoud, Assaad; and Marks, Renee U. (AD-9-US-57) An Assessment of Certain Medical Aspects of Death Certificate Data for Epidemiologic Study of Arter- iosclerotic Heart Disease JOURNAL OF CHRONIC DISEASES 16:249-262, 1963 The population for this study consisted of 611 consecutively autopsied patients 45 years of age and over who died in the Albany Medical Center Hospital between March 1955 and September 1957. The underlying cause of death was determined by coding the death certificates according to the ICD. Of the patients who were shown at autopsy to have arterio- sclerotic heart disease, 50 percent were coded to the 420 ICD rubric (thus the sensitivity is 50 percent), while 80 percent of the patients found not to have arteriosclerotic heart disease were not coded in the 420 category (thus the specificity is 80 percent). Note that the study showed how often death certifi- cate cause of death reflected the presence of the condition, but did not measure the accuracy of the certified cause of death. The authors then used the same procedure for malignant neoplasms as cause of death. They found that the sensitivity and specificity were high for that disease category. They concluded that malignant neoplasms death certificates could be useful in epidemiologic studies involving associa- tions. However, because of low sensitivity in arterio- sclerotic heart disease diagnosis, any associations that might be found in a study involving death certificate data could be spurious. 14. Bonser, Georgiana M. and Thomas, Gretta M. (ACD-2-UK-54) An Investigation of the Validity of Death Certifica- tion of Cancer of the Lung in Leeds BRITISH JOURNAL OF CANCER 13(1):1-12, 1959 In a previous report concerning lung cancer in three regions of Britain, the authors had noted that more deaths were recorded than cases were diagnosed in hospitals around the Aberdeen-Leeds region. The possibility that sex was related to the tendency to overrecord was of particular interest. A total of 1,036 death certificates covering the years 1950-1954 were obtained from the files of the Statistics Depart- ment of the Medical Officer of Health for Leeds in which the cause of death was cancer of the trachea, pleura, lungs, or bronchi. These were matched where- ever possible with hospital records, clinical notes, and autopsy records. When the certificate of death was compared with the clinical findings, 96.5 percent of the certificates that had lung cancer recorded as the cause of death were supported by clinical or autopsy findings. When the accuracy of the clinical diagnoses was questioned, 15 percent of them were in ““. . .con- siderable doubt. ..”; however, for these cases the condition was generally confirmed at autopsy. With a hospital diagnosis of cancer of the lung, the diagnosis was compared with the cause of death coded by the Statistics Department of the Medical Office of Health for Leeds. This comparison showed a 92.5 percent agreement for the 879 hospital diagnoses of cancer. 15. Bourke, Geoffrey J. and Hall, Michael A. (BD-9-IR-63) A Study of Some Certified Causes of Death and Age of the Certifying Doctor JOURNAL OF THE IRISH MEDICAL ASSOCIA- TION 61(370):115-122, 1968 (Author’s summary) “The study reports the results of the examination of domiciliary death certificates in the Republic of Ireland in 1963. Five mortality causes were chosen to determine whether or not a certified cause of death could be related to the age of the certifying doctor. A clear relationship exists between older doctors and deaths classified as being due to ‘other myocardial degeneration’ (ICD, 422) for male patients only, and ‘senility without mention of psychosis’ (ICD, 794) for both sexes. The results are discussed and it is suggested that the use of the latter term be abandoned. With regard to the former term it would appear that the majority of such deaths would be more appropriately classified 8 under the heading ‘arteriosclerotic heart disease in- cluding coronary disease’ (ICD, 420). Further study in this field would be worthwhile, not only in Ireland but in other countries also, since accuracy of death certification has important epidemiological impli- cations.” Reprinted with permission of the Irish Medical Journal. 16. Briggs, Robert C. (AS-9-US-75) Quality of Death Certificate Diagnosis as Compared to Autopsy Findings ARIZONA MEDICINE 32(8):617-619, 1975 The author asserts that the death certificate is not valid in any case except those for which the certifying physician has access to the autopsy report. Estimating that the U.S. autopsy rate is 20 percent, the remaining 80 percent bear low validity. The author compares his personal 20-year series of 260 autopsied patients with the United States mortality statistics. Diseases of the heart are overestimated by at least threefold; tuber- culosis remains a leading cause of death. On this basis, he concludes that U.S. vital statistics on causes of death bear little relation to the actual state of affairs. 17. Britton, Mona (AC-9-SW-71) Diagnostic Errors Discovered at Autopsy ACTA MEDICA SCANDINAVICA 196(3):203-210, 1974 (Author’s abstract) ‘It has been questioned whether routine autopsies are needed any longer for control and correction of causes of death, specially in clear-cut cases. This question was therefore studied in connection with deaths in a department of internal medicine. Among 400 consecutive deaths autopsy was performed in 383 (96 percent). Causes of death diagnosed before autopsy were compared with those established by the same clinicians after autopsy. As main cause of death the clinical diagnosis was thereby confirmed as correct in 57 percent of cases, and as erroneous in 30 percent. In the remaining 13 percent it had not been possible to make a definite diagnosis ante mortem. Fewer diagnostic errors were encoun- tered among patients below than above 70 years of age. There were also fewer errors when clinical diag- noses had been considered fairly certain than when estimated only as probable. However, even in the case of deceased patients below 70 years of age with fairly certain diagnoses, 15 percent were revealed to be erroneous at autopsy. The main cause of death was a circulatory disorder in 67 percent of cases, and a neoplastic in 17 percent. All other groups of diseases together accounted for the remaining 16 percent of deaths. Clinical diagnoses of neoplasms were more seldom found to be erroneous than diagnoses of other groups of diseases. Contributory causes of death were clinically underestimated. Of the disorders established as contributory after post-mortem 46 percent had been unrecognized before death. It is concluded that autopsies are still needed for control and correction of causes of death, also in ‘clear-cut’ cases.” Reprinted with permission of Acta Medica Scandinavica. 18. Britton, Mona (AG9-SW-71) Clinical Diagnostics: Experience From 383 Autopsied Cases ACTA MEDICA SCANDINAVICA 196(3):211-219, 1974 (Author’s abstract) ‘The aim of the present study was to investigate whether the experience of clinical diagnostics could still further be enriched through routine autopsies. The question was studied by comparing diagnoses made by the same clinicians before and after autopsy in 383 subjects. Clinical misinterpretations thereby revealed were further analysed and described. Acute myocardial infarction (AMI) was the most common main cause of death. The diagnosis was seldom disproved when clinically considered fairly certain, but the disorder had often been missed ante mortem, especially among patients with known chronic ischemic heart disease (IHD). Hidden behind this latter label were also cases with valvular lesions or with lung disorders and right heart failure. Apart from chronic IHD, cerebrovascular diseases were clinically overdiagnosed as main cause of death; sometimes recent myocardial infarcts or malignant neoplasms were instead disclosed at au- topsy. Clinical diagnoses of neoplastic disorders were seldom found to be erroneous, but malignancy should be more often considered clinically. In several cases where it had been impossible to establish a definite diagnosis on clinical grounds acute abdominal dis- orders were revealed post mortem. An increased suspicion as regards these diseases seems warranted in obscure cases. The misinterpretations were frequently a consequence of our tendency to stick to earlier diagnoses and to overlook the development of new signs and symptoms. The same mechanism might partly explain why disorders contributing to death had often been unrecognized clinically—most fre- quently pulmonary embolism, AMI, cirrhosis of liver, and ulcer of stomach or duodenum. It is con- cluded that the Latin epigram regarding autopsies is still valid: ‘Mortui vivos docent’, let the dead teach the living.” Reprinted with permission of Acta Medica Scandinavica. 19. Burch, Thomas A. (BCD-9-US-73) Hawaii Mortality Followback Study: I. Introduction HAWAII STATE DEPARTMENT, RAND S REPORT ISSUE NO. 16, RESEARCH AND STATISTICS OFFICE, 1977 This paper is a preliminary report on the design of the Hawaii followback study, which explored the pos- sibility of collecting information on the extent of disability and the level of health care services received during the year preceding death. The author presents the questionnaires that were sent to physicians, hos- pitals, and informants. Selection criteria for the sam- ple are specified, and tables are presented to indicate characteristics of the sample population. For the results and conclusions of the study, see No. 113 in this report. 20. Burrows, Stanley (AC-9-US-73) The Postmortem Examination: Scientific Necessity or Folly? JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 233(5):441-443, 1975 To study whether recent advances in diagnostic techniques have lessened the need for time-consuming, expensive autopsies, postmortem findings were com- pared with clinical diagnoses for 252 adults who died while inpatients at Cooper Hospital in Camden, New Jersey. The major postmortem findings essentially agreed with the clinical diagnoses in 88 percent of the cases, with no significant differences by sex, age, or length of hospitalization. Of the 30 cases in which the postmortem findings differed substantially from the clinical diagnosis, a change of therapy would have been effective for about one-third. The autopsies were most productive in the cases of older adults hospitalized one day or less and patients with clinical problems involving the kidney, liver, gastrointestinal system, central nervous system, or of uncertain nature. This indicates that cases should be selected carefully for autopsy, to avoid unnecessary expense. 21. Carter, Ann P. and Lee, John A. H. (DS-2-US-67) Death Certification of Malignant Melanoma: A Prob- lem in Epidemiology AMERICAN JOURNAL OF EPIDEMIOLOGY 93(2): 77-78, 1971 For malignant melanoma from 1958 to 1967, the percent of deaths with primary site coded as un- known varied from 61 percent to 71 percent in the United States. This variation is in contrast to that of other skin cancer deaths in the United States during the same period of time, where the primary site was unknown in only 18-28 percent. In England and Wales from 1958 to 1967, 11-21 percent of melanoma deaths were coded as unknown primary site, and only 1-8 percent of the primary sites were unknown for other skin cancers, even though the same ICD classifi- cation was used. The authors suggest that the proportion of mela- nomas coded as unknown primary site could be reduced by adopting the procedure used successfully in England and Wales, where a followup postcard is mailed to the certifier when more detailed informa- tion is needed. 22. Carucci, Peter M. (CD-9-US-72) Reliability of Statistical and Medical Information Reported on Birth and Death Certificates NEW YORK STATE DEPARTMENT OF HEALTH MONOGRAPH NO. 15, 1979 In a study of the reliability of death certificates, 2,480 hospital records for patients discharged dead in 1972 were microfilmed in 96 upstate New York hospitals. The reliability of the cause-of-death state- ments were measured in terms of agreement with the hospital records. Agreement on underlying cause of death between the two sources at the following ICDA code levels—four digit, three digit, the disease category, and the major disease category—were 66, 74, 84, and 94 percent, respectively. Percent agreement at the disease category level is also shown for each disease category. Agreement at the disease category level for malignant neoplasms and cardio- vascular-renal disease were 92 and 84 percent, respec- tively. The author concludes that information from death certificates agreed well with information from hospital records. A parallel study on the quality of birth certificate information was conducted simul- taneously and is discussed thoroughly in this article. 10 23. Clarke, Cyril and Whitfield, A. G. W. (CD-8-UK-77) Deaths From Rhesus Haemolytic Disease in England and Wales in 1977: Accuracy of Records and Assess- ment of Anti-D Prophylaxis BRITISH MEDICAL JOURNAL 1669, 1979 1(6179):1665- Eight years after prophylactic anti-D gamma- globulin became generally available to prevent immunization of rhesus-negative women, deaths still occurred from haemolytic disease of the newborn (HDN). To investigate the circumstances in which these deaths occurred, copies of death certificates were obtained for all deaths that occurred in 1977 in which HDN was listed as a cause of death (54 live- born cases and 101 stillbirths). The causes of death listed on the certificates were compared with causes determined from a study of clinical records obtained from hospitals and obstetricians who had provided care. Overreporting of haemolytic disease occurred in about one-fourth of the liveborn cases because cases due to hydrops automatically were coded to HDN under the Eighth Revision of the ICD. HDN was underreported as the underlying cause of death in the liveborn cases because it was incorrectly recorded as a contributory cause on the death certificates. Among the 101 stillbirths, 31 were found to be due to a cause other than HDN. These errors in cause of death statistics create an inaccurate picture of the efficacy of anti-D prophylaxis. 24. Cooper, Edward S.; Cooper, Jean W.; and Schnabel, Truman G., Jr. (ACD-1-US-63) Pitfalls in the Diagnosis of Bacterial Endocarditis. A Review of 159 Patients, with Emphasis on 96 With Autopsy ARCHIVES OF INTERNAL MEDICINE 118:55-61, 1966 The report reviews associated clinical findings in 159 patients with bacterial endocarditis at the Phila- delphia General Hospital between 1954 and 1963. Of the 96 patients autopsied, 28 (29 percent) were cor- rectly diagnosed ante mortem. Of the clinically un- recognized, approximately one-half had endocarditis as a complication of an associated disease. Case reports are presented to illustrate difficulties in diag- nosis, and blood culture results are discussed. 25. Davis, Joseph H. (AC-8-US-64) Medical Examiner Problems Arising From Neuro- surgical Cases CLINICAL NEUROSURGERY 12:300-311, 1964 The author discusses the problems of clinically diagnosing intracranial hematomas, poisoning in the presence of head trauma, problems of alcohol, and certain other neurological causes of death. Several case studies are presented. 26. Dean, Geoffrey (DS-9-IR-66) The Need for Accurate Certification of the Cause of Death and for More Autopsies JOURNAL OF THE IRISH MEDICAL ASSOCIA- TION 62(386):273-278, 1969 In carrying out a study of lung cancer and bron- chitis in the Republic of Ireland, the author en- counters serious deficiencies in the system for certify- ing cause of death in the Republic of Ireland. He cites several examples of incorrect certification procedures and recommends specific changes in the system to improve the accuracy of the certified cause of death. A brief discussion of recent mortality trends is also included. 27. de Faire, Ulf; Friberg, Lars; Lorich, Ulla; and Lundman, Torbjorn (CD-9-SW-73) A Validation of Cause-of-Death Certification in 1156 Deaths ACTA MEDICA SCANDINAVICA 200:223-228, 1976 (Author’s abstract) “Swedish twins have been followed for mortality since 1961, when the Swedish Twin Registry was formed. During the years 1961-73 there were 1,290 deaths among twins born in 1901- 25. In 1,156 cases the cause of death could be estab- lished from collected records and classified according to the 1965 Revision of ICD. Using the review of records as the standard, rates of detection and con- firmation relating to the death certificate diagnoses were calculated. It is concluded that Swedish death certificate data are fairly valid for use in epidemiolo- gical studies and mortality statistics with regard to most cancer forms, cerebrovascular disease, ischemic heart disease, bronchitis, asthma and emphysema, accidents and suicides, but not for diabetes mellitus, alcoholism, mental diseases, rheumatic heart diseases and other heart diseases. However, in selected clinical- epidemiological studies it is often necessary to collect all available documents prior to judging the cause of death.” Reprinted with permission of Acta Medica Scandinavica. 28. Dorn, Harold F. and Cutler, Sidney J. (CD-2-US-47) Comparison of Death Certificates and Case Reports PUBLIC HEALTH MONOGRAPH NO. 56: Morbidity From Cancer In the United States, pp. 117-124. 1958. In 1938 the National Cancer Institute initiated a series of cancer morbidity surveys in 10 metropolitan areas, representing different geographic regions of the United States. These same areas were resurveyed in 1947 and 1948. The resurvey covered approxi- mately 10 percent of the total and 15 percent of the urban population of the United States. Data were col- lected from every physician, hospital, and clinic in the survey areas on every cancer patient diagnosed, observed, or treated during the 1947 calendar year. In addition, the records of local vital statistics offices were searched for information on deaths from cancer as a partial check upon the completeness of reporting. The primary purpose of the study was to estimate incidence, prevalence, and mortality rates for cancer. However, the study data permitted an evaluation of the validity of the cause of death entered on the death certificate. The primary site of cancer from the case report and the underlying cause of death from the death certificate were com- pared for each resident cancer patient reported as dying during the survey year or 6 months thereafter. The study group included 22,681 cancer cases. Detailed tables are presented showing agreement between the two records by primary site. Overall, 82 percent showed the same major site group, with the best agreement for leukemia (95 percent) and the digestive system (90 percent) and the poorest agree- ment for other and unspecified sites (33 percent) and soft tissue (42 percent). Survey results are compared for the 1947 and 1937 studies, showing little overall change in agreement in the major site group, but significant improvement for certain sites (respiratory system, skin, brain, and bone). 11 29. Ehrlich; Dov Li-Sik, Marcel; and Modan, Baruch (AD-2-IS-69) Some Factors Affecting the Accuracy of Cancer Diagnosis JOURNAL OF CHRONIC DISEASES 28:359-364, 1975 , From January 1, 1968 to December 31, 1969, 1,212 consecutive autopsy records of patients who died at the Chaim Sheba Medical Center, Tel Hashomer were screened. These represented 49 percent of the deaths in this hospital during the 2 year period. Subsequently, records of all cases with a diagnosis of a malignant neoplasm on either the post- autopsy record, or on the preautopsy request were selected and comprehensively reviewed. The cases were divided into three categories: (1) confirmed positives (226 cases), (2) false positives (43 cases), and (3) false negatives (28 cases). An advanced age, non-European ethnic origin, short length of terminal hospitalization, and lower rate of performance of laboratory tests were each found to be of major importance in misdiagnosis. 30. Engel, Linda W.; Strauchen, James A.; Chiazze, Leonard, Jr.; and Heid, Marian (ACD-9-US-70) Accuracy of Death Certification in an Autopsied Population With Specific Attention to Malignant Neoplasms and Vascular Diseases AMERICAN JOURNAL OF EPIDEMIOLOGY 111 (1): 99-112, 1980 (Author’s abstract) “Accuracy of certification of underlying cause of death and implications for US mortality statistics were assessed among 257 autopsied cases collected during the calendar year 1970 at a shortstay general hospital in Atlanta, GA. Clinicopathologic cause of death (CPCD) certi- ficates, with assignment of underlying cause of death based on autopsy findings in combination with pertinent clinical data, were prepared by a pathologist and were employed as a standard of comparison against which the accuracy of the underlying cause of death on the original death certificate was measured. Results suggest that autopsy findings are not nec- 12 essarily used to supplement clinical data in filling out death certificates. Improper recording of under- lying cause of death was found in 42 percent of the autopsied cases. Malignant neoplasms were found to be underreported and vascular diseases overreported, each by approximately 10 percent, when original certificates were compared to CPCD certificates. The confirmation rate for original death certificate diagnoses was 89 percent. In the case of a confirmed diagnosis, the underlying cause of death was sub- stantiated by postmortem findings as having existed regardless of its role in the sequence of events leading to death. The underlying cause of death as assigned by the pathologist was listed on the original death certificate among the sequence of events leading to or contributing to death at the rate of 72 percent (i.e., this rate measures the sensitivity of the death certifi- cate).” Reprinted with permission of the American Journal of Epidemiology. 31. Erhardt, Carl L.; Weiner, Louis; and McAvoy, Grace (AD-9-US-56) Pathological Reports for Mortality Statistics JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 171(1):119-122, 1959 Since 1945, New York City has required that autopsy reports be submitted to the vital statistics office to revise the original cause of death. However, a 1955 study showed poor compliance with this requirement. To determine whether the requirement should be enforced, a study was made of autopsy reports submitted in 1956, when special requests to hospitals resulted in almost twice as many autopsy reports (5,217) being filed as in 1955. Cause-of-death codes based on autopsy reports were compared with original coded causes, and tallies were made of cases in which assignment to another cause group was indicated by the autopsy report. A review of the 13 cause groups in which 20 or more changes were indicated showed that, numerically and proportion- ately, the changes between groups were small; only 15.2 percent of the reports indicated any change in the causes of death. The authors concluded that the expenditure in time, money, and effort to fulfill the requirement was not justified by the small im- provement in the accuracy of the mortality statistics. 32. Fedrick, Jean and Butler, N. R. (ACD-9-UK-58) Accuracy of Registered Causes of Neonatal Deaths in 1958 BRITISH JOURNAL OF PREVENTIVE AND SOCIAL MEDICINE 26:101-105, 1972 The 1958 British Perinatal Mortality Survey (BPMS) included a detailed study of the causes of 1,656 neonatal deaths that occurred in England, Scotland, and Wales during March, April, and May 1958. To investigate the shortcomings of the Seventh Revision of the ICD, causes ascribed by the BPMS were compared with causes coded by the General Register Office and with causes entered on death certificates. Anencephalus and rhesus isoimmuniza- tion were coded most accurately (89 percent); the greatest discrepancies were found for major renal malformations and pulmonary lesions. Even obvious congenital malformations such as spina bifida were found to be coded to other causes. Errors were attributed to the coding system and to the inatten- tion of physicians completing death certificates. 33. Florey, Charles Du V.; Senter, Margaret G.; and Acheson, Roy M. (DS-4-US-64) A Study of the Validity of the Diagnosis of Stroke in Mortality Data 1. Certificate Analysis THE YALE JOURNAL OF BIOLOGY & MEDICINE 40(2):148-163, 1967 A study was made of all death certificates of New Haven residents over 15 years of age who died during 1959-64 with cerebrovascular accidents (CVA) listed on the certificates as underlying (Group 1) or con- tributory (Group 2) causes of death. All causes of death listed on the certificates were coded by the authors, and the codes were compared with the underlying cause of death codes originally assigned by coders for published national mortality statistics. Of the 1,038 certificates in Group 1, only 10 CVA codes assigned by the authors differed from the original codes—seven were Health Department coding errors, and three were differences in opinion as to the most important cerebral event among several mentioned on the certificate. More vague CVA codes were assigned as contributory causes of death than as underlying causes, because coders generally selected more precise diagnoses as underlying causes. Some biases to which certain CVA diagnoses were subject also were noted. 34. Florey, Charles Du V.; Senter, Margaret G.; and Acheson, Roy M. (ACD-4-US-64) A Study of the Validity of the Diagnosis of Stroke in Mortality Data II. Comparison by Computer of Autopsy and Clinical Records with Death Certificates AMERICAN JOURNAL OF EPIDEMIOLOGY 89(1): 15-24, 1969 (Author’s abstract) “Clinical records of residents of New Haven, Connecticut, who died from cerebro- vascular disease between 1962-1964 were abstracted for data concerning the final strokes. Of 690 possible records, 620 were abstracted. The method by which the records were reviewed by computer is described in detail along with the criteria for diagnosis of the different types of strokes. The diagnoses for each decedent found in the autopsy report, clinical record, death certificate and by computer were compared with each other and graded as to whether they were identical, similar or dissimilar. The diagnoses from the autopsies and records (133 cases) agreed in 79 percent of cases; from autopsies and certificates in 65 percent of cases; and of diagnoses from all the records abstracted and which were not for decedents who had definitely died in the absence of cerebro- vascular disease (607 cases), 74 percent agreed with the certificates. The reliability of the diagnoses was graded into five categories ranging from autopsy confirmation to no evidence available. Evidence from autopsy diagnoses and good clinical records indicated that cerebral hemorrhage was over-diagnosed on death certificates at the expense of thrombo- embolism. The hemorrhage-thrombosis ratio from certificate data was 2.7:1 and from autopsy and clinical data .78:1.” Reprinted with permission of the American Journal of Epidemiology 35. Garcia-Palmieri, Mario R.; Feliberti, Manuel; Costas, Raul Jr.; Benson, Herbert; Blanton, James H.; and Aixala, Ramon (ABCD-3-US-63) Coronary Heart Disease Mortality—A Death Certif- cate Study JOURNAL OF CHRONIC DISEASES 18:1317-1323, 1965 To evaluate the apparent lower mortality from coronary heart disease (Seventh Revision of the ICD, 420) for males 45-64 years of age in Puerto Rico 13 compared with the United States, the authors col lected diagnostic data on 674 deaths occurring between May 1, and October 31, 1963 in San Juan of persons 20-64 years of age. Autopsy and hospital records were reviewed; physician and family inter- views were conducted independently and a new cause of death was chosen, recoded by routine coders, and compared with the one on the original death certif- icate. A net gain in death rates from coronary heart disease based on the revised cause of death statements did not explain the difference in rates in the United States and Puerto Rico. Eight tables detail the sources of data, autopsies, age and sex distribution of deaths, and corrected coronary heart disease rates for Puerto Rico and the United States. 36. Gittelsohn, Alan (CDS-9-US-72) Notes on Data Quality Cooperative Health Information Center of Vermont, Inc., Burlington, 1974. Unpublished. The author examines questions of the reliability of health data sets by using hospital and mortality records maintained by the Cooperative Health Infor- mation Center of Vermont State Department of Health. The 94,248 patient discharge abstracts contained in the 1972 Vermont hospital file were studied to determine the frequency of occurrence of impossible code combinations. For sex and diagnosis incompatibilities, the rate was 137 errors per 100,000 entries; for sex and procedure incompatibilities, the rate was 63. Over half of the errors were accounted for by two hospitals contributing less than 20 percent of the case load. Hospital abstracts for the years 1969-71 were also reviewed to evaluate the consistency of the final diagnosis explaining admission with the types of surgical procedures performed. For the eight types of surgery studied, the final diagnosis was consistent with the surgery performed in over 95 percent of the cases. Finally, hospital and death records were linked for 5,824 patients with discharge status of expired during 1969-71. The underlying cause of death from the death record was compared with all diagnoses coded on the hospital record, and the degree of agreement was coded on a scale of 1 (agreement at the 3-digit level of the ICDA) to 4 (no agreement). Major discrepancies between hospital and death records were found. Discrepancies were related to cause of death, size of hospital, and length of stay, but were not related to physician specialty or number of diagnoses. 14 37. Gittelsohn, Alan and Senning, John (CD-9-US-75) Studies on the Reliability of Vital and Health Records: I. Comparison of Cause of Death and Hospital Record Diagnoses AMERICAN JOURNAL OF PUBLIC HEALTH 69(7):680-689, 1979 (Author’s abstract) “Based on computer linkage of death records and hospital discharge abstracts, underlying cause of death and discharge diagnoses are compared for 9,724 Vermont resident in-hospital deaths occurring between 1969 and 1975. The agreement between the diagnoses recorded in the two data systems provides a measure of the reproduci- bility of recording, abstracting, and coding practices. Using the first three digits of the International Classification of Diseases, the agreement between cause and closest medical record diagnosis was 72 percent. Concordance declined by patient age and length of hospital stay and varied significantly by coded cause of death. A major source of variation was the hospital of death where agreement levels ranged between 45 and 84 percent. The latter finding is regarded as a potential starting point for targeting investigation of sources of discrepancy and initiating efforts to improve diagnosis recording and coding in the two record systems. The value of both depends on continuing efforts to improve and maintain data quality.” Reprinted with permission of the American Journal of Public Health. 38. Goldacre, M. J. (CD-1-UK-73) Accuracy of Death Certification for Acute Bacterial Meningitis PUBLIC HEALTH, LONDON 91:279-281, 1977 Case records were collected and studied for every identifiable case of acute bacterial meningitis or meningococcal disease occurring in children under 10 years of age in the Northwest Metropolitan Region between 1969 and 1973. Of the 94 deaths identified in this search, 90 percent were coded under one of the rubrics for meningitis or meningococcal disease, while the remaining deaths were attributed to laryn- gotracheitis, prematurity, gastroenteritis, septicemia, bronchopneumonia, chronic lymphoid deficiency, and otitis media. The acute meningitis that had been diagnosed in these children was not mentioned on the death certificate in six of the nine cases assigned another cause. It was found that many inaccuracies could have been avoided by recording available information on the death certificate. 39. Gorham, L. Whittington (AC-3-US-56) A Study of Pulmonary Embolism. Part 1. A Clinico- pathological Investigation of 100 Cases of Massive Embolism of the Pulmonary Artery: Diagnosis by Physical Signs and Differentiation from Acute Myo- cardial Infarction ARCHIVES OF INTERNAL MEDICINE 108:76- 90, 1961 One hundred cases of massive pulmonary embo- lism were selected for study from 5,700 autopsies performed between 1934 and 1956 in New York Hospital. A review of autopsy protocols and clinical records indicated that myocardial infarction and strokes tended to .be clinically diagnosed at the expense of embolization. The author describes 12 physical signs that he believes could be used to improve the accuracy of a pulmonary embolism clinical diagnosis. 40. Green, Adele and Donald, K. J. (AD-9-AU-73) Necropsy As a Control of Death Certification: Some Unexpected Findings MEDICAL JOURNAL OF AUSTRALIA 2(4):131- 132, 1976 (Author’s abstract) “A comparison of death certificates and necropsy findings in a group of premenopausal women suggests that a number of diseases are either underdiagnosed [or overdiagnosed] in life. Atypical cases of intracerebral haemorrhage are frequently misdiagnosed. These occur in the frontal, temporal or parietal lobes in nonhypertensive women and may be suitable for surgical treatment. Their aetiology remains obscure. Pulmonary embolus is habitually underdiagnosed in premenopausal women and myocardial infarcts appear to be over- diagnosed. The study reemphasizes that death certif- icates are inaccurate and that low necropsy rates render accurate statistics of diseases in the commu- nity difficult to obtain.” NOTE: The study group included all women residents of Brisbane City, Australia, between the ages of 35 and 49 who died during the years 1969 through 1973. Reprinted with permission of the Medical Journal of Australia. 41. Griffith, G. Wynne (S-2-UK-73) Cancer Surveillance with Particular Reference to the Uses of Mortality Data INTERNATIONAL JOURNAL OF EPIDEMIOLOGY 5(1):69-76, 1976 As a section of a larger report, the accuracy of cancer mortality statistics is reviewed. Citing several references and presenting a table of the percent cer- tified cancer deaths in 12 cities before and after a review of diagnostic evidence, the author conjectures that cancer deaths may be underestimated by an average of 5 percent. On the basis of other reports, the author concludes that changes in the ICD classifi- cation of cancer deaths over the years has not re- sulted in any major changes in “secular trends” of cancer mortality rates. However, he did warn that as the ICD coding of cancer sites becomes specific, errors in ICD classification become more pronounced. 42. Griffith, G. Wynne and Morgan, G. A. V. (S-9-UK-53) Diagnostic Precision as a Factor in Male Mortality Data BRITISH JOURNAL OF PREVENTIVE AND SOCIAL MEDICINE 15:68-78, 1961 To investigate relationships between causes of death, a factor analysis was performed on the crude male mortality rates for 167 of the administrative areas of England and Wales for the 4-year period 1950-53. The first principal component identified in the factor analysis, which accounted for 64.3 percent of the total variance, was found to be concerned with the size of population by age. Two components that accounted for much of the remaining variance for cer- tain causes of death were thought to be aspects of diagnostic precision. The basis for and implications of this interpretation of the two components are dis- cussed. 43. Gwynne, J. F. (AD-9-NZ-73) Death Certification in Dunedin Hospitals Report to the Medical Research Council of New Zealand, Unpublished Report, 1976 Death certificates and autopsy protocols were compared for a study group comprised of 643 post mortem examinations performed between October 1, 1971 and September 30, 1973 in Dunedin, New 15 Zealand. Stillbirths, emergency room deaths, and coroner’s cases were excluded. Frequency distribu- tions of both cause-of-death categories and demo- graphic characteristics of the study population are presented in tables. The causes of death are sorted into 17 major disease groups. Errors in diagnosing specific disorders within each group are tabulated by type of diagnostic error. The author reports that, overall, 57.5 percent of the certificate causes of death had significant errors when compared with the post mortem findings. In conclusion, he presents four recommendations for improving mortality statistics. 44. Haenszel, William; Loveland, Donald B.; and Sirken, Monroe G. (BD-2-US-58) Lung-Cancer Mortality as Related to Residence and Smoking Histories. 1. White Males JOURNAL OF THE NATIONAL CANCER INSTI- TUTE 28(4):947-1001, 1962 To study variations in lung cancer mortality by lifetime residence history, controlling for smoking history, a 10-percent sample was selected of all white male lung-cancer deaths in the United States during 1958. Residence and smoking histories were collected from family informants and additional diagnostic details from certifying physicians. Residence and smoking histories were also obtained for a sample of the general population from the U.S. Bureau of the Census Current Population Survey. Certifying physicians were queried by mail about the methods used to establish the diagnosis, whether tissue had been microscopically examined, and the histologic type. Seventy-six percent of the 2,381 sample death certifications were based on microscopic examination of tissue, and 66 percent were based on tissue from the primary site. Most of the remaining cases (18 percent) relied on X-ray findings. The diagnosis was histologically confirmed more often for decedents under 65 and for decedents who had lived in metropolitan counties, although the differ- ences by residence were minor. 16 45. Hallén, J. and Nordén, J. (AC-8-SW-60) Liver Cirrhosis Unsuspected During Life: A Series of 79 Cases JOURNAL OF CHRONIC DISEASES 17:951-958, 1964 During the 10-year period 1951-60, 360 cases of liver cirrhosis were diagnosed at Malmo General Hospital in Malmo, Sweden. Of these cases, 115 were diagnoses at necropsy and had not been suspected during life. Seventy-nine of the 115 unsuspected cases presented a histologically unequivocal picture and, therefore, were selected for further study. Clinical symptoms and abnormal laboratory findings for these cases were rare and often could be explained by co- existing diseases of the aged. Liver cirrhosis was found to be the main cause of death in 14 percent of the 79 cases. 46. Heasman, M. A. and Lipworth, L. (ACD-9-UK-59) Accuracy of Certification of Cause of Death GENERAL REGISTER OFFICE, STUDIES ON MEDICAL AND POPULATION SUBJECTS, NO. 20, Her Majesty’s Stationery Office, London, 1966 This study of 9,501 deaths in 75 hospitals in England and Wales in mid-1959 was designed to estimate the effect of autopsy examinations on mor- tality statistics. A dummy death certificate was com- pleted by the clinician without reference to autopsy findings, and a second one was done by the patholo- gist. They were compared after coding by staff of the General Register Office and complete agreement (all four digits of the Sixth Revison of ICD) was found for 45.3 percent of the deaths. The authors note that disagreements of fact (i.e., where the underlying cause on one certificate was not mentioned on the other) occurred in 25 percent. The underlying cause of death is presented by ICD categories, and estimates of the alterations in mortality statistics are calculated. The authors acknowledge that autopsies are not un- biased or free from error as a standard for accuracy. These findings apply only to hospitalized deaths (roughly half of all deaths in England and Wales) and may have been biased toward the group of deaths posing difficulty diagnostically. The study design assumes that all clinician’s certificates are completed without benefit of post mortem examination, though the authors recognize this is not the case. The authors note that extension of these results to national statis- tics requires knowledge of the proportion dying out of the hospital, the number autopsied, age of death, and the cause of death, since the accuracy varies greatly for these characteristics. 47. Hewitt, David; Milner, Jean; and Csima, Adele (S-9-US-59) Some Proposed ‘‘Comparability Areas” for U.S. Sta- tistics on Cause of Death PUBLIC HEALTH REPORTS 84(10):857-863, 1969 (Author’s summary) “After becoming qualified at a particular medical school, physicians do not disperse uniformly all over the United States but tend to take up practice in circumscribed regions. Because of variations in diagnostic preferences and in the medical vocabulary among medical schools, and consequently among their graduates, these geographic patterns of physician settlement can give rise to spurious differ- ences between States in statistics on causes of death. An index is therefore proposed for measuring the degree of comparability between any pair of States, together with a method for building up ‘compar- ability areas’ in which interstate comparisons will have some assurance of validity. Fourteen compar- ability areas are proposed, based on the known geo- graphic distributions of medical school alumni in 1959. All but 13 States have a place in one or more of these areas.” 48. Hickey, Noel and Mulcahy, Risteard (S-3-IR-65) The Significance of Changes in Certified Coronary Heart Disease Mortality Rates in Ireland IRISH JOURNAL OF MEDICAL SCIENCE 3(4): 163-168, 1970 To determine whether the documented increase in male mortality from coronary heart disease (CHD) in the Republic of Ireland was real, a study was made of the age-adjusted standardized death rates for 1952- 65. A comparison of mortality data for males 35-69 years of age for the two periods 1952-54 and 1963-65 revealed that the increase in deaths assigned to CHD (ICD codes 420-422) was significantly greater than the increase in deaths assigned to all codes that could include deaths from CHD (ICD codes 334, 401416, 420-422, 434, 441-446, and 794). The corresponding decrease in male deaths ascribed to heart disease other than 420-422 was not sufficient to counter- balance the increase in deaths ascribed to 420-422. The authors conclude that the increase in deaths from CHD is real and that the increase is caused to some extent by certain risk factors found in an affluent society. 49. Hoffman, Paul M. and Brody, Jacob A. (CD-8-US-63) The Reliability of Death Certificate Reporting for Amyotrophic Lateral Sclerosis JOURNAL OF CHRONIC DISEASES 24:5-8, 1971 Hospital records of North Carolina residents with a definite diagnosis of amyotrophic lateral sclerosis (ALS) in three hospitals for the period 1958-63 were reviewed. Death certificates were obtained for 72 of the 85 cases. Forty-eight had ALS as the underlying cause of death and four listed ALS as a contributory cause. Of the 20 certificates with no mention of ALS, 8 were coded as multiple sclerosis and 4 as muscular dystrophy. The authors conclude that mortality data, even in a disease with a distinct clinical picture, should be interpreted with great caution. 50. Holding, T. A. and Barraclough, Brian, M. (CD-7-UK-70) Psychiatric Morbidity in a Sample of a London Coroner’s Open Verdicts BRITISH JOURNAL OF PSYCHIATRY 127:133- 143, 1975 (Author’s summary) “One hundred and thirty- four deaths recorded as open verdicts in the Inner West London Coroner’s District during 1969 and 1970 have been surveyed for evidence of mental ill- ness. For 110 (82 percent) of these deaths the prob- able verdicts were suicide or accident and they were reclassified as undetermined deaths. Of these deaths 73 percent were diagnosed as mentally ill, 54 percent were receiving medical treatment for psychological symptoms before death, 42 percent had a history of psychiatric care and 24 percent had made a previous 17 suicide attempt. In these respects undetermined deaths and suicide deaths resemble each other; both are drawn predominantly from the mentally ill.” Reprinted with permission of the British Journal of Psychiatry. 51. Holler, Jacob W. and De Morgan, Nicholas P. (AC-9-US-99) A Retrospective Study of 200 Post-Mortem Exami- nations JOURNAL OF MEDICAL EDUCATION 45:168-170, 1970 In a review of 212 autopsies conducted in a university-affiliated teaching hospital, the authors compare clinical with autopsy findings. Complete agreement of clinical-autopsy diagnoses was noted for 52 percent of the cases. In an additional 24 percent, the major diagnosis was correct and minor diagnoses were either missed or wrong: 16 percent of the cases had a correct major diagnosis with confirmation of questionable diagnoses by autopsy and 8 percent had errors in major diagnoses. The authors conclude that in half the cases reviewed, autopsy had provided useful information. 52. Hook, Ernest B.; Farina, Matthew A.; and Hoff, Margaret B. (CD-3-US-73) Death Certificate Reports of Cardiovascular Dis- orders in Children: Comparison with Diagnoses in a Pediatric Cardiology Registry JOURNAL OF CHRONIC DISEASES 30:383-391, 1977 Death certificate diagnoses, clinical diagnoses, and the codes of underlying causes of death were com- pared for 294 deceased children listed in a pediatric cardiology registry. The registry listed all individuals evaluated by the Division of Pediatric Cardiology at Albany Medical Center between midyear 1967 and December 31, 1973. Based on a review of clinical records, it was judged that in 206 cases, the cardio- 18 vascular disorder probably contributed to the death. Of these, the mention of a cardiovascular disorder was made on 90.3 percent of the certificates; how- ever, the major cardiac defect present was specified in only 39.3 percent. 53. Howell, Trevor H. (AC-9-UK-99) Causation of Diagnostic Errors in Octogenarians: A Clinico-Pathological Study JOURNAL OF THE AMERICAN GERIATRICS SOCIETY 14(1):41-47, 1966 To emphasize the difficulties in formulating diagnoses in geriatric medicine, clinical diagnoses and pathological verdicts were compared for 50 octogenarians coming to autopsy. The two sources did not coincide in 50 percent of the cases. Cause of death was clinically ascribed most often to cerebral arteriosclerosis (24 percent), diagnosis uncertain (18 percent), senile dementia (16 percent), and cerebro- vascular accident (14 percent). Pathologists most often attributed the cause of death to atherosclerosis (28 percent), bronchopneumonia (24 percent), myocardial degeneration (18 percent), and cancer in some organ (10 percent). The factors that contrib- uted to incorrent clinical diagnoses are discussed. 54. Israel, Robert A. and Klebba, A. Joan (DS-9-US-68) A Preliminary Report of the Effect of Eighth Revi- sion ICDA on Cause of Death Statistics AMERICAN JOURNAL OF PUBLIC HEALTH 59(9):1651-1660, 1969 (Author’s abstract) “The introduction of the Eighth Revision of the International Classification of the Causes of Death has resulted in breaks in mortality trends for many causes. The reasons for these breaks are discussed. To deal with time trends without distortion due to changes in classification, comparability ratios have been prepared and these are explained here.” Reprinted with permission of the American Journal of Public Health. 55. Jablon, S.; Angevine, D. M.; Matsumoto, Y. S.; and Ishida, M. (AD-9-JP-62) On the Significance of Cause of Death as Recorded on Death Certificates in Hiroshima and Nagasaki, Japan NATIONAL CANCER INSTITUTE MONOGRAPH NO. 19: Epidemiological Approaches to the Study of Cancer and Other Chronic Diseases, 1966. pp 445- 465. The Japanese National Institute of Health— Atomic Bomb Casualty Commission Life Span Study, begun in 1947, was concerned with studying survivors of the atomic bombings of Hiroshima and Nagasaki. The sample included a study group of survivors who were within 2,500 meters of the hypocenter at the time of the bombings and two control groups. The sample was followed until death. Autopsies were performed on 1,215 of the study deaths between 1950 and 1962, and the principal autopsy diagnosis was compared with the underlying cause of death recorded on the death certificate. Agreement was good for deaths from malignant neoplams (although these were somewhat underdiagnosed) and for deaths from major cardiovascular-renal diseases when these diseases were considered as a class. Tuberculosis as a certified cause of death became increasingly unreli- able over time; however, accuracy of diagnosis of malignancies seemed to improve. 56. Jensen, Ole Moller; Mosbech, Johannes; Salaspuro, Mikko; and Jhamaki, Timo (CD-2-22-71) A Comparative Study of the Diagnostic Basis for Cancer of the Colon and Cancer of the Rectum in Denmark and Finland INTERNATIONAL JOURNAL OF EPIDEMIOLOGY 3(2):183-186, 1974 The incidence of and mortality from cancer of the colon and cancer of the rectum are reported to be 2-2Y% times higher in Denmark than in Finland. To investigate these differences all death certificates were examined that listed either of these cancers as primary or secondary cause of death and where death had occurred between May and July 1971. No differences were found in rates of registration of other gastrointestinal diseases. Furthermore, the diagnostic criteria for these two cancers were found to be identical for the two countries. The authors assert that the recorded difference between Finland and Denmark is real and not a statistical artifact. 57. Kagan, Aubrey; (ACS-4-Z7-66) Katsuki, Shibanosuke; Sternby, Nils; and Vanécek, Rudolf Reliability of Death Certificate Data on Vascular Lesions Affecting the Central Nervous System BULLETIN OF THE WORLD HEALTH ORGANI- ZATION 37:477-481, 1967 (Author’s abstract) “Although it is generally stated that national mortality statistics are to some extent unreliable, it is difficult to ascertain their degree of unreliability. In studies carried out in limited areas of Czechoslovakia, Japan and Sweden it has been possible to determine the cause of death at autopsy in a large series of cases, and the findings relevant to ‘vascular lesions affecting the central nervous system’ (CNS) have been compared with the national mortality statistics for the same causes and with the clinical findings. “It was found that death rates for vascular lesions affecting the CNS, taken as a whole, obtained from the autopsy studies were close to the national figures but that the ratio of cerebral haemorrhage to cerebral infarction was lower in the autopsy data, indicating that classification into cerebral haemorrhage and cerebral thrombosis is incorrect in the national figures and that the former is over-diagnosed at the expense of cerebral thrombosis. “Comparison of the autopsy data with clinical diagnoses in Prague and Malmo showed that only 70 percent of the cases of fresh cerebral vascular accidents regarded as the principal cause of death by the pathologist were diagnosed before autopsy, and that many cases found by the pathologist but not considered to be the principal cause of death were not suspected clinically. Less than 1 percent of cases of cerebrovascular accident regarded as the principal cause of death by the clinician were not found by the pathologist.” Reprinted with permission of the Bulletin of the World Health Organization. 19 58. Khoury, Sami A. (ACD-1-US-67) Death Certificates and Tuberculosis Register Cards. A Correlation Study of 108 Cases AMERICAN REVIEW OF RESPIRATORY DISEASE 104(6):936-937, 1971 The United States has a tuberculosis death rate of approximately 3 per 100,000 yearly. The study was designed to investigate the diagnostic accuracy of deaths coded as tuberculosis. The study group consisted of 108 patients who died between January 1, 1966 and December 31, 1967 and whose death certificates were matched with a tuberculosis register card. Forty-six percent of the deaths with tubercu- losis as immediate or underlying cause were inactive cases on the registry. The authors conclude that tuberculosis tends to be overreported as a cause of death. 59. Knight, Mark V. (CD-2-US-99) The Accuracy of Mortality Statistics Bureau of Health Statistics, Division of Health, Wisconsin Department of Health and Social Services, Working papers, Sept. 17, 1976, and Feb. 21, 1977. Unpublished. A sample of 550 death certificates with cancer as the underlying cause were selected from certificates filed with the Wisconsin Bureau of Health Statistics (years of data not given). The certifiers were queried by mail regarding (1) the methods used prior to death to diagnose the cancer, (2) the methods used at or after death in determining that cancer was the under- lying cause, and (3) the certifier’s confidence in the diagnosis of cancer as the underlying cause. Eighty- seven percent of the 513 returned questionnaires indicated that histopathological confirmation of the cancer diagnosis was obtained at some time. However, only one-third received histopathological confirma- tion at or after death. The quality of the diagnostic information was related to age of decedent and place of death, but not to site. A second sample of 300 certificates was drawn from those certificates with an underlying cause other than cancer. Certifiers were asked whether there was a history of cancer in the deceased. Twenty-one (7.5 percent) said yes, (but for 18 of these, the cancer was under control), 288 said no, and 30 said they did not know. 20 60. Krueger, Dean E. (D-9-US-62) Hypertensive and Chronic Respiratory Disease Mortality: Confirmation of Trends by Multiple Cause of Death Data PUBLIC HEALTH REPORTS 81(2):197-198, 1966 This study investigates a possible cause of the trends in hypertensive and chronic respiratory disease rates since 1949. Death certificates of veterans holding life insurance policies issued before 1940 and who died between July 1954 and the end of 1962 were collected; up to three causes of death were coded for each death. Average annual age-specific death rates were calculated for each of three sub- divisions of an 8% year period, for hypertensive disease and for bronchitis and emphysema as under- lying causes and as associated causes. Age-adjusted rates are presented. For hypertensive disease, both the underlying cause and the associated cause rates decreased. The rates for bronchitis and emphysema as the underlying cause increased, as did the rates for these conditions as an associated cause. The author concludes that the trends for these conditions are not simply caused by a shift in reporting these diseases from associated causes to underlying causes, and offers explanations for the trends. 61. Kuller, Lewis H.; Anderson, Herbert; Peterson, Donald; et al. (ACD-4-US-65) Nationwide Cerebrovascular Disease Morbidity Study STROKE 1:86-98, 1970 This study, designed to investigate geographic differences in stroke morbidity and mortality rates, was carried out in eight of the nine areas that partici- pated in the Nationwide Mortality Study (see Number 63 in this report). Stroke cases were ascer- tained through the diagnostic indexes of participating hospitals for white persons aged 45-69 years during 1965, and through case lists used in the Nationwide Mortality Study. Analysis was limited to the six areas in which at least 90 percent of the hospital records were available for review. The high mortality areas in the Southeastern United States apparently also have a higher incidence of stroke than the low stroke mortal- ity areas. About 80 percent of the stroke diagnoses could be verified by an autopsy report, arteriography, hemorrhagic spinal fluid, hemiplegia, or coma on admission. 62. Kuller, Lewis H.; Blanch, Thomas; and Havlik, Richard (CD-4-US-65) Analysis of the Validity of Cerebrovascular Disease Mortality Statistics in Maryland JOURNAL OF CHRONIC DISEASES 20:841-851 1967 To study the validity of cerebrovascular disease diagnoses, a sample of death certificates, stratified according to underlying cause of death, were selected from certificates for deaths of persons aged 40 to 75 who died in Baltimore City in the first 6 months of 1965 and in the eastern and western areas of Maryland in all of 1965. Information about the deaths was gathered from several sources, including hospitals, attending physicians, and medical examiners. A review of the available information showed that of the 364 deaths in hospitals for which cerebrovascular disease was mentioned on the certificate, the disease was probably not present in 19.8 percent. Further- more, cerebrovascular disease was not mentioned for 19.8 percent of the 356 hospital deaths that were attributed to the disease as a result of the review. Numerous other statistics relating to the validity of the certified cause of death were presented. 63. Kuller, Lewis H.; Bolker, Abraham; Saslaw, Milton, S.; et al (ABCD-4-US-65) Nationwide Cerebrovascular Disease Mortality Study. I. Methods and Analysis of Death Certificates AMERCIAN JOURNAL OF EPIDEMIOLOGY 90(6): 536-544, 1969 (Author’s abstract), “Large differences in cere- brovascular disease (CVD) mortality among the geographic areas of the United States have been reported. In order to determine whether these geo- graphic differences might be due to differences in certification practices or accuracy of the diagnosis of stroke, a study of death certificates in 9 areas of the United States — 3 with high, 3 intermediate and 3 with low cerebrovascular disease death rates for white males ages 35-74—was completed. A stratified sample of 6,314 death certificates was included in the study. The information on the death certificate was then compared with clinical data from hospital records, physicians reports and medical examiner's records. Cerebrovascular disease had been listed as underlying cause of death on 1,232 death certificates in the original sample and 1,310 after adjusting for sampling. Stroke was listed as underlying cause for 1,310 (66.8 percent) of the 1960 certificates listing stroke as either underlying or contributing cause of death. There were no substantial differences among the areas. The age and sex distribution, place of death, type of stroke listed on the certificate and the frequency with which stroke appeared on the same certificate with hypertension, arteriosclerotic heart disease and diabetes were similar in the high, low and intermediate cerebrovascular death rate areas. The variations in death rates could not be explained by differences in certification practices such as choice of the underlying cause of death among all death certificates listing stroke.” Reprinted with permission of the American Journal of Epidemiology. 64. Kuller, Lewis H.; Bolker, Abraham; Saslaw, Milton S.; et al. (ABCD-4-US-65) Nationwide Cerebrovascular Disease Mortality Study. II. Comparison of Clinical Records and Death Certi- ficates AMERICAN JOURNAL OF EPIDEMIOLOGY 90(6): 545-555, 1969 (Author’s abstract) “The large differences in cerebrovascular disease mortality among geographic areas of the United States cannot be explained by variations in certification practices such as the choice of the underlying cause of death. There may be differences in the frequency that clinical stroke diagnoses on hospital records, physician’s and medical examiner’s reports or from a family interview were listed on the death certificate. In order to measure the relationship between clinical stroke diagnosis and information on the death certificate, pertinent clinical records were reviewed for every death certi- ficate included in the nationwide cerebrovascular disease mortality study. Clinical information or a family interview was obtained for 96.5 percent of the 16,956 death certificates in the adjusted sample. When all sources of stroke diagnosis were pooled together, it was found that 18.9 percent of the deaths had a stroke diagnosis. The frequency was highest in the high cerebrovascular disease death rate areas. Approximately 60 percent of all stroke diagnoses were listed on the death certificate including 42.2 percent as the underlying cause of death. Differences among the areas were comparatively small. The sensitivity of the death certificate was highest for hospital deaths and lowest for those certified by the medical examiner, while the specificity was the same irrespective of the place of death. Differences in the frequency in which clinical stroke diagnoses were 21 listed on the certificate did not explain the geo- graphic differences in death rates.” Reprinted with permission of the American Journal of Epidemiology. 65. Kuller, Lewis, H.; Bolker, Abraham; Saslaw, Milton S.; et al. (ACD-4-US-65) Nationwide Cerebrovascular Disease Mortality Study. III. Accuracy of the Clinical Diagnosis of Cerebro- vascular Disease AMERICAN JOURNAL OF EPIDEMIOLOGY 90(6): 556-566, 1969 (Author’s abstract) “The large differences in cerebrovascular disease mortality among geographic areas of the United States may be due to differences in the accuracy of the diagnosis of stroke. The accuracy of the diagnosis of stroke among areas of the United States with high, intermediate or low cerebrovascular mortality rates was studied by reviewing clinical records for every death included in the nationwide cerebrovascular disease mortality study. The frequency of most symptoms of stroke was similar among the areas. For stroke deaths in a hospital, hemiplegia was listed on 55 percent of the hospital charts and coma on 65.7 percent. Approxi- mately 85 percent of the hospital stroke deaths could be validated by either an autopsy, arteriogram, hemorrhagic spinal fluid, hemiplegia or coma on admission. There were no differences among the high, low and intermediate cerebrovascular disease mortal- ity areas. When stroke was the underlying cause on the death certificate, 69.5 percent of the clinical records from hospitals, physicians or medical examiners reported coma; 49.1 percent, hemiplegia; 25.8 percent, an autopsy of the brain; 33.4 percent, a spinal puncture; 7.9 percent, an arteriogram and 2.4 percent, a craniotomy. Differences in the accuracy of the diagnosis of stroke apparently did not account 22 for geographic variations in cerebrovascular disease mortality among areas of the United States.” Reprinted with permission of the American Journal of Epidemiology. 66. Kuller, Lewis H.; Lilienfeld, Abraham; and Fisher, Russell (ABCD-3-US-65) Sudden and Unexpected Deaths Due to Natural Causes in Adults — A Comparison of Deaths Certified and Not Certified by the Medical Examiner ARCHIVES OF ENVIRONMENTAL HEALTH 13: 236-242, 1966 (Author’s summary) “A study of sudden and unexpected nontraumatic deaths between June 15, 1964, and June 14, 1965, was completed in Balti- more. The deaths were studied by reviewing all avail- able medical information [for a stratified sample of death certificates for Baltimore residents aged 20-64] in order to determine: (1) whether the death was possibly sudden or not; and (2) the accuracy of the diagnosis as reported on the death certificate. “There were 1,857 deaths in the original sample, and 589 were sudden and unexpected; after adjusting for sampling it was estimated that 1,178 (32 percent) of 3,648 deaths were sudden. “The medical examiner certifies more sudden deaths in the younger age group, in Negroes, and in males. While 77 percent of unwitnessed sudden deaths are certified by the medical examiner, only 23.1 percent of those witnessed between 2 and 24 hours in the 40-64 year age group were certified. Sudden deaths not due to arteriosclerotic heart disease were more likely to have been certified by the medical examiner. Sudden deaths in the upper socioeconomic classes, and especially those deaths with a history of either a recent visit to a physician or heart disease, were usually not certified by the medical examiner. Therefore, adequate studies of sudden death must include both deaths certified and not certified by the medical examiner.” Reprinted with permission of Archives of Environmental Health. 67. Kuller, Lewis H.; Lilienfeld, Abraham; and Fisher, Russell (ABCD-3-US-65) Epidemiological Study of Sudden and Unexpected Deaths Due to Arteriosclerotic Heart Disease CIRCULATION 34:1056-1068, 1966 A sample of all nontraumatic deaths occurring from June 15, 1964 to June 14, 1965 of Baltimore residents aged 20-64 years was studied to determine the frequency and cause of sudden death in a defined population. Of the original 1857 deaths selected for study, 589 were classified as sudden and unexpected. For comparison, information was obtained on two additional groups; a probability sample of deaths of white male Baltimore residents aged 40-64 years and all arteriosclerotic heart disease (ASHD) deaths that were ot sudden. Hospital records, autopsy protocols, physician reports, and interviews with witnesses were studied to evaluate the accuracy of the cause-of-death diagnosis. Based on this evidence, ASHD was selected as the most likely cause of death in 489 deaths. Of these deaths, 92.6 percent were correctly assigned to ASHD on the death certificate, and the remaining cases listed ASHD as either immediate or contributing cause. An investigation was also made of (1) differ- ences in frequency and cause of sudden death be- tween racial and socioeconomic groups, (2) the relationship of ASHD sudden deaths to prior history of heart or cardiovascular disease and to prior medical treatment, and (3) the circumstances surrounding the death. 68. Kuller, Lewis H.; Lilienfeld, Abraham; and Fisher, Russell (ABCD-3-US-65) Quality of Death Certificate Diagnoses of Arterio- sclerotic Heart Disease PUBLIC HEALTH REPORTS 82(4):339-346, 1967 (Author’s summary) ‘“Nontraumatic deaths of Baltimore residents aged 20-64 years, occurring be- tween June 15, 1964 and June 14, 1965, were investigated. The accuracy of the diagnosis listed on the death certificate was determined by reviewing available medical information and from interviews of next-of-kin or other relatives or friends of the dead person. “A stratified sample of approximately 50 percent (1,857) of the total deaths was reviewed. In 553 (29.8 percent) of these 1,857 deaths arteriosclerotic heart disease was considered to be the principal cause of death. Of the 553 deaths due to arteriosclerotic heart disease 488 (88.4 percent) occurred within Baltimore City. In 452 (92.6 percent) arteriosclerotic heart disease (rubrics 420 and 422) had been con- sidered as the underlying cause of death and in the other 36 (7.4 percent) as either an immediate or contributing cause. Because of the rapidity of the events leading up to the deaths attributed to arterio- sclerotic heart disease, the accuracy of the diagnosis is often based only on a history of heart disease, sud- denness of the death, and the absence of other significant diseases.” 69. Kunitz, S. J. and Edland, J. F. (DS-9-US-70) The Epidemiology of Autopsies in Monroe County, New York JOURNAL OF FORENSIC SCIENCES 18(4):370- 379, 1973 In an effort to assess the validity of mortality statistics for Monroe County, New York, death cert- ificates were obtained for all county residents who died in the 11-year period 1960-70. Autopsy rates were calculated from the autopsy information recorded on the certificate, and the rates were ana- lyzed by age, sex, and race of the decedent, and by year, cause, and place of death. Autopsy rates were disproportionately higher for younger and for black persons. The authors feel that this is justified to some extent by the unnaturalness of these deaths; however, they recommend that more effort be spent in under- standing the chronic and degenerative diseases of the steadily growing population of elderly persons. 70. Lombard, Herbert L.; Huyck, Evelyn P.; and Snegireff, Leonid S. (CD-2-US-58) An Appraisal of the Cancer Death Record PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA 48(12):2059-2062, 1962 The authors cite four factors that make death record studies of cancer by site unsatisfactory: (1) errors in diagnosis, (2) recording of the site of metas- tasis rather than the primary site, (3) imprecise reporting of site, and (4) recording of the correct noncancer cause of death when cancer was also present. To study these problems, death records were obtained for 13,246 individuals who had attended Massachusetts cancer clinics from 1946 through 1958 and who died after 1948. The death certificates were compared with clinical records, and problems of 23 understatement and overstatement were studied for 10 cancer sites. The problems as they pertain to lung cancer are discussed in detail. The authors found that these errors were compensating to some extent, but were skeptical of the value of death records for epidemiological purposes. 71. Lombard, Herbert L. and Joslin, Elliott P. (CD-8-US-57) Underlying Causes of Death of 1,000 Patients with Diabetes NEW ENGLAND JOURNAL OF MEDICINE 259: 924-926, 1958 To investigate the change in diabetes mortality rates caused by the introduction of the sixth revision of the ICD, death certificates of 1,000 patients with diabetes diagnosed at the Joslin Clinic were studied. All patients had died in the period 1950-57. The findings were compared with findings of a previous study of the coding of 1,000 deaths of diabetics. Only 33 percent had diabetes as the cause of death under the sixth revision, a decrease from 66 percent. In the earlier study, 10 percent had diabetes men- tioned on the certificate but another cause was regis- tered, compared with 44 percent for the current study. Diabetes was not mentioned on the certificates of the remaining cases. 72. Lundberg, George D. and Voigt, Gerhard E. (A-9-SW-76) Reliability of a Presumptive Diagnosis in Sudden Unexpected Death in Adults: The Case for the Autopsy JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 242(21):2328-2330, 1979 One hundred cases of sudden unexpected death in adults (SUDA) were studied to determine the value of autopsy in assigning the cause of death for cases of SUDA. The study cases were 100 consecutively autopsied case of SUDA, occurring in Sweden in July or August, 1976, in which no cause of death was apparent. A distribution of the cases by autopsy cause of death is shown, followed by a discussion of the errors that might have arisen had the cause of death been assigned without autopsy. 24 73. Markush, Robert E. (BD-5-US-64) National Chronic Respiratory Disease Mortality Study I. Prevalence and Severity at Death of Chronic Res- piratory Diseases in the United States, 1963 JOURNAL OF CHRONIC DISEASES 21:129-141, 1968 (Author’s summary) “The certifiers of 3193 U.S. deaths, aged 35-74 were queried by mail on the presence and severity of several chronic respiratory diseases (excluding tuberculosis, lung cancer and occupational pneumoconioses), in order to evaluate published vital statistics which indicate a marked increased since 1950 in the U.S. death rate from these diseases. “The prevalence at death of the chronic respira- tory diseases was found to be more than nine times greater than its underlying cause death rate. It was also found that U.S. death certifiers in 1963 listed on the death certificates only half of the severe chronic respiratory disease present at death, suggesting that vital statistics based on death certificates may seri- ously underestimate the contribution of chronic respiratory diseases to U.S. deaths. Such epidemi- ologic observations on the chronic respiratory dis- eases as their rapidly increasing U.S. mortality rate, their relatively small contribution to mortality in the U.S. when compared to the U.K., and their relatively high mortality rates in several states could be statis- tical artifacts arising from variation in habits of death certification.” NOTE: The sample included all (837) deaths of white persons aged 35-74 years, from the 10 percent Current Mortality Sample for July through September 1963 with mention of chronic respiratory disease anywhere on the death certificate. The remaining 2,356 deaths were selected systematically (every 200th death) from the death certificates of white, U.S. native-born persons, aged 35-74 years from 12 States on file at the National Center for Health Statistics for January through March 1963 and January 1964. 74. Markush, Robert E.; Schaaf, William E.; and Seigel, Daniel G. (BD-9-US-62) The Influence of the Death Certifier on the Results of Epidemiologic Studies JOURNAL OF THE NATIONAL MEDICAL ASSO- CIATION 59(2):105-113, 1967 (Author’s summary) ‘‘This report investigates the bias introduced into epidemiological studies by differences in the distributions of selected character- istics of the certifiers of deaths. “It was found that a potential source of bias in studies of the effects of urbanicity may derive from differing proportions of certifiers who were board specialists inside and outside metropolitan areas. Similarly, cigarette smoking studies may be distorted by different lengths of certifier attendance and by different proportions of medicolegal deaths among smokers and nonsmokers. The alternative interpreta- tion that observed differences in the distributions of diagnoses among certifiers are the result of differences in the kinds of patients that they see is not dismissed. It is suggested that giving more attention to the characteristics of certifiers of deaths would increase the reliability of epidemiological studies of mortal- ity.” NOTE: The study group on which this pilot study was based consisted of a sample of 507 death certificates from 12 states obtained from the National Vital Statistics Division for the months of November and December 1962. The sample was restricted to white male decedents aged 35-74 years and white female decedents aged 30-74 years. A questionnaire was sent to the informant listed on each death certificate in order to obtain the decedent’s smoking history. Characteristics of the certifier of the death were obtained from the death certificates and from the 22nd edition, 1963, of the American Medical Directory, published by the American Medical Association. 75. Marshall, Thomas K. (AB-9-IR-67) The Value of the Necropsy in Ascertaining the True Cause of a Non-Criminal Death JOURNAL OF FORENSIC SCIENCES 15(1):28-33, 1970 (Author’s summary) ‘This paper reports the errors in diagnosing the cause and category of death, whether homicide, accident, suicide, or natural causes, likely to be introduced when a clinical assess- ment is relied upon instead of the cause and circum- stances of the death being confirmed by necropsy. An analysis of 1,000 consecutive coroner’s necropsies [performed between January 1966 and January 1967 in Northern Ireland] was carried out and the diag- nosis made from the history alone was found to be quite wrong in 11.3 percent cases when checked against the necropsy findings. This error was a min- imum one, likely to have been greater if the diagnoses had been made by someone inexperienced in forensic pathology. The significance of the misdiagnoses is commented on.” Reprinted with permission of the Journal of Forensic Sciences. 76. McCarthy, P. Desmond and (BD-7-IR-68) Walsh, Dermot Suicide in Dublin: I. The Under-reporting of Suicide and the Consequences for National Statistics BRITISH JOURNAL OF PSYCHIATRY 126:301- 308, 1975 (Author’s summary) “This study of suicide in Dublin during 1964-1968 from coroners’ records was undertaken to estimate the discrepancy between coroners’ verdicts, the national suicide statistics compiled from them and the clinical assessment of probability of suicide by psychiatrists examining the same records. The large difference in numbers of suicides deriving from the two approaches has con- siderable implications for national suicide statistics, and these have been briefly considered. From the findings presented we believe that we are justified in concluding that: (a) there are real differences in national suicide rates, at least between Ireland, England and Wales, and Scotland, and (b) the Irish suicide rate is low, though not as low as official statistics suggest, and (c) the discrepancy between official and ‘true’ suicide rates in Ireland is greater than in England and Wales and in Scotland.” Reprinted with permission of the British Journal of Psychiatry. 77. Mitchell, Roger S.; Maisel, John C.; Dart, Gladys A.; and Silvers, G. Wayne (AD-5-US-70) The Accuracy of the Death Certificate in Reporting Cause of Death in Adult: With Special Reference to Chronic Bronchitis and Emphysema AMERICAN REVIEW OF RESPIRATORY DISEASE 104:844-850, 1971 A study was made of 634 persons 40 years of age or over who died and were autopsied at either Colorado General or Denver Veterans Administration Hospital between September 1959 and May 1970. In 578 of the autopsies, special attention was given to the heart and lungs. The autopsy findings and clinical records were studied to determine the most likely cause of death for each study case. Underreporting of chronic bronchitis and emphysema was found to decrease over the 12-year study period. Also, there 25 was overreporting of these conditions as underlying cause of death when death had actually been caused by another condition. Fifty-six percent of the un- autopsied cases listed chronic bronchitis or emphy- sema on the death certificate, compared with an expected rate of 74 percent estimated from the autopsy study. These findings indicate that chronic airway obstruction is underreported as a cause of death. 78. Mitchell, Roger S.; Walker, Strother H.; Silvers, G. Wayne; Dart, Gladys A.; and Maisel, John C. (AD-5-US-66) The Causes of Death in Chronic Airway Obstruction. I. The Unreliability of Death Certificates and Routine Autopsies AMERICAN REVIEW OF RESPIRATORY DISEASE 98(4):601-610, 1968 To determine whether mortality statistics for chronic airway obstruction (CAO) are accurate, a study was made of 263 subjects with a history of dyspnea on exertion, a physical examination revealing generalized impairment of airflow, and, in most cases, physiologic evidence of diffuse airway obstruction. These subjects died between September 1959 and December 1966 in Colorado. Autopsies were per- formed on 173, with special studies of lungs done on 134 of the subjects from three Denver hospitals. Also included in the study were 253 autopsied subjects without clinical evidence of CAO and on whom special lung studies were done. It was determined that 101 of the 134 CAO patients on whom lung studies were performed died as a direct result of chronic airway obstruction. Only 79 of the 101 patients had CAO reported as a cause of death. In the other 33 patients in whom it was determined that CAO was not the underlying cause of death, three were re- ported as if CAO was the underlying cause. The authors compared autopsied and unautopsied patients with CAO and found that the rates for myocardial infarction, cerebrovascular accident, and arterioscle- rotic heart disease were inflated for the unautopsied patients. Comparison of the routine autopsy with the special lung studies showed important discrepancies in one-fourth of the cases. The authors conclude that there was an underreporting of the CAO on death certificates, causing the mortality rates for CAO to be underestimated. 26 79. Modan, Baruch; Sharon, Ezra; and Jelin, Navah (AC-3-IS-64) Factors Contributing to the Incorrect Diagnosis of Pulmonary Embolic Disease CHEST 62(4):388-393, 1972 (Author’s abstract) “A study of 2,107 consecutive patients on whom an autopsy had been performed in one major medical center during a 4-year period revealed 545 who had had either a clinical or patho- logic diagnosis or both of a pulmonary embolus. The frequency of a false negative diagnosis was 66.6 percent, and the frequency of false positives of all cases with a clinical diagnosis of pulmonary embolus was 61.9 percent. The frequencies were unrelated to age, sex, or ethnic origin, but were slightly lower among patients who died on the surgical ward. A variety of clinical factors and in particular the under- lying disorder, played a role in both false positive and false negative diagnoses. Lack of ECG and chest x-ray examinations increased the false negative rate, but had no effect on the false positive rate. The cases with false negative diagnoses differed from those with correct diagnoses in the location of the embolus and in a lower frequency as cause of death.” Reprinted with permission of Chest. 80. Moriyama, Iwao M. (D-9-ZZ-99) Comparison of Cause-of-Death Coding: Canada, England and Wales, United States of America World Health Organization report 1958. Unpublished. To investigate the variation among countries in the assignment and coding of the cause of death, three countries (Canada, England and Wales, and the United States) were asked to participate in the following study. Each country selected 1,000 death certificates, typical of that nation’s death records with respect to causes of death and the type of certificate that might be encountered in that country. Each country coded the 3,000 death certificates, using their standard procedures. The United States National Office of Vital Statistics submitted two sets of codes, one coded by using routine procedures and the second including a further check by a coding in- structor in consultation with the coding supervisor. A comparison of the four sets of codes indicated that coding errors were made regularly, but that the errors were compensating to some extent within groups of categories. Errors were found that would be corrected when the seventh revision of the ICD was introduced. Many of the errors resulted from different views on what constitutes a ‘‘highly improbable sequence” of causes of death being listed on the death certificate. The WHO proposed a further investigation to test the effect of the seventh revision of the ICD and the Rules of Selection. 81. Moriyama, Iwao M. (CD-3-US-56) Factors in Diagnosis and Classification of Deaths from CVR Diseases PUBLIC HEALTH REPORTS 75(3):189-195, 1960 A sample of 1,837 death certificates of patients dying in Pennsylvania during a 3-month period in 1956 were studied to determine the problems of diag- nosis, reporting, and classification of cardiovascular- renal (CVR) diseases. (For a summary of study methods, see No. 82 in this report.) Using supple- mentary information obtained from the certifying physician, the author found that for the 1,406 CVR deaths, 47 percent of the records had supporting diagnostic information that was sketchy in type and amount. In 80 percent of the 1,194 CVR deaths certified by a physician, the diagnoses were reason- able or solidly established. When a physician had decided on the diagnosis, it was found that the under- lying cause and contributory cause of death were correctly worded and ordered on 94 percent of the certificates. The author concludes that a universal adoption of diagnostic criteria would improve CVR mortality statistics, but that CVR mortality statistics were accurate enough to show general trends in mortality. 82. Moriyama, Iwao M.; (CD-9-US-56) Baum, William S.; Haenszel, William M.; and Mattison, Berwyn F. Inquiry into Diagnostic Evidence Supporting Medical Certifications of Death AMERICAN JOURNAL OF PUBLIC HEALTH 48 (10):1376-1387, 1958 A subsample of the 10 percent Current Mortality Sample (CMS) of deaths registered in the State of Pennsylvania, during a 3-month period in 1956, was used to ascertain the quality of cause-of-death certifi- cation. The subsample, which totaled 1,837 deaths, excluded deaths due to accidents, suicides, and homo- cides, included only a portion of deaths due to cardiovascular-renal causes, and was augmented by inclusion of all deaths from cancer of the respiratory system occurring during the study period. A follow- back study was conducted with the certifying physi- cian to obtain information on (1) diagnostic methods and findings on which the certified cause of death was based, (2) an expression of his certainty of the diagnosis, and (3) a revised medical certification if his opinion had changed. Coroners and medical exam- iners who certified deaths were queried in a similar manner, except that item (2) was excluded. For the 1,837 sample deaths, 4 percent had no support of the diagnosis because additional information was not obtained. In 38 percent the supporting data were sketchy. The remaining 58 percent of the diagnoses had very good or good support. The certifications were rated with respect to consistency with the diagnostic information reported, showing that for 79 percent the certified cause was the most probable diagnosis, for 13 percent another diagnosis was equally probable, and for 5 percent another diagnosis was preferred. 83. Moriyama, Iwao M.; (CD-3-US-60) Dawber, Thomas R.; and Kannel, William B. Evaluation of Diagnostic Information Supporting Medical Certification of Deaths from Cardiovascular Disease NATIONAL CANCER INSTITUTE MONOGRAPH No. 19: Epidemiological Approaches to the Study of Cancer and Other Chronic Diseases, 1966. pp. 405-419. (Author’s summary) “Questionnaires were sent to the medical certifiers and others with knowledge of the case on a national sample of 1,362 cardiovascular- renal disease deaths [occurring in July and August, 1960] to secure information on diagnostic methods used and pertinent findings on which the medical certification of death was based. Data were also obtained on sudden and unexpected deaths and presence or absence of associated diseases. The returns were reviewed and determination made as to whether the assigned cause of death was supported by diagnostic data provided by the certifier. “It is estimated that from 70 to 75 percent of deaths classified as cardiovascular disease in the United States may be considered as a reasonable inference or better. With a higher rate of usable response, this estimate would have been somewhat higher. About 10 percent of deaths from cardio- vascular diseases were sudden and unexpected deaths. “Large proportions of related diseases such as diabetes and certain chronic bronchopulmonary 27 diseases were not reported on death certificates as contributing to death.” 84. Murphy, Gordon K. (A-2-US-99) Cancer and the Coroner JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 237(8):786-788, 1977 Thirteen hundred consecutive forensic autopsies performed by the author in Dayton, Ohio and Balti- more, Maryland were reviewed (1,287 included brain examination), and 22 malignancies (1.7 percent) were found. Cancer was incidental in nine cases, the cause of sudden death in six, an unusual presentation in three, and associated with suicide in four. This review did not include coroner’s cases where cancer was known to exist. The author comments on the impor- tance of accurate autopsies in the medicolegal setting. 85. Murphy, Thomas (CDS-2-IR-64) Certification of Death from Cancer of the Uterus JOURNAL OF THE IRISH MEDICAL ASSOCIA- TION 60(364):385-388, 1967 Using the sixth revision ICD codes for uterine cancer — (171) malignant neoplasm of the cervix, (172) malignant neoplasm of the corpus uteri, (173) malignant neoplasm of other parts of the uterus including chorioepithelioma, and (174) malignant neoplasm of the uterus—unspecified—the authors present a frequency distribution, by coded cause, of women dying of uterine cancer in 1964 in Ireland. For 149 of the 157 deaths coded to ICD 171-174, the author collected additional information about the cause of death and reduced the number of cases in the ICD 174 category from 95 to 16. Most were reclassified as cancer of the cervix or corpus. The ratio of cancer of the cervix to corpus uteri was reduced significantly as a result. The author concludes that if followup were instituted for all causes of death, the resultant mortality statistics would be more accurate. 86. Murphy, Thomas and Joyce, Nessa (S-2-IR-68) Certification of Death from Cancer of the Uterus JOURNAL OF THE IRISH MEDICAL ASSOCIA- TION 63(396):235-236, 1970 In a review of cancer of the uterus in Ireland, the author presents tables showing a decline in the use 28 of the indefinite category ‘Malignant neoplasm of the uterus, unspecified” between 1961 and 1968. Additional tables are presented that show similar patterns of certification for other indefinite. categories of the ICD selected from codes for Arteriosclerotic and Degenerative Heart Disease (420-422) and Symptoms, Senility, and Ill-defined Conditions (730- 795). The authors conclude that accuracy in the certification of death has improved but that even greater accuracy is needed. 87. Najem, G. Reza; Riley, Harris D., Jr.; and Najem, Leila I. (CD-3-US-70) Reliability of Heart Disease Diagnoses JOURNAL OF THE OKLAHOMA STATE MEDICAL ASSOCIATION 68(12):452-457, 1975 A study of the reliability of heart disease diag- noses was made using death certificates for Oklahoma City for 1969 and 1970. The study sample included all certificates listing hypertensive and chronic rheumatic heart disease as immediate cause of death, and 10 percent of those listing ischemic heart disease as immediate cause. For 150 of the 159 selected cases, clinical information was collected and evalu- ated using predetermined clinical criteria. When sudden deaths were excluded, the reliability of the cause of death was found to be 71 percent for chronic rheumatic heart disease, 49 percent for is- chemic heart disease, and 47 percent for hyper- tensive heart disease. The reliability of heart disease diagnoses was shown to depend on sex, type of medical center where patient died, race, and marital status. The authors concluded that death certificate information must be refined and adjusted for the proportion of incorrect assignments of cause of death before the data can be used for research purposes. 88. Nelson, Franklyn L.; Farberow, Norman L.; and MacKinnon, Douglas R. (B-7-US-75) The Certification of Suicide in Eleven Western States: An Inquiry into the Validity of Reported Suicide Rates SUICIDE AND LIFE-THREATENING BEHAVIOR 8(2):75-88, 1978, Human Sciences Press (Author’s abstract) “From Durkheim’s time to the present social researchers interested in the problem of suicide have relied upon officially reported rates of suicide to develop and test their theories. Despite the fact that the validity of any theory rests upon the accuracy of its underlying data, the relative accuracy of reported suicide rates have rarely been questioned or systematically evaluated. This paper investigates the process of death certification as practiced by a sample of 191 coroners in 11 western states. Findings indicate extensive variation in the backgrounds, pro- fessional resources, operating procedures, and govern- ing statutes of coroners and coroners’ offices and in policies concerning the use of the suicide mode. Since the coroner is generally charged with the official re- sponsibility for certifying the mode of death when unnatural mode is suspect, the extent of variation found here calls into question the validity and com- parability of reported suicide rates.” Reprinted with permission of Suicide and Life-Threaten- ing Behavior. 89. Newhouse, M. L. and Wagner, J. C. (ACD-2-UK-64) Validation of Death Certificates in Asbestos Workers BRITISH JOURNAL OF INDUSTRIAL MEDICINE 26:302-307, 1969 A study to validate certified causes of death was made of 158 asbestos factory workers who died before 1964 and for whom necropsy reports were available. In 84 of the cases, histological material also was obtained. Four additional cases of carcinoma of the bronchus were identified in the group for which only necropsy reports were available, and another four cases of carcinoma of the bronchus and 15 mesotheliomata were discovered in the group for which histological material also was reviewed. Some degree of asbestosis was found in 60 of the 67 cases for which lung sections were available. 90. Padley, Richard (DS-9-CY-56) Cause-of-Death Statements in Ceylon: A Study in Levels of Diagnostic Reporting BULLETIN OF THE WORLD HEALTH ORGANIZ- ATION 20:677-695, 1959 (Author’s abstract) “The author presents a critical discussion of the current system of registra- tion of deaths in Ceylon, pointing out the difficulties inherent in collecting and analysing data relating to causes of death in a country where many of the cause-of-death statements are made by lay registars whose knowledge of modern medical terminology is slight or non-existent. He suggests various ways in which the present system might be improved, stress- ing particularly the need for continuous scrutiny of the Sinhalese or Tamil terms commonly used in reporting causes of death in rural areas and for the preparation of simple instructions to lay registrars regarding a system of priorities to be followed in selecting the symptoms to report as a result of in- terrogation of the relatives notifying the death.” Reprinted with permission of the Bulletin of the World Health Organization. 91. Penttild, Antti and Ahonen, Antti (DS-3-FN-68) Arteriosclerotic and Other Degenerative Heart Diseases in Finland. I. A Death Certificate Study of the Frequency of Degenerative Heart Diseases among Males and Females SCANDINAVIAN JOURNAL OF SOCIAL MEDI- CINE 3(2):61-67, 1975 (Author’s abstract) ‘All available information recorded on the death certificates of 12,973 Finnish persons who, according to the official Finnish mor- tality statistics, died in 1968 from arteriosclerotic and other degenerative heart diseases (ADHD, rubrics 420-422 in ICD) comprised the material of the pres- ent study. The mortality of males from ADHD analysed by age and place of residence was very high when compared with various national rates of inter- national WHO statistics. The degree of urbanization of the domicile did not have any statistically signifi- cant effect on the mortality of ADHD. Significant differences between various provinces were found in the mortality of males from ADHD. The male popula- tion living in the eastern provinces of Finland showed a highly significantly higher mortality from degenera- tive heart diseases than the male population living on the west coast. A highly significant difference was found in mortality between various subgroups of the Finnish male and female populations analysed by age, place of residence, and type of community. The uni- form difference between the mortality of various male and female subgroups of the Finnish population, which was obtained using the present statistical survey of death certificates, and the fairly uniform distribution of high rate of mortality of males from degenerative heart diseases in most regions of the country lend further support to the reliability of cause-of-death statistics, since certification of deaths can then be regarded to occur uniformly and with about the same accuracy in different parts of the country.” Reprinted with permission of the Scandinavian Journal of Social Medicine. 29 92. Penittila, Antti and Ahonen, Antti (DS-3-FN-68) Arteriosclerotic and Other Degenerative Heart Diseases in Finland. II. A Death Certificate Study of the Examination of the Cause of Death from Degenera- tive Heart Diseases SCANDINAVIAN JOURNAL OF SOCIAL MEDI- CINE 3(2):69-74, 1975 (Author’s abstract) “A statistical survey of death certificates was made to analyse the ante-mortem and post-mortem medical and medico-legal examinations used in the determination of the cause of death of 12,973 decedents who were recorded officially to have died of arteriosclerotic and other degenerative heart diseases in Finland in 1968. The relationship between the regional autopsy rate and the rate of mortality from degenerative heart diseases was studied in particular. The survey indicated that there was no systematic relationship between the type of ante-mortem and post-mortem cause-of-death exam- inations, including medical and medico-legal autopsies, and the rate of mortality from arteriosclerotic and other degenerative heart diseases in various groups of the Finnish population analysed by age, sex and dom- ‘icile. This was concluded to be an indication of the reliability of Finnish cause-of-death statistics of degenerative heart diseases which show a generally high rate of mortality and prominent regional differ- ences in the rate of deaths from those diseases among the Finnish male population.” Reprinted with permission of the Scandinavian Journal of Social Medicine. 93. Peterson, D. R. (BCD-9-US-73) Final Contract Report to the National Institutes of Health: Cardiovascular and Respiratory Disease Mortality Contract Report, NIH-NHLI-72-2952(C), Unpublished To test the utility of mail questionnaires for ob- taining additional information on deaths due to heart and lung causes, the National Heart and Lung Institute conducted a feasibility study based on a sample of 928 deaths of residents 35-74 years of age in Wash- ington State between October 1, 1972 and October 1, 1973. Deaths due to cancer, accident, poisoning, or violence were excluded. Death record informants, physicians, funeral directors, and hospitals were queried by mail and by personal interview to obtain information on medical attention in the year prior to death, functional limitation in the same period, 30 suddenness of death, and place of death. Tables are included that show completion rates and agreement in reporting of specific items by various sources. The quality of the new information obtained was ques- tionable since the death certificate information agreed with data from supplementary sources only 50-80 percent of the time. The author concludes that the death certificate followback study is a relatively expensive method of obtaining information that is of dubious quality. 94. Preston, Samuel H.; Keyfitz, Nathan; and Schoen, Robert (S-9-ZZ-64) Causesof Death: Life Tables for National Populations, Chapter III: Accuracy and Comparability Seminar Press, New York and London, 1972 This book presents mortality data from 180 pop- ulations representing 48 nations by age and sex for all years from 1861 to 1964 for which both (1) numbers of deaths by cause, age, and sex, and (2) numbers of persons alive by age and sex were available. The authors discuss classification of cause of death and derive life-table parameters. In Section IIIB the authors compare mortality statistics by country and point out specific problems associated with different cause-of-death categories. Their treatment focuses on the declining proportions of deaths assigned to ill-defined and unknown causes, and the consequent increased proportion of assignment to well-defined diseases. However, the ill-defined classification still accounts for a large proportion of deaths in statis- tically poor countries. These phenomena made comparison between countries and comparisons over time difficult. The authors also summarize the major problems of accuracy and comparability for each major disease category. 95. Puffer, Ruth Rice (ABCD-9-2Z-64) Study of Multiple Causes of Death Pan American Health Organization Report, Un- published In analyzing the data from the Inter-American Investigation of Mortality, death certificates, autopsy reports, and clinical records were compared for all (3,506) study deaths from San Francisco, California and Bristol, England for which clinical and autopsy records were available. (For further description of the study, see No. 96 in this report.) In San Francisco, 94.0 percent of the underlying causes of death on the death certificate were con- sidered present on autopsy, 4.8 percent were not stated on the autopsy report but were given on clinical records, and 1.2 percent were not present on either autopsy or clinical records. The percentages for Bristol were 81.9, 9.2 and 8.9, respectively. The underlying cause stated on the death certif- icate was also compared with the final assignment of cause based on all available information collected in the investigation. The final assignment of cause was not given on the death certificate for 12.5 percent of the Bristol deaths and 7.9 percent of the San Francisco deaths. All comparisons were based on 232 groups of causes of death; that is, the underlying cause on the certificate was considered confirmed if the cause of death was finally assigned to any cause in the same group. 96. Puffer, Ruth Rice and Griffith, G. Wynne (ABCD-9-ZZ-64) Patterns of Urban Mortality — Report of the Inter- American Investigation of Mortality: Chapter XV, Changes in Assignments of Causes of Death PAN AMERICAN HEALTH ORGANIZATION, SCI- ENTIFIC PUBLICATION No. 151, 1967 In the Pan American Health Organization study of geographic variations in mortality, 12 cities in the United States, Central and South America, and the United Kingdom were selected for investigation. A systematic sample of deaths from the death registry in the age group 15-74 years was selected in each city for further study. The final sample included 43,298 deaths that occurred during 1962-64. Additional in- formation was collected on clinical examinations, X-rays, laboratory examinations, biopsies, autopsies, and household interviews. The assembled histories were reviewed and recoded as to cause, thereby providing a measure of the reliability that can be attached to death certificate information. The 3,865 deaths in San Francisco and the 4,262 deaths in Bristol are most relevant to mortality statistics in the United States. Using a broad grouping of 74 causes, 26 percent of the deaths in San Francisco and 22 percent in Bristol showed a change in classification with the more definite information. For total respi- ratory disease cases in Bristol per 100 final assign- ments, 36 percent were excluded and 33 percent added. Over all 12 cities, 41 of every 100 final assignments to diseases of the digestive system were not originally in the same division of the 74-cause list; 23 of the 41 were transfers from other digestive dis- eases; 18 were transfers from outside the digestive disease group. The same type of result is presented for cardiovascular diseases, infective and parasitic diseases, maternal causes, malignant neoplasms, accidents and violence, and all other causes. The authors conclude that medical certification can be improved and point out the desirability of devising record linkages to insure the use of avail- able hospital and autopsy information. 97. Puffer, Ruth Rice and Serrano, Carlos V. (ABCD-9-ZZ-71) Patterns of Mortality in Childhood—Report of the Inter-American Investigation of Mortality in Child- hood: Chapter XVII, Changes in Assignments of Causes of Death PAN AMERICAN HEALTH ORGANIZATION, SCI- ENTIFIC PUBLICATION No. 262, 1973 The PAHO Inter-American Investigation of Mor- tality in Childhood, in which 35,095 deaths of chil- dren under 5 years of age were studied, was conducted in 15 areas of North, Central, and South America. The objective was to establish death rates for these areas that would be as accurate and comparable as possible, taking into account biological as well as nutritional, sociological, and environmental factors. The investigation was conducted in Latin America in 1969-71, in California during 1969-70, and in Quebec Province, Canada during 1970-71. Data from house- hold interviews, clinical records, death certificates, and autopsy records were studied to determine the true cause and circumstances of death. Clinical and other information in addition to the death certificate was available for 27,082 deaths from 14 projects. For slightly more than one-half of the deaths (52.5 percent) the final assignment of underlying cause was in agreement with the underlying cause derived from the death certificate. In nearly one-third of the deaths the underlying cause assigned on the death certificate was considered an associated cause on final assign- ment. Agreement rates are shown by project and by age at death. For a common communicable disease of childhood such as measles, only 55.4 percent of the deaths could have been known from death certifi- cates, and in several projects the numbers of measle cases were more than doubled in the final assignment. Specific types of nutritional deficiency were also missed; a fivefold increase was noted for the more severe form, kwashiorkor. 31 98. Quinn, Robert W.; Sprague, Homer A.; and Quinn, Julia P. (ACD-3-US-65) Mortality Rates for Rheumatic Fever and Rheumatic Heart Disease, 1940-65 PUBLIC HEALTH REPORTS 85(12):1091-1101, 1970 Over the period 1940-65, death rates due to rheumatic fever (RF) and rheumatic heart disease (RHD) decreased markedly. To determine if the trend was real or artificial, all deaths due to RHD and RF during the years 1940-65 in Nashville and David- son County, Tennessee were selected for study. The underlying cause of death was determined by the ICD fifth revision for 1940-48, the sixth revision for 1949- 58, and the seventh revision for 1959-65. Autopsy results and clinical records were obtained whenever possible, and after comprehensive evaluation of all available information, each underlying cause was either verified or reclassified. Only 61 percent of the causes of death due to RF or RHD were validated. The authors computed new death rates using the validated data rates and found only a slight decreasing trend during the period. Age, sex, and race differ- ences in the rates were shown to diminish. The. authors conclude that the downward trend in the RF and RHD death rates was an artificial decline brought about by the ICD changes in assignment of the under- lying cause of death and by better diagnostic proce- dures for RF and RHD. 99. Ransom, Don (DS-9-US-75) Quality and Accuracy of Vital Statistics at the Okla- homa State Department of Health Unpublished paper A discussion of the procedures for coding and validating vital records in Oklahoma is presented, along with estimated error rates for cause of death codes and several demographic items on the vital records. 100. Reid, D. D. and Rose, G. A. (ACD-9-ZZ-58) Assessing the Comparability of Mortality Statistics BRITISH MEDICAL JOURNAL 2:1437-1439, 1964 The reported death rates from coronary disease and bronchitis of middle-aged men are quite different for the United States, the United Kingdom, and 32 Norway. The authors proposed and pretested a meth- od of determining to what extent the mortality data only reflect international differences in diagnostic habits and concepts about the classification of various forms of cardiorespiratory disorder. Ten deaths were randomly selected from all deaths in two London hospitals in 1958 of males 40-64 years of age where death was certified as being due to one of a number of cardiovascular, renal, and respiratory diseases. All relevant clinical and autopsy information was ab- stracted and sent to doctors in Boston (24 doctors), Norway (16), and the United Kingdom (30) who determined the causes of death. Returns were then coded according to the ICD at the General Register Office in London. The classification of the cases into five broad categories was very close; however, within these groups, patterns of classifying deaths appeared which were in line with the pattern of prevailing mortality rates from these diseases in the three countries. 101. Riccitelli, M. L. (S-3-US-99) Myocardial Disease in the Aged, Including a Review of the Literature JOURNAL OF THE AMERICAN GERIATICS SO- CIETY 14(4):366-379, 1966 The author discusses the problems of measuring the frequency of clinical myocarditis in the living population and notes that even the pathologist is unable to make this diagnosis without histological examination of the gross specimen. Data are presented from literature showing diseases associated with myocarditis and the types of lesions found. Diagno- sis, treatment, and prognosis of myocarditis are summarized. The author concludes that the literature review indicates that myocarditis occurs more frequently than clinicians realize, and it is associated with a wide variety of diseases, and that the diagnosis of myocard- itis is seldom entered on either a hospital chart or a death certificate. 102. Rigdon, R. H. and Kirchoff, Helen (AD-9-US-61) Vital Statistics and the Frequency of Disease TEXAS STATE JOURNAL OF MEDICINE 59: 317-324, 1963 To examine the accuracy of cause of death statistics the authors selected 294 autopsy records for study; 100 cases routinely reviewed by the senior author during 1959-61 in a general teaching hospital in Galveston and 194 cases of malignancies from the autopsy records made between 1945 and 1958 from a Houston hospital which treats only malignancies. The clinical diagnosis, the diagnosis coded for vital statistics, and the post-mortem diagnosis were com- pared for those 294 cases. For 20 percent of the cancer cases and 32 percent of the general hospital cases, the diagnoses coded for vital statistics were different from the death certificate or post-mortem diagnoses. Lack of information or errors in filling out the death certificate produced many of the discrepancies. The authors conclude that the fre- quency of diseases as shown by vital statistics, at best, is only a crude index of the frequency of disease. 103. Rose, Geoffrey (S-3-UK-64) Errors in the Classification of Fatal Pericarditis THE LANCET II: 851, 1966 The death rates for rheumatic fever and active rheumatic pericarditis in young people have declined rapidly in the period 1950-64. For adults a similar, but less rapid, decline has been observed for rheumatic fever. However, the death rate for active rheumatic pericarditis in adults has remained constant. The author believes that deaths in adults currently classified as acute rheumatic pericarditis (ICD 401.0) are probably not rheumatic. The confusion may result from including in this category acute peri- carditis, not otherwise specified. It is recommended that, in future revisions of the ICD, such deaths should be coded separately because they may be responsible for as much as a threefold inflation of the total mortality attributed to rheumatic fever. 104. Rosenblatt, Milton B.; Teng, Peter K.; and Kerpe, Stase (AC-2-US-71) Diagnostic Accuracy in Cancer as Determined by Post Mortem Examination PROGRESS IN CLINICAL CANCER 5:71-81, 1973 A comparison of clinical and post-mortem diag- noses was made for all patients (1,000) autopsied in Doctors’ Hospital in New York during the period January 1, 1960 through February 6, 1971. During this period, there were 1,216 cases with clinical diag- noses of malignant disease and 493 cases with autopsy diagnoses of malignancy. The overall autopsy rate was 36 percent. For the common neoplasms defined clinically, most of the diagnoses were confirmed by autopsy (95-100 percent), except for carcinoma of the pancreas (80 percent) and for lung cancer where less than half of the cases were confirmed by autopsy. While the confirmation rate was high for most clin- ically detected carcinomas, the specific site of origin had been missed in a considerable number of cases. Between one-third and one-half of the carcinomas of the colon, pancreas, stomach, and ovary detected at autopsy were not suspected during life. Particular problems in the clinical diagnosis of primary lung cancer were discussed, explaining the marked dispar- ity between cases established clinically and those found at autopsy. 105. Rosenblatt, Milton B.; Teng, Peter K.; Kerpe, Stase; and Beck, Irene (AC-9-US-71) Causes of Death in 1,000 Consecutive Autopsies NEW YORK STATE JOURNAL OF MEDICINE 71: 2189-2193, 1971 One thousand consecutive autopsies conducted at Doctor’s Hospital in New York from January 1, 1960 through February 6, 1971 were studied to obtain a more accurate perspective on the current causes of death than that generally obtained from clinical data. Autopsy confirmation of 493 clinically diagnosed neoplasms ranged between 80 and 100 percent for carcinomas of the colon, breast, pancreas, stomach, and ovary, but many additional cases were diagnosed at autopsy which had not been suspected clinically. Carcinoma of the lung was the only neoplasm that was greatly overdiagnosed clinically, and no unsus- pected cases were found at autopsy. The distribution of autopsy diagnoses for the remaining 507 patients with nonmalignant causes of death is presented. 106. Ross, Olivia and Kreitman, Norman (DS-7-UK-99) A Further Investigation of Differences in the Suicide Rates of England and Wales and of Scotland BRITISH JOURNAL OF PSYCHIATRY 127:575- 582, 1975 (Author’s summary) “National samples of case records of suicidal-type deaths from England and Wales and from Scotland were reassessed by officials in the other country. It emerged that similar criteria for suicide existed in both countries, and that there was no age-related tendency to misclassify cases. The lower official suicide rate amongst the old in Scotland 33 was therefore considered not to result from ascertain- ment differences. It was also concluded that Scottish records were not so briefly documented as to prevent the conclusive ascertainment of cause by England and Wales coroners. Cases which were designated ‘un- determined’ in Scotland tended to be classified ‘accidental’ by coroners. Reasons for the lower inci- dence of suicide in Scotland are discussed.” Reprinted with permission of the British Journal of Psychiatry. 107. Rossman, Isadore (DS-3-US-73) True Incidence of Pulmonary Embolization and Vital Statistics JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 230(12):1677-1679, 1974 To examine the discrepancy between autopsy incidence and vital statistics reporting of fatal pul- monary embolization, all of the coded New York City death certificates for April 1973 (6,385) were selected for study. Pulmonary embolus was listed as a major or contributing cause on 207 death certifi- cates. Of the 182 deaths in which death was ascribed directly to pulmonary embolus, only one-sixth was coded to this cause for vital statistics purposes. The remaining cases were coded most often to cancer, chronic heart disease, and acute myocardial infarct. Furthermore, there is substantial evidence from past studies that pulmonary embolization is clinically unrecognized and, therefore, underreported as a cause of death in the absence of autopsy. 108. Rossman, Isadore; (AC-9-US-71) Rodstein, Manuel; and Bornstein, Alfred Undiagnosed Diseases in an Aging Population: Pul- monary Embolism and Bronchopneumonia ARCHIVES OF INTERNAL MEDICINE 133:366- 369, 1974 To study the accuracy of cause-of-death state- ments for the elderly, chronically ill, institutionalized population, a study was conducted of 250 consecu- tively autopsied patients who died between 1966 and 1971 in the Kingsbridge Division of the Jewish Home and Hospital for Aged (JHHA). Comparisons are made between the autopsy and clinical diagnoses with emphasis on primary cause of death, associated conditions, and premortem diagnoses. Accuracy of diagnosis was highest for ischemic heart disease and for cerebral vascular accidents (both 100 percent) and lowest for pulmonary embolism. The six chief causes 34 of death in the JHHA group were contrasted with those in a comparison group, all over 80 years of age who were autopsied in New York City general hospi- tals in 1970. The JHHA group showed death rates of 22.8 percent for bronchopneumonia and 6.4 percent for pulmonary embolism, compared with rates of 8 percent and 2 percent, respectively, for these condi- tions in the general hospital group. Comments are included on the diagnosis and treatment of these two conditions. 109. Samuelson, Wayne M.; Williams, Roger R.; and Maness, A. Timothy (CD-3-US-75) Accuracy of Death Certificates in Utah for Myo- cardial Infarction University of Utah College of Medicine, 1979, Un- published (Author’s abstract) “From a computer file of 37,387 Utah deaths attributed to myocardial infarc- tion during the years 1956-1975, a sample of 387 in- hospital deaths were assessed for verification of the death certificate diagnosis. Hospital charts were reviewed for 387 persons who died in 32 Utah hospitals during 1956-1975 and whose death certi- ficates listed acute myocardial infarction as the underlying cause of death. Charts were rated accord- ing to objective data present, and a composite ‘verifi- cation score’ was used to test for differences between subgroups. “Objective data in hospital charts supported the diagnosis of MI for 81 percent of the corresponding death certificates. Deaths attributed to myocardial infarction were better supported in males than in females, in larger hospitals than in smaller hospitals, and in the years 1966-1975 than in the years 1956- 1965. Somewhat better verification was observed among younger MI deaths, but this difference lacked statistical significance.” 110. Sauer, Herbert, I. (S-9-US-72) Cause Specific Death Rates as a Measure of Need Presented at the National Center for Health Statis- tics Data Use Conference, Dallas, Texas, March 1977, Unpublished (Author’s summary) ‘Individuals in various settings have a tendency to draw conclusions from data as if the data collection, tabulation and analysis were done perfectly, or else go to the other extreme of assuming data to be worthless. Each extreme is usually an untenable position. When any of the categories containing ill-defined deaths (especially ‘Symptoms and ill-defined, ICDA 780-796," ‘Other heart disease, ICDA 420-429, ‘Malignant neoplasm, site secondary or unspecified, ICDA 195-199, and ‘Neoplasm, benign or not specified ICDA 210-239’) is substantially in excess of the U.S. rate, then it is appropriate to question whether the counts for specific causes are reasonably complete. When the rates for these ill-defined (or ‘wastebasket’) categories are low, there is less opportunity for specific cause categories/rates to be incomplete. “Specific cause categories may also be incomplete because, in a specific area, there may be a tendency to give greater weight to the role of one cause, and in another area to give more weight to another cause. Available evidence suggests that generally these shortcomings do not present serious limitations in the data. However, it seems both feasible and desir- able, before using death data, to consider the evi- dence in support of a hypothesis of limited data accuracy. “The count of the number of deaths and crude (all ages) death rates have value for program planning and evaluation, even when they are a function of the age distribution of the population. For more intensive planning and evaluation purposes, age-sex-race- specific death rates for chronic diseases are also likely to be needed.” 111. Sauer, Herbert I. and Enterline, Philip E. (S-3-US-51) Are Geographic Variations in Death Rates for the Cardiovascular Diseases Real? JOURNAL OF CHRONIC DISEASES 10(6):513-524, 1959 The authors present 1949-51 United States mor- tality data by State for white males aged 45-64 years for all causes and for cardiovascular diseases. Various aspects of death rates are analyzed, including the magnitude of differences between States, the corre- lation between the death rates and physician supply, and the correlation between cardiovascular and non- cardiovascular disease death rates and between coronary and other cardiovascular disease death rates. The authors conclude that considerable geographic differences do exist between States. 112. Schoenberg, Bruce S. and (CD-4-US-65) Powell, James Meyers, Jr. Statistics on Stroke: A Pilot Study of the Clinical Evi- dence Justifying the Reporting of Stroke on Death Certificates in Alameda County, California CALIFORNIA MEDICINE 109(1):19-23, 1968 In April 1965, 94 certificates of death in Alameda County had stroke (rubrics 330-334) listed as the immediate, contributory, or underlying cause of death. Because of the necessity for accurate informa- tion in a short period of time concerning those deaths, only patients who died in hospitals or nursing homes and who had clinical records present were further studied. Each case was matched to a control death on the basis of age, sex, race, and place of death. Con- trols were chosen from the deaths classified to the rubric 420-459, because it was felt that most patients with missed diagnoses would be in this category. Twenty-two percent (11) of the hospital deaths that had no mention of stroke on the certificate were found to have had a stroke by the. authors’ criteria (false negative), while 20 percent (3) of the control group’s nursing home deaths were false negatives. The false-positive percent of the nursing home cases (47 percent), on the other hand, was much higher than the false-positive percent for the cases who died in hospitals (2 percent). False positives were de- fined as cases where the clinical data failed to support the death certificate diagnosis of stroke. It was con- cluded that if immediate and contributory causes of death were included in the death rate tabulations, then the resulting stroke rates would be higher and might reflect more accurately the true experience of stroke cases in the United States. 113. Schwartz, Charles J. (CD-9-US-74) The Hawaii Mortality Follow-back Study: 2. An Evaluation of Medical Certification of the Cause of Death Through the Use of Hospital Discharge Diag- noses HAWAII STATE DEPARTMENT, R AND S REPORT ISSUE NO. 17, RESEARCH AND STATISTICS OFFICE, 1977 As part of the Hawaii followback study, estimates were made of the errors in certified causes of death. 35 (See Number 19 in this report.) The study included 715 deaths of Oahu residents dying in Honolulu from February 1972 through January 1974 for whom a hospital record was available. Deaths due to homi- cide, suicide, accidents, or cancer were excluded. Using the hospital diagnosis as the standard, the author computed the error rates of death certificate classification. Significant differences were found between the two cause-of-death classifications, the largest errors being in the categories of hyperten- sion and hypertensive disease, chronic ischemic heart disease, and other heart disease. The gross error in classification of cause of death on the death certificate was computed to be 20 percent. Agreement of the death certificate diagnosis with the hospital diagnosis varied considerably by disease. Reliability rates for various classifications were derived and possible sources of error were identified. 114. Shipley, Paul W.; Norris, Frank D.; Wray, Jo Ann; Alberton, Paul G.; and De Fisher, Duke (BD-9-US-62) Final Report of California Medical Certification of Death Study California Department of Public Health, 1964, final grant report, Unpublished. In developing protocols for a cause-of-death validation study, two projects were carried out; (1) California death certificates for the years 1952 and 1962 were studied, and (2) a pretest was conducted with 42 physicians concerning their understanding of certification procedures. The death certificate study yielded data on the following topics: autopsy rates, characteristics of autopsied cases, characteristics of the certifier, extent of use of coding rules in selecting underlying cause, apparent changes in coding cardiovascular renal disease by the Los Angeles County Coroner’s Office, and the use of affidavits to correct medical informa- tion on the death certificate. The pretest interviews conducted with 29 physi- cians in the bay area and 13 physicians in Los Angeles County covered such topics as the physician’s under- standing of the concepts and procedures involved in completing death certificates, his background and training, and suggestions on instructional material that would be useful. Physicians reported consider- able difficulty with the sequential aspect of the cause- of-death statement and preferred simply to list sign- ificant conditions present at death. A need was indicated for standardized instruction in filling out death certificates. 36 115. Skyring, A.; Modan, B.; Crocetti, A.; and Hammerstrom, C. (ACD-3-US-57) Some Epidemiological and Familial Aspects of Coro- nary Heart Disease: Report of a Pilot Study JOURNAL OF CHRONIC DISEASES 16:1267-1279, 1963 A list of all deaths of persons under 46 years of age classified as caused by arteriosclerotic heart dis- ease, chronic endocarditis not specified as rheumatic and myocardial degeneration (International Statistical Classification of Causes of Death Rubric Nos. 420, 421, and 422, respectively) for the years 1954-57 was obtained from the Baltimore City Health Department. Death certificates were obtained and an attempt was made to verify the cause-of-death statement by a review of hospital records, autopsy records in hospitals and in medical examiners’ offices and private physi- cians’ records. Criteria were specified for classifying the deaths as proved cases of coronary occlusion and myocardial infarction, unproved but probable cases, and definitely not coronary occlusion, myo- cardial infarction, or both. A table is presented showing that 16 of 324 deaths ascribed to arteriosclerotic heart disease were not coronary artery disease. None of the nine deaths certified as chronic endocarditis not specified as rheumatic were proved or probable cases of coronary heart disease, but 26 of the 80 deaths certified as resulting from myocardial degeneration definitely had coronary heart disease and an additional 18 were designated as probable. The authors suggest that the latter findings in- dicate the need for searching this category for deaths from myocardial infarction in future studies. 116. Spain, David M.; Bradess, Victoria A.; and Mohr, Charles (A-9-US-59) Coronary Atherosclerosis as a Cause of Unexpected and Unexplained Death: An Autopsy Study from 1949-1959 JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 174(4):384-388, 1960 All autopsies of persons over 30 years of age per- formed between 1949 and 1959 in Westchester County, New York, were studied to determine the frequency with which coronary atherosclerotic heart disease appeared as a cause of sudden unexplained and unexpected death in adults. These 1,329 deaths were classified into three groups according to the duration of the fatal episode: (1) less than 1 hour, (2) between 1 and 3 hours, and (3) duration un- known (unwitnessed). Coronary atherosclerotic heart disease was the cause of death for 91 percent of the 463 white males and for 48 percent of the white females in group one. For the unwitnessed deaths (group 3), 61 percent of the white males and 35 percent of the white females died from coronary artery disease. The frequency of spontaneous intra- cerebral or subarachnoid hemorrhage as cause of death was also studied. The author points out that assignment of a cause of death without an autopsy in cases of sudden unexplained or unexpected deaths in adults can seriously distort the vital statistics. 117. Steer, Arthur; (AD-9-JP-70) Land, Charles E.; Moriyama, Iwao M.; Yamamoto, Tsutomu; Asano, Masahide; and Sanefuji, Hayato Accuracy of Diagnosis of Cancer Among Autopsy Cases: JNIH-ABCC Population for Hiroshima and Nagasaki GANN JOURNAL 67(5):625-632, 1976 (Author’s abstract) “The accuracy of death cer- tificate diagnoses of cancer in the fixed population of about 100,000 samples in Hiroshima and Nagasaki was determined for the period 1961-1970 by com- parison with autopsy findings. In general, when the death certificate listed cancer as a cause of death it was found at autopsy in a high proportion of cases. However, cancer was not always reported on death certificates, indicating that cancer occurs more frequently than recorded by official mortality statis- tics. Older persons, persons who die at home, and persons with certain cancers are more likely not to have cancer named on their death certificates. It is estimated that in the 10,749 deaths occurring at home or in hospital, there were 32 percent more deaths due to cancer than certified on death certifi- cates (3,095 vs. 2,345) and for persons aged 70 or more dying at home it is estimated there were 55 percent more stomach cancer (269 estimated vs. 174 listed) and 244 percent more lung cancer (141 estimated vs. 41 listed) than were certified on death certificates. “The death certificate is not a good source of information for cancer of the cervix because many cases of this disease are reported on death certificates as cancer of the uterus. This practice needs to be taken into account in the use of mortality data for cervical cancer in Japan.” Reprinted with permission of the Gann Journal. 118. Steinitz, Ruth and Costin, Corina (BD-2-1S-65) Cancer Mortality - Vital Statistics Versus Cancer Registry ISRAEL JOURNAL OF MEDICAL SCIENCES 7(12): 1405-1412, 1971 Epidemiological conclusions about cancer mor- tality among Israel’s Jewish population have been based on death certificate data supplied by the Central Bureau of Statistics (CBS). To see whether these conclusions would be different if the CBS data were checked against data from the Israel Cancer Registry (ICR), standardized cancer mortality rates for Jews dying in 1964-65 were computed by site and sex from ICR data and compared with rates based on CBS data published for the same time period. Frequency distributions by age and site are also compared for the two data sources. Observed differ- ences and possible explanations for the differences are discussed briefly. 119. Virkkunen, M.; Penttila, A.; Tenhu, M.; Huittinen, V.-M.; Lehti, H.; Rissanen, V.;and Uotila, U. (ABC-9-FN-99) Comparative Study on the Underlying Cause and Mode of Death Established Prior to and After Medi- colegal Autopsy FORENSIC SCIENCE 5(1):73-79, 1975 To investigate the value of external medicolegal examination of decedents as an alternative to medico- legal autopsy, 600 selected cases on which a medico- legal autopsy had been performed in the city of Helsinki were studied. Six physicians of the Depart- ment of Forensic Medicine reviewed all information that had been available for these cases before autopsy and assigned causes of death according to the Eighth Revision ICD, and modes of death according to the international rules of deaths from natural causes, accidents, suicides, homicides, and deaths in which the mode could not be determined. Comparison of the cause and mode of death with those established after medicolegal autopsy revealed that the mode of death differed in 60 cases, primarily in the groups of accidental deaths, suicides, and natural deaths. The underlying cause of death diverged in about 29.5 percent of the 600 cases. For deaths from natural causes, the cause of death was misclassified most 37 often in deaths due to subarachnoidal hemorrhage, cerebral hemorrhage, or malignant tumors. For violent deaths, misclassification occurred most often when death was due to the toxic effects of drugs or alcohol. The authors conclude that the medicolegal autopsy can significantly improve the accuracy of assignment of the underlying cause of death. 120. Waaler, Erik and Grimstvedt, Magne (AC-2-NW-54) The Clinical Diagnoses of the Causes of Death and Their Reliability ACTA PATHOLOGICA ET MICROBIOLOGICA SCANDINAVICA 43(4):330-338, 1958 The authors compared autopsy and clinical findings for the 783 malignancy deaths in Bergen, Norway for the period 1950 through 1954. Using the autopsy findings as definitive, premortem clinical diagnoses were classified as correct (63 percent), doubtful (10 percent), and wrong (19 percent). In 6 percent the malignant tumors were an incidental finding not related to the cause of death and in 13 cases (1.6 percent) the primary tumor could not be found with certainty at the post-mortem examination. The autopsy rate for malignancy cases in Bergen was 30 percent during the period. Carcinomas of the breast, cervix and corpus of the uterus, and leukemias were virtually all correct (97-100 percent), while the percent for less accessible tumors ranged from 0 percent for papilla of vater and duodenum and 13 percent for liver to 60 percent for colon and 62 percent for stomach. 121. Waldron, H. A. and Vickerstaff, Lorna (ACD-9-UK-76) Intimations of Quality: Ante-Mortem and Post- Mortem Diagnoses Nuffield Provincial Hospitals Trust, London, 1977 To examine the degree of agreement between the ante mortem and post mortem causes of death, a prospective study modeled on the Heasman and Lipworth study was carried out in the West Midlands and the Trent regions. Hospitals in these regions were sent two-part forms; the first was to indicate the clinician’s diagnosis prior to autopsy, and the second was to reflect the cause of death determined by the pathologist, based on both autopsy results and discussion with the clinician. All deaths autopsied in the participating hospitals from January 1975 through April 1976 were included in the study. 38 Of the 1,117 cases for which both ante mortem and post mortem diagnoses were available, 531 (47.5 percent) showed complete agreement between the clinical diagnosis and autopsy diagnosis, 295 (26.4 percent) showed partial agreement, and 291 (26.1 percent) indicated disagreement between the clinical and autopsy diagnoses. Disagreements were more likely for females, the elderly, nonteaching- hospital patients, those with a clinical diagnosis of respiratory disease, and those in which the clinical diagnosis was uncertain. 122. Walford, P. A. (CD-3-UK-99) Sudden Death in Coronary Thrombosis: A Study of the Accuracy of Death Certification JOURNAL OF THE ROYAL COLLEGE OF GEN- ERAL PRACTITIONERS 21:654-656, 1971 Eleven general practitioners supplied information on all cases (142 total) encountered in their practices of sudden deaths classified as due to coronary throm- bosis. They were asked to give the bases for the final diagnoses and to rate the certainty of the diagnoses. In 57 of the 142 cases, there was no ante mortem evidence of coronary disease and no necropsy had been performed. The author concludes that general practitioners are introducing into mortality statistics a considerable though unquantifiable error. 123. Walsh, Brendan; Walsh, Dermot; and Whelan, Brendan (B-7-IR-68) Suicide in Dublin: II. The Influence of Some Social and Medical Factors on Coroners’ Verdicts BRITISH JOURNAL OF PSYCHIATRY 126:309- 312,1975 (Author’s summary) ‘“This paper presents an analysis of the factors which influence coroners in their decision to classify some deaths as suicides and others as accidental or ‘open’. The most important influence on coroners’ behavior was seen to be the manner by which the person died. Those who died by cutting, hanging, drugs or gas were significantly more likely to receive a suicide verdict than those whose deaths were due to drowning, jumping, shooting, or poisoning. If the deceased left any intimation of a suicidal intent, this increased the likelihood that a suicide verdict would be returned. Finally, persons aged under 40 were significantly more likely to be returned as suicides than older victims, especially those aged over 70. All of these results show that coroners operate by observing the law as it defines suicide, that is, by looking for evidence of intent of self-inflicted death. “Our findings concerning the factors associated with the suicide verdict help to clarify the meaning of the official data on suicides in Ireland, and illuminate the reasons why, using clinical rather than legal criteria, a much higher rate is obtained.” NOTE: The study was based on 201 coroners’ cases in Dublin during 1964-68 which were judged to be suicides in a previous study. (See Number 76 in this report.) Reprinted with permission of the British Journal of Psychiatry. 124. Warshauer, M. Ellen and Monk, Mary (BDS-7-US-70) Problems in Suicide Statistics for Whites and Blacks AMERICAN JOURNAL OF PUBLIC HEALTH 68(4):383-388, 1978 (Author’s abstract) “The accuracy of suicide statistics was assessed by comparing published Health Department suicide rates for an area of New York City with Medical Examiner records. For the period 1968-1970, records from the Medical Examiner’s Office were searched to determine all deaths classified as definite suicides. Another group of deaths was considered suicide by the Medical Examiner but never classified as such. These deaths we labeled ‘assigned suicides’. “When definite suicides were compared with all deaths considered suicide by the Medical Examiner (definite and assigned suicides), black suicide was underestimated by 80 percent and white suicide by 42 percent. Underestimation was the same for males and females but varied by age group. “In 1968, when the seventh revision of the Inter- national Classification of Deaths (ICD) was used, Health Department suicide rates for blacks were almost identical to Medical Examiner rates, while white rates were underestimated by 25 percent. In 1969-1970, when the eighth revision was used, Health Department statistics underestimated black suicides by 82 percent and white suicides by 66 percent. “Reasons for the underestimations were related to the methods used in committing suicide by the two ethnic groups and to the ways that suicide classification has changed from the seventh to eighth revision. Implications for research using official death certificate reports are discussed.” Reprinted by permission of the American Journal of Public Health. 125. Weiss, Noel S.; Green, Delray; and Krueger, Dean E. (DS-8-US-68) Problems in the Use of Death Certificates to Identify Sudden Unexpected Infant Deaths HEALTH SERVICES REPORTS 88(6):555-558, 1973 (Author’s abstract) “Death certificates on all U.S. infants who died at 3 months of age in 1968 were examined to see if sufficient information was avail- able to identify those whose death was sudden and unexpected. Of the 2,954 deaths, 371 were coded for official mortality statistics on underlying cause of death to nonspecific causes implying, or compatible with, sudden unexpected death (ICDA eighth revision, 795, 796.2, 796.3, 796.9). An additional 151 deaths were described on the death certificate as sudden and unexpected, but because of the presence of other information in the cause of death section, they were assigned to various specific causes of death. Based on incidence rates from several earlier studies, it was estimated that there remained at least several hundred additional sudden unexpected infant deaths coded to other causes. Nonetheless, other items that might have been helpful in identifying these deaths, such as approximate interval between onset and death, place of death, and type of certifier, were either in- frequently recorded or not sufficiently discriminating to establish criteria for sudden unexpected death that were both sensitive and specific. “Unless changes are made in the construction of the death certificate, the completeness and accuracy with which it is filled out, or the coding of under- lying cause of death, it is unlikely that accurate rates of sudden unexpected infant death will be routinely produced in the United States.” 39 126. Wilson, Robert R. (AC-9-UK-58) In Defense of the Autopsy JOURNAL OF THE AMERICAN MEDICAL ASSO- CIATION 196(11):1011-1012, 1966 To assess the usefulness of autopsies, the clinical diagnosis was compared with the post mortem diag- nosis for 265 adult decedents (excluding cases “brought in dead”) autopsied in 1958 in Paddington General Hospital, London. Using the post mortem diagnosis as the standard, the author found that the clinical diagnosis was absolutely correct in 53 percent of the cases, partly correct in 40 percent, and totally wrong in 7 percent. 127. Worth, R. M.; Kato, H.; Rhoads, G. G.; Kagan, A.; and Syme, S. L. (ACD-9-Z2-72) Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Mortality AMERICAN JOURNAL OF EPIDEMIOLOGY 102(6):481-490, 1975 (Author’s abstract) “Stroke, coronary heart disease (CHD) and total mortality are evaluated from death certificates in enumerated cohorts of 45- 64-year-old Japanese men in Hiroshima and Nagasaki (1965-1970), in Honolulu (1966-1970), and in the San Francisco area (1968-1972). Total mortality is highest in Japan with no consistent differences between Japanese Americans in Honolulu and San Francisco. Age-specific CHD death rates are marked- ly lower in all three Japanese groups than in Amer- ican whites. The CHD rates are consistently and significantly lower in Japan than in American Japan- ese. Stroke death rates for American Japanese men appear equivalent to figures for U.S. white men of the 40 same age, but are significantly lower than in the Japan cohort for the 60-64-year-old group. The number of stroke deaths below that age are too few as yet for analysis. Validation of mortality ascertain- ment and of the accuracy of death certification had been carried out in Japan and in Hawaii. The inter- national differences in mortality do not appear to be due to certification or other methodologic artifact.” NOTE: In Japan, comparison of the certified cause of death to autopsy findings for about one-third of the study group yielded estimates of 20 percent overreporting for CHD and 5 percent overreporting for stroke. In Honolulu, comparison of the certified cause of death with data from hospital records, the medical examiner’s office, autopsy reports, and attending physicians yielded estimates of 22 percent overreporting of CHD deaths and 15 percent underreporting of stroke deaths. Reprinted with permission of the American Journal of Epidemiology. 128. Wright, D. J. M. (CD-8-UK-66) Inaccuracy and Under-Reporting in Certification of Death Following Urethral Stricture MEDICINE, SCIENCE, AND THE LAW 9:205-207, 1969 In a comparison of death certificates and clinical case records, the author obtained from the General Register Office all 247 death records for the years 1964-66 pertaining to urethral stricture. Useful case notes were obtained for 165 of the 180 inhospital deaths. The underlying cause of death was misdiag- nosed as urethral stricture in 12 percent of the 165 cases. Errors in assignment of the direct cause of death were found in 18 percent of the cases. Incom- pleteness of case notes was a serious problem: in a comparison with case notes for 85 current patients with urethral stricture at two teaching hospitals, the site and cause were recorded 20 percent less often in the case notes for the study sample. Using estimates by Caine (1954) and Dunlop (1961), the author con- cludes that over 90 percent of urethral strictures are unreported on the death certificate. Index Indexes are presented for each of the four clas- sification variables. There are three indexes by data source and one index each by cause of death, country, and latest year of data. 1. Articles indexed by data source A. Articles that compare death certificates (D) with autopsy results (A), clinical records (C), mortality or autopsy statistics (S), and/or other sources of information (B): Data source code AD BD CD DS ACD BCD BDS CDS ABCD Article numbers 13, 29, 31, 40, 43, 53, 77, 78,102,117, 15,44,73,74,76,114, 118 5 6, 8, 22, 23,127,128, 37, 38, 49, 50, 52,56,59,62,70, 71, 81, 82, 83, 87, 109, 112, 113,122,128 21, 26, 33, 54,69,90,91,92, 99, 106, 107, 125 14, 24, 30, 32, 34, 46, 58, 61, 65, 89,98, 100, 115, 121, 127 3,19,93 124 36, 85 35, 63, 64, 66, 67, 68,95, 96, 97 B. Articles that compare autopsy results (A) with clinical records (C), mortality or autopsy statistics (S), and/or other data sources (B): Data source . Article numbers code AB 11,75 AC 12, 17, 18, 20, 25, 39,45, 51, 53, 79, 104, 105, 108, 120, 126 AS 16 ABC 119 ACS 57 C. Articles that are based on only one data source; that is, autopsies (A), death certifi- cates (D), clinical records (C), mortality or autopsy statistics, (S), or other sources (B): Data source . Article numbers code A 72,84,116 B 88,123 D 1,2, 60, 80 S 4,7,9,10,41,42,47, 48, 86, 94,101,103,110,111 41 2. Articles indexed by cause of death Cause of death code 1 2 Article numbers 24, 38,58 8, 12, 14, 21, 28, 29, 41, 44, 56, 59, 70, 84, 85, 86, 89, 104,118, 120 35, 39,48, 52,66,67,63,79, 81, 83, 87, 91, 92, 98, 101, 103,107,109, 111,115,122 3, 33, 34, 57,61, 62, 63, 44, 65,112 6,7,73,71,78 9,10,11,50,76,88, 106,123, 124 4,23,25,45,49,71,125,128 1,2,5,13,15,16,17,18,19, 20, 22, 26, 27, 30, 31, 32, 36, 37, 40, 42, 43,46,47, 51,53, 54, 55, 60, 69,72,74,75, 80, 82,90, 93, 94, 95, 96,97,99, 100, 102, 105, 108, 110, 113, 114, 116, 117, 119, 121, 126, 127 3. Articles indexed by country 42 Country code AU CN CY FN IR IS Jr NW NZ Article number 40 6,7,8 90 91,92, 119 15, 26, 48, 75, 76, 85, 86, 123 29,79,118 55,117 120 43 SW Us ZZ Year code 45-50 51-55 56-60 61-65 66-70 71-75 76-79 99 17, 18,27,45,72 4, 5, 10, 11, 14, 23, 32, 38, 41, 42, 46, 50, 53, 89, 103, 106, 121, 122,126, 128 2,3,12,13,16,19,20, 21,22, 24,25, 28, 30,31,33, 34, 35, 36, 37, 39, 44, 47,49, 51, 52, 54, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68,69,70,71,73, 74,1717, 78, 81, 82, 83, 84, 87, 38, 93, 98, 99, 101, 102, 104, 105, 107, 108, 109,110, 111, 112, 113, 114, 115, 116, 124, 125 1,9,56,57,80,94,95, 96,97, 100, 127 4. Articles indexed by latest year of data Articles 28 14,42,111, 120 7. 8, 13, 31, 32, 39, 44, 45, 46, 47, 70, 71, 81, 82, 83,90, 100, 115, 116, 126 5,6,12,15,24,25,33,34, 35, 48, 49, 55, 60, 61, 62, 63, 64, 65, 66, 67, 68,73,74,79, 85, 89, 94, 95, 96, 98, 102, 103, 112,114,118 3,9,10,11,21,26,29, 30,50, 54,57, 58,692,75,76,77, 78, 86, 87,91, 92,117, 123, 124, 125, 128 2,4,.16,17,18,19,20,22,27, 36, 37, 38, 40,41,43,52, 56, 88, 93, 97, 99, 104, 105, 107, 108, 109, 110, 113, 127 23,72,121 1,51,53,59, 80,84, 101, 106, 119,122 U.S. GOVERNMENT PRINTING OFFICE: 1982-361-161:516 Vital and Health Statistics series descriptions SERIES 1. SERIES 2. SERIES 3. SERIES 4. SERIES 10. SERIES 11. SERIES 12. SERIES 13. Programs and Collection Procedures.—Reports describing the general programs of the National Center for Health Statistics and its offices and divisions and the data col- lection methods used. They also include definitions and other material necessary for understanding the data. Data Evaluation and Methods Research.—Studies of new statistical methodology including experimental tests of new survey methods, studies of vital statistics collection methods, new analytical techniques, objective evaluations of reliability of collected data, and contributions to sta- tistical theory. Analytical and Epidemiological Studies.—Reports pre- senting analytical or interpretive studies based on vital and health statistics, carrying the analysis further than the expository types of reports in the other series. Documents and Committee Reports.—Final reports of major committees concerned with vital and health sta- tistics and documents such as recommended model vital registration laws and revised birth and death certificates. Data from the National Health Interview Survey.—Statis- tics on illness, accidental injuries, disability, use of hos- pital, medical, dental, and other services, and other health-related topics, all based on data collected in the continuing national household interview survey. Data From the National Health Examination Survey and the National Health and Nutrition Examination Survey.— Data from direct examination, testing, and measurement of national samples of the civilian noninstitutionalized population provide the basis for (1) estimates of the medically defined prevalence of specific diseases in the United States and the distributions of the population with respect to physical, physiological, and psychological characteristics and (2) analysis of relationships among the various measurements without reference to an explicit finite universe of persons. Data From the Institutionalized Population Surveys.—Dis- continued in 1975. Reports from these surveys are in- cluded in Series 13. Data on Health Resources Utilization.—Statistics on the utilization of health manpower and facilities providing SERIES 14. SERIES 15. SERIES 20. SERIES 21. SERIES 22. SERIES 23. long-term care, ambulatory care, hospital care, and family planning services. Data on Health Resources: Manpower and Facilities.— Statistics on the numbers, geographic distribution, and characteristics of health resources including physicians, dentists, nurses, other health occupations, hospitals, nursing homes, and outpatient facilities. Data From Special Surveys.—Statistics on health and health-related topics collected in special surveys that are not a part of the continuing data systems of the National Center for Health Statistics. Data on Mortality.— Various statistics on mortality other than as included in regular annual or monthly reports. Special analyses by cause of death, age, and other demo- graphic variables; geographic and time series analyses; and statistics on characteristics of deaths not available from the vital records based on sample surveys of those records. Data on Natality, Marriage, and Divorce.—Various sta- tistics on natality, marriage, and divorce other than as included in regular annual or monthly reports. Special analyses by demographic variables; geographic and time series analyses; studies of fertjlity; and statistics on characteristics of births not available from the vital records based on sample surveys of those records. Data From the National Mortality and Natality Surveys.— Discontinued in 1975. Reports from these sample surveys based on vital records are included in Series 20 and 21, respectively. Data From the National Survey of Family Growth.— Statistics on fertility, family formation and dissolution, family planning, and related maternal and infant health topics derived from a periodic survey of a nationwide probability sample of ever-married women 15-44 years of age. For a list of titles of reports published in these series, write to: Scientific and Technical Information Branch National Center for Health Statistics Public Health Service Hyattsville, Md. 20782 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics 3700 East-West Highway Hyattsville, Maryland 20782 POSTAGE AND FEES PAID U.S. DEPARTMENT OF HHS HHS 396 THIRD CLASS EE U.S.MAIL Em OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 HRST From the Office of Health Research, Statistics, and Technology DHHS Publication No. (PHS) 82-1363. Series 2, No. 89 For listings of publications in the VITAL AND HEALTH STATISTICS series, call 301-436-NCHS U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES e Public Health Service ® National Center for Health Statistics ® Series 2, No. 9( ~~ The Collection and = | Processing of Drug gl (lH | National Ambulatory CT Co Re TTY ' United States, 1980 Data Evaluation and Methods Research Series 2, No. 90 SUGGESTED CITATION National Center for Health Statistics, H. Koch: The collection and processing of Drug information: National Ambulatory Medical Care Survey, United States, 1980. Vital and Health Statistics. Series 2, No. 90. DHHS Pub. No. (PHS) 82-1364. Public Health Service. Washington. U.S. Government Printing Office, March, 1982. Library of Congress Cataloging in Publication Data Koch, Hugo K. The collection and processing cf drug information in the national ambulatory medical care survey, United States, 1980. (Data evaluation and methods research, series 2 ; no. 90) (DHHS publication ; no. (PHS) 82-1364). Includes biographical references. Supt. of Docs. no.: HE 20.6209: 2/90 1. Drugs, Prescribing—United States. 2. Drugs, Code numbers. 3. Ambulatory medical care—United States. 4. Health surveys—United States. |. Campbell, William H. (William Howard), 1942- . Il. Vital and health statistics, Series 2, Data evaluation and methods research ; no. 90. Ill. Series: DHHS publication ; no. (PHS) 82-1364. RA409.U45 no. 90 [RM138] 312'0723s 81-607138 ISBN 0-8406-0242-1 [615°.1'0723] AACR2 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 VIALE HERLTHOIATR] The Collection and Processing of Drug Information: National Ambulatory Medical Care Survey United States, 1980 The report describes the method and instru- ments used to collect and process drug infor- mation for the 1980 National Ambulatory Medical Care Survey. It explains in detail the development of the system by which the drug information was classified and coded, and offers a complete set of coding instructions along with an alphabetized list of 7,227 drugs that office-based physicians might prescribe. Finally, it presents the plan for analyzing the drug data and reporting it to potential users. Data Evaluation and Methods Research Series 2, No. 90 DHHS Publication No. (PHS) 82-1364 U.S. Department of Health and Human Services Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics Hyattsville, Md. March 1982 National Center for Health Statistics DOROTHY P. RICE, Director ROBERT A. ISRAEL, Deputy Director JACOB J. FELDMAN, Ph.D., Associate Director for Analysis and Epidemiology GAIL F. FISHER, Ph.D., Associate Director for the Cooperative Health Statistics System GARRIE J. LOSEE, Associate Director for Data Processing and Services ALVAN 0. ZARATE, Ph.D., Assistant Director for International Statistics E. EARL BRYANT, Associate Director for Interview and Examination Statistics ROBERT C. HUBER, Associate Director for Management MONROE G. SIRKEN, Ph.D., Associate Director for Research and Methodology PETER L. HURLEY, Associate Director for Vital and Health Care Statistics ALICE HAYWOOD, Information Officer Division of Health Care Statistics W. EDWARD BACON, Ph.D., Director JOAN F. VAN NOSTRAND, Deputy Director JAMES E. DELOZIER, Chief, Ambulatory Care Statistics Branch MANOOCHEHR K. NOZARY, Chief, Technical Services Branch Acknowledgements The authors acknowledge a special debt to P. May Douglas, Coordinator for the Drug Data Processing Systems and Services of the American Society of Hospital Pharmacists, for her advice on this project. We are also grateful for the contributions of the following persons: Walter Anyan, M.D. David W. Bailey, M.D. Charles E. Brackett, M.D. Saul Boyarsky, M.D. Herbert Brendler, M.D. John C. Byrne, M.D. Lynn P. Carmichael, M.D. John A. D. Cooper, M.D. Edward P. Crowell, D.O. John P. Connelly, M.D. Robert H. Drachman, M.D. Mary B. Forbes Todd M. Frazier Harriette Gibson Charles Given, Ph.D. Irving D. Goldberg Birt Harvey, M.D. Reginald T. Harris, M.D. Richard J. Hannigan Charles V. Heck, M.D. James I. Hudson, M.D. Arthur P. Herrmann, Jr., Pharm. D. Deane E. Knapp, Ph.D. Ralph E. Kauffman, M.D. Charles E. Lewis, M.D. Robert H. Moser, M.D. Ervin E. Nichols, M.D. Nelson G. Richards, M.D. C. H. William Ruhe, M.D. Melvin Sabshin, M.D. David L. Starbuck, M.D. Hugh C. Thompson, M.D. James B. Tenney, M.D. Richard A Terselic Herb L. Tindall, M.D. Peyton E. Weary, M.D. Kerr L. White, M.D. Maurice Wood, M.D. na . * ~ ; , Contents AcknOWISAUBIMBINLS « « + o.5 sv wa 55 51.0 5.0% 5.5 5 815 BE 25 00 9 0008 #50) 8000508, 00000 88) BIR we 0 ee 0c 8 81K bk 1 0 1] EEC OI OTIC eto ee ed ler ie ie eel ef Gerben hears ed ie RAEN Rare ie red ree Beet re Te le Tees Cee ch ert re CoE CELL ve eT ee he] here Ee ne ele Le he oe Hee 1 Drug identification and desCription . wu. sus savas nes sv av ams s sis armts sao sas sabe dossenusasssisnessss 8 Classifying and coding deviCe «ose var ov ves rar rmassssnssrassnpnss 50 0800 6 HL 6 REE 1" CONG DLO0BES ovis a vos 480s 58% 40a EE E30 0 EWE ESS GERD HA B80 Rad) Fw Bp ip 0 08H GE BW 08 ER 16 Appendixes COAIENGS ov 0 5 a FR 0 fr ed hh 2 Cr BT Ah rate Pee ve aria ery re re rit Rete ren ve were sn 2) |. Response to precoded questions on evaluation INTEIVIBW . ov vv vv vv vt tt ttt t tit tent t tenses nenenens 22 Il. Inventoryofgenericnames ...........oo0viennnrans BALE A WR oR wae 25 111. Medication Code List, NAMCS 1980 ................ ERE REE E dE we Ep pg 32 IV. Coding procedures for medication entries, NAMCS 1980 ......... oath tt EL Peta a oh her 4 reer oet en amet 2 er ver om en aver . 73 V. American Hospital Formulary Service classification system and therapeutic category codes ...........covuvuv.n 20 List of text figures 1. Patient Record, 1078... iii i iit itt ttt e ttt intents ens enon nsennens taser eenns 2 2. Patient Record, Version A, 1980 . . «vv ss mai sims sn sas ai@ sss 5058s 5 eas@ se awniedse se svees sane 3 3. Pationt Record, Version B, TOB0 ... vase vs fio snes a sans ass awed ola malin ss oe mes +5 vem dawson ones 4 4. Patient Record and Instruction on Patient Record jacket . ......... citi iinet inti 7 8. AnalYSs pian: POBMIBI CONIA . «vcs vnmin sve s 3aid ses Da000 0 F060 0 ® 50 ESSE 5 EEE se uw ® Esse 20 List of text tables A: DPIF record TOMA «vs ss 00500 a5 5 5.50553 5 8 #50030 5 0 8 5.5 Fee 4 (5.505 bn 5 06 #50 5140 50x 5 00 fo 0500s hac 104181 1 1 0,050 10 wt rca 12 B. NAMCS drug file format... .. iii tit tit ttt et ttt tees e eee e tintin enna 13 C. Coding choices provided and USE, MCL. TOBD . J + caine sts vmmm v8 5m ms «5 6 sew 56 8 SES Es 6 0668 0 Ewe 19 - - o = Bh . : al x =, . FB H 2 x : . ) ) . Te . B - - y In wl } . } } E - ® » 5 . . % oat . . oo . & e i oo. “ B & y . } } od oo g a } : . | . - . N . } - - a “ B . - a § - . . + . . y ) ] ) | | - | . T Ge =e : i . oo I ) rr BE | | | . ) . ) >) =) - a \ ) ) 3 | | ) } " -: # | | ) E [ - ] " . ot . a . - ) . , ° - - . } » | Bh ’ + ) : . . x ) : } . ) » - . . Lt £ I . oo : N - * . The Collection and Processing of Drug Information: National Ambulatory Medical Care Survey by Hugo Koch, M.H.A., Division of Health Care Statistics and William H. Campbell, Ph.D., School of Pharmacy, University of Washington at Seattle Background This report describes the method and instruments used to collect and process drug information for the 1980 National Ambulatory Medical Care Survey (NAMCS). In this report, the term drug is inter- changeable with the term medication and includes all immunizing and desensitizing agents. The use of the term prescribing a drug or of the physician’s prescrib- ing habits, is used in its broader sense, namely to mean the ordering or provision of any medication— prescription (Rx) or nonprescription (OTC)—in the course of the office visit. The relationship between prescribing and using drugs is inferential, subject to the patient’s compliance with the doctor’s decision. The purpose of NAMCS is to collect and report descriptive data about the medical care provided in the physician’s office. Each year beginning with 1973, NAMCS has sampled 3,000 of the Nation’s roughly 200,000 doctors of medicine and osteopathy who are principally engaged in office-based, patient- care practice. The sample excludes: physicians who are federally employed; physicians who practice in Alaska and Hawaii; and physicians who, though office-based, specialize in anesthesiology, pathology, or radiology. Among eligible physicians, NAMCS uses a multistage probability design that distributes its sample geographically by primary sampling units (PSU’s) and functionally by the specialties of the physicians who practice within the PSU’s. Sampled physicians complete records (figure 1) for a system- atic random sample of their office visits over a weekly reporting period. In 1978, the year of the Patient Record shown in figure 1, sampled physicians responded at a rate of 72.8 percent. This produced a raw data base of 47,291 patient records, from which an estimated 584,498,000 total visits was projected. In designing the Patient Record each year, the prime motivating factor has been to spare the physi- cian time and effort. The approach, typified by the 1978 Record, has been as indicated below. 1. To ensure that the record format does not exceed one side of one page, the overall number of items must be limited. 2. To make sure that the items are as self-explana- tory as possible, or place clarifying information on the face of the Record form. If necessary, special instructions or definitions are made readily available to the recording physician by either printing them directly on the jacket in which the Patient Records are filed or on a card which is inserted in the jacket. 3. To hold write-in responses to a minimum, placing relatively more reliance on the checkbox. Until the 1980 NAMCS, every Patient Record contained checkbox items for nonspecific drug therapy ordered or provided during the visit. In brevity and content, all the items closely resembled the drug question that appeared on the 1978 Record under Item 12. Two subitems, Item 12.2 and 12.3, elicit relevant drug response. Though this approach to gathering drug informa- tion agreed with the NAMCS tradition by requiring a laudable economy of respondent effort, it unfortu- nately resulted in a regrettable paucity of descriptive detail about the role played by drugs in office-based medical care. For example, statements such as this could be made. “In 1978 there were an estimated 348,262,000 drug visits (visits at which at least one drug was prescribed). Drug visits amounted to about 60 percent of all office visits.” It could be shown how the frequency of drug visits varied with such influencing factors as physician specialty, primary diagnosis, age or race of patient, etc. But beyond this limited description, additional information about the drug therapy was not available. For example, specific drug products could not be 1 ASSURANCE OF CONFIDENTIALITY —AIl information which would permit identification of an individual, a practice, or an establishment will be held confidential, will be used only by persons engaged in and for D 6 1 0 4 3 i the purposes of the survey and will not be disclosed or released to other persons or used for any other purpose. 1. DATE OF VISIT PATIENT RECORD Lot NATIONAL AMBULATORY MEDICAL CARE SURVEY Mo/Day/ Yr 2. DATE OF BIRTH 3. SEX 4. COLOR OR 5. WAS PATIENT 6. PATIENT'S COMPLAINT(S), SYMPTOM(S), OR OTHER RACE REFERRED FOR REASON(S) FOR THIS VISIT THIS VISIT BY (In patient’s own words) +O wHITE ANOTHER wTof— 1 0 FEMALE 2 [J NEGRO/ PHYSICIAN? a. MOST TIME OF Mo /Day / Yr BLACK IMPORTANT 2 0 maLE 10 ves VISIT » J OTHER b. OTHER + [J UNKNOWN : Ono em. | 7. TIME SINCE ONSET 8. PHYSICIAN'S DIAGNOSES 9. HAVE YOU SEEN 10. SERIOUSNESS OF eesttettserssmmmmmsmnn OF COMPLAINT/ PATIENT BEFORE? CONDITION IN m SYMPTOM IN ITEM 6a a. PRINCIPAL DIAGNOSIS/PROBLEM ASSOCIATED WITH ITEM 8a (Check one) B (Check one) ITEM 6a 1 3 ves 2 [J NO F— 0 1 [J VERY SERIOUS 1 [J LESS THAN 1 DAY am. DA is Yes, FOR THE t [J SERDUS 23 LE DAYS ONDITION IN Bn b. OTHER SIGNIFICANT CURRENT DIAGNOSES ITEM 8a? SLIGHTLY p.m, 3 [J 1-3 WEEKS SERIOUS + [J 1-3 MONTHS + [J NOT SERIOUS tO ves 20n0 am. s [] MORE THAN 3 MONTHS p.m. ¢« [J NOT APPLICABLE on 11. DIAGNOSTIC SERVICES THIS 12. THERAPEUTIC SERVICES THIS 13. DISPOSITION THIS VISIT 14. DURATION OF Lai VISIT (Check all ordered or provided) VISIT (Check all ordered or provided) (Check all that apply) THIS VISIT er smn (Time actually 1 OJ NONE + 0 None Og : NO FOLLOW-UP PLANNED t with M1, O LIMITED EXAM/HISTORY 2 0 iIMMUNIZATION/ sO » a po] 3 [J GENERAL EXAM/HISTORY DESENSITIZATION RETURN AT SPECIFIZO TIME « 0 PAP TEST » [J DRUGS (PRESCRIPTION/ * [J RETURN IF NEEDED, P.R.N. Bm. NONPRESCRIPTION) » |_| s OcLINICAL LAB TEST * [J DIET COUNSELING [J TELEPHONE FOLLOW-UP PLANNED « O x-RAY s [J FAMILY PLANNING * 0 REFERRED TO OTHER PHYSICIAN pol Oexe a MINUTES . MEDICAL COUNSELING ‘ RETURNED TO REFERRING s [J viIsION TEST 7 [J PHYSIOTHERAPY PHYSICIAN s» [J ENDOSCOPY s [J OFFICE SURGERY 7 [J] ADMIT TO HOSPITAL o [J BLOOD PRESSURE CHECK 9 J PSYCHOTHERAPY/ . ; 1 OJ OTHER (Specify) THERAPEUTIC LISTENING O oTHER (Specify) 10 [J OTHER (Specify) hase? DEPARTMENT OF HEALTH, EDUCATION AND WELFARE O.M.B. #68-R1498 PUBLIC HEALTH SERVICE HEALTH RESOURCES ADMINISTRATION NATIONAL CENTER FOR HEALTH STATISTICS Figure 1. identified by name and counted. Certain facts that bear critically on a full understanding of drug utiliza- tion were left unexplored, such as: the extent of generic prescribing, the relative reliance on combina- tion drugs, and the use of usually-less-costly nonpre- scription drugs. Except for the immunizing and desensitizing agents, there were only inferential clues to the therapeutic effect that the physician desired in using the drug. Sensitive to the limited utility of the NAMCS drug data, its panel of advisory experts had for several years urged a more ambitious effort. Two concerns delayed the implementation of their suggestion. One, it was thought that a more detailed probe into the prescribing habits of respondents might act negatively on their response rate. Drug therapy is cur- rently one of the more controversial facets of health care. Although the confidentiality of data provided in NAMCS is protected by statute, NAMCS planners feared that physicians might assume that their re- sponses would be used to evaluate the quality of their Patient Record, 1978 practice, monitor their Medicare or Medicaid billings, or be otherwise interpreted to their disadvantage. Two, the problems of classifying and coding the detailed drug data might prove too costly and com- plex for practical management. However, realizing that subjective reservations should not by themselves be allowed to debar a po- tentially useful action, NAMCS planners began in 1978 to pursue a course of action that would objec- tify the problems implicit in an extended drug inquiry, and to solve as many of these problems as possible, with the tentative aim of incorporating the revised approach in the 1980 survey. This course of action led to the undertaking of the following specific tasks: 1. A formal study that would evaluate the feasibility of obtaining more specific drug information and indicate the format that such a revised drug item should take. 2. A search for a system by which drug data could ASSURANCE OF CONFIDENTIALITY — All information which would permit identification of an individual, a practice, or an establishment will be held confidential, will be used only by persons engaged in and for the purposes of the survey and will not be disclosed or released to other persons or used for any other purpose. IE 311319 Y. PATECFVISH PATIENT RECORD Ve LY NATIONAL AMBULATORY MEDICAL CARE SURVEY 2. DATE OF 3. SEX 4. COLOR OR B. ETHNICITY 6. WAS PATIENT 1. PATIENT'S COMPLAINT(S), SYMPTOM(S), BIRTH RACE BE Ane OR OTHER REASON(S) FOR THIS VISIT 1 O WHITE 1 0 HISPANIC THIS j jent’ 20) BAGH HISPAN VISIT BY [in patient's own words] 3 [0 ASIAN/PACIFIC ANOTHER . MOST Wo. Gey vis | 1osemaLe OSU ANOER 20 Not PHYSICIAN? * IMPORTANT "| 2O0mALE | 4D AMERICAN INDIAN/ HISPANIC ALASKAN NATIVE 1 > YES b. OTHER 20 NO 8. MAJOR REASON FOR THIS VISIT [Check One] 1 0 ACUTE PROBLEM 2 0 CHRONIC PROBLEM, ROUTINE 3 [0 CHRONIC PROBLEM, FLAREUP 4 [0 POST SURGERY INJURY 5 [0 NON-ILLNESS CARE (ROUTINE PRENATAL, GENERAL EXAM, WELL BABY, ETC.) 1D 6 0 DIAGNOSTIC SERVICES THIS VISIT [Check all ordered or provided) NONE 2 O LIMITED HISTORY/EXAM 3 OJ GENERAL HISTORY/EXAM 10 OO ENDOSCOPY 4 O PAP TEST X-RAY 8 OJ EKG 9 J VISION TEST 11 J MENTAL STATUS EXAM 5 OJ CLINICAL LAB TEST 12 [J OTHER (Specify) 7 0 BLOOD PRESSURE CHECK 10. PHYSICIAN'S DIAGNOSES a. PRINCIPAL DIAGNOSIS/PROBLEM ASSOCIATED WITH ITEM 7a 11. HAVE YOU * SEEN PATIENT BEFORE? tvs 2ONO IF YES. FOR THE CONDITION IN b. OTHER SIGNIFICANT CURRENT DIAGNOSES ITEM 10a? 1OYES 20NO 12 DISPOSITION THIS VISIT * [Check all that apply] 1 00 NO FOLLOW-UP PLANNED 2 00 RETURN AT SPECIFIED TIME 3 0 RETURN IF NEEDED, P.R.N. 4 0 TELEPHONE FOLLOW-UP PLANNED 5 [0 REFERRED TO OTHER PHYSICIAN 6 [J RETURNED TO REFERRING PHYSICIAN 7 OJ ADMIT TO HOSPITAL 8 [J OTHER (Specify) 13. DURATION OF THIS VISIT [Time actually spent with physician) MINUTES 14. NON-MEDICATION THERAPY [Check all services ordered or provided this visit] 1 O NONE 2 [J DIET COUNSELING 3 [OJ FAMILY PLANNING 4 [J MEDICAL COUNSELING 5 [J PHYSIOTHERAPY 6 0 70 8 0 OFFICE SURGERY PSYCHOTHERAPY/ THERAPEUTIC LISTENING OTHER (Specify) 15. MEDICATION THERAPY THIS VISIT WAS ANY NEW OR CONTIN- UING MEDICATION ORDERED, INJECTED, ADMINISTERED, OR OTHERWISE PROVIDED THIS VISIT? 1 O ves 2 ONO PROCEED TO ITEM 16 STOP OW ONO HEWN = ANALGESIC (Systemic) ANALGESIC (Local) ANTI-DEPRESSANT ANTI-DIABETIC ANTI-DIARRHEA ANTI-EMETIC ANTI-HISTAMINE -HYPERTENSIVE -INFECTIVE (Systemic) INFECTIVE (Local) - INFLAMMATORY 000oooooo 12 J ANTI-NAUSEA PHS T-478 Rev. 5/79 be most efficiently and economically classified and coded. enter entre DEPARTMENT OF HEALTH, EDUCATION AND WELFARE 130) 1a. 15 16 OJ 17 0 18 OJ 19 0 200 210 220 2z 0] 24 16. THERAPEUTIC EFFECTS DESIRED [Check all that apply] ANTI-PARASITIC ANTI-PYRETIC 2s [J 26 OJ ANTI-SPASMODIC zz LC] ANTI-TUSSIVE 28 APPETITE SUPPRESSANT 29 [J CARDIAC EFFECT 30 [] CONTRACEPTIVE, ORAL DECONGESTANT 3 J DESENSITIZATION 32] DIURETIC i EXPECTORANT FUNGICIDE HORMONE, SEX HORMONE, THYROID HYPNOTIC (Sleep producing) IMMUNIZATION LAXATIVE SEDATIVE OR TRANQUILIZER VASCULAR EFFECT VITAMIN OTHER (Specify) PUBLIC HEALTH SERVICE NATIONAL CENTER FOR HEALTH STATISTICS Figure 2. Patient Record, Version A, 1980 The first task was undertaken in July 1979 by the NAMCS contractor, The National Opinion Research Center. A small-scale feasibility study was designed to during the patient visit? investigate two major questions: 1. Were physicians willing and able to report more specific information about medications prescribed 0.M.B. #68-R1498 2. What were the advantages of open-ended (ver- batim) recording of drugs versus a more modest extension of the drug inquiry, expressed as a closed checklist of therapeutic effects desired by the use of the drugs? To put these questions to empirical test, two ex- perimental forms of the Patient Record were con- structed. Version A (figure 2) had a precoded format on which physicians were first asked whether any 1. DATE OF VISIT ASSURANCE OF CONFIDENTIALITY — All information which would permit identification of an individual, a practice, or an establishment will be held confidential, will be used only by persons engaged in and for the purposes of the survey and will not be disclosed or released to other persons or used for any other purpose. | 8 311319 PATIENT RECORD NATIONAL AMBULATORY MEDICAL CARE SURVEY Mo. Day Yr. 2. DATE OF 3. SEX (4, COLOROR |B ETHNICITY |, WAS PATIENT | 7, PATIENT'S COMPLAINT(S), SYMPTOM(S), BIRTH B RACE REFERRED OR OTHER REASON(S) FOR THIS VISIT 1 0 WHITE 1 0 HISPANIC THIS in patient’ GE 2 [0 BLACK o ORIGIN dd . [in patient's own words] 3 [J ASIAN/PACIFIC . MOST Mo. I Day Ivr. | ' oremace | °0 ISLANDER 20 Not PHYSICIAN? * IMPORTANT 200MALE | 40) AMERICAN INDIAN/ HISPANIC 10 ALASKAN NATIVE o os b. OTHER 2 8. MAJOR REASON 9. FOR THIS VISIT [Check One) 1 [J ACUTE PROBLEM 2 [0 CHRONIC PROBLEM, ROUTINE [Check all ordered or provided) 8 OEKG 9 [J VISION TEST 1 O NONE 2 OJ LIMITED HISTORY/EXAM 7 0 BLOOD PRESSURE CHECK WELL BABY, ETC.) DIAGNOSTIC SERVICES THIS VISIT 3 CJ GENERAL HISTORY/EXAM 10 OJ ENDOSCOPY 10, PHYSICIAN'S DIAGNOSES 11. HAVE YOU SEEN PATIENT a. PRINCIPAL DIAGNOSIS/ PROBLEM BEFORE? ASSOCIATED WITH ITEM 7a avs 20NO 30] GHAONIC PROBLEM, 4 CPAP TEST 11 CJ MENTAL STATUS IF YES, FOR THE 4 0] POST SURGERY/ EXAM CONDITION IN ok 5 CY CLINCAL LAB TEST 5" OTHER SIGNIFICANT CURRENT CONDITIO 50) NON-ILLNESS CARE ‘ (ROUTINE PRENATAL, | 6 0 X-RAY 12 [J OTHER (Specity) 10YEs 20NO GENERAL EXAM, 12. MEDICATION THERAPY THIS VISIT O NONE Include immunizing and desensitizing agents] a. FOR PRINCIPAL DIAGNOSIS i [Using brand or generic names, record all new and continued medications ordered, injected, administered, or otherwise provided. b. FOR ALL OTHER REASONS 1. 2 2 3. 3. 4 4 13. NON-MEDICATION THERAPY 14, DISPOSITION THIS VISIT 15. DURATION [Check all services ordered or provided this visit] [Check all that apply) OF This 1 0 NO FOLLOW-UP PLANNED (Tim : y 1 ONONE 6 OJ OFFICE SURGERY 2 0 RETURN AT SPECIFIED TIME actually ih 2 CI DIET COUNSELING 7 0 PSYCHOTHERAPY/ 30 RETURN IF NEEDED, P.R.N. $per) " t 3 [J FAMILY PLANNING THERAPEUTIC LISTENING 0 TELEPHONE FOLLOWUP PLANNED physician} 4 DIMEDICAL COUNSELING 6 CJ RETURNED TO REFERRING 5 CJ PHYSIOTHERAPY 8 [J OTHER (Specify) PHYSICIAN 7 0 ADMIT TO HOSPITAL MINUTES 8 C OTHER (Specify) DEPARTMENT OF HEALTH, ATION AND WELFA ig Lars PUBLIC HEALTH SERVICE O:M.8, fea Rie NATIONAL CENTER FOR HEALTH STATISTICS Figure 3. Patient Record, Version B, 1980 medication was ordered or provided. If the answer was yes, they were asked to check off the therapeutic effect(s) desired from a list of 33 precoded response categories. Version B (figure 3) employed an open- ended format that required the physician to list ver- batim all new or continued medications prescribed during the visit, using either generic or brand names, and recording nonprescription as well as prescription drugs. The study design called for two matched samples of 20 physicians each, one sample assigned Version A, 4 the other Version B. The assignment was random. The samples were selected from physicians who had recently participated in NAMCS. Since these physi- cians were already comfortable with NAMCS proce- dures and with the content of its Patient Record, it was felt that they could better direct their critical attention to the new medication item without suffer- ing the possibly negative distraction which might—in newly recruited physicians—be evoked by new pro- cedures, by other items on the Patient Record, or both. The sample was dispersed geographically and in- cluded all the major specialties. Each sampled physi- cian was asked to fill out 10 Patient Records a day for 3 days. Following the 3-day period, interviewers visited the doctor’s office to pick up the completed forms and to conduct a brief evaluation interview. This interview probed for reactions to the medication items with particular reference to any difficulties ex- perienced in obtaining and recording the data. During the interview, physicians were shown the alternative version of the Patient Record (precoded checkbox or open-ended write-in) and asked to indicate a prefer- ence. The findings of the evaluation interviews are tabulated in appendix I. Finally, respondents were asked whether in their judgment any substantial num- ber of physicians might not be willing to provide this kind of drug information. Patient Records and completed evaluation inter- views were received from 35 of the 40 physicians. In the sample that used Version A, 17 physicians partici- pated, producing 407 Patient Records. In the sample that used Version B, 18 physicians participated, pro- ducing 384 Patient Records on which a total of 706 drug entries were listed. None of the five nonrespond- ents demurred because of a negative reaction to the drug item. Summarizing the results of the evaluation inter- views, it may be said that physicians had little or no difficulty using either version of the form, expressed confidence that they had included all prescription drugs ordered or provided (but thought they may have missed some nonprescription drugs), and had no strong preference for either the precoded checklist or the open-ended write-in. It deserves emphasis that there was absolutely no overt refusal to participate in the study or criticism of the medication items on the grounds that the infor- mation would be used to evaluate the quality of the doctor’s practice, monitor Medicare or Medicaid billings, or otherwise be used to the physician’s disad- vantage. However, though they themselves expressed no concerns, 10 of the 35 pretest respondents did raise possible concerns that—they conjectured—other physicians might have, for example: “Maybe if they’re handing out amphetamines.” “Certain drugs are not recognized for the condi- tion on the Patient Record. If they’re using those, they might not want to say so.” “Some don’t like to admit the number of drugs they are prescribing.” “They might fear it will open the door to the gov- ernment, but probably not unless they’re doing something unethical. There may be abuses they do not want to reveal.” “Some prescribe diet and pain pills in large quan- tities.” “It could be used by government agencies as a means of review.” “If the doctor has not kept up with current trends he wouldn’t want to reveal it.” The contractor for the study put his conclusions and recommendations in these words. “In our opinion, this pretest has demon- strated the feasibility of obtaining data on medi- cations ordered or provided to ambulatory care in physicians’ offices. We encountered no refusal to participate in the survey because this item was included, no specific objections to recording the information, and no immediate visible evidence that responses were distorted. The fact that ten of the participating physicians could think of reasons why other physicians might object to recording the information is not surprising to us; the speculation is a plausible one. But we have no evidence that such attitudes would significantly affect the data. “Either the checklist type of item or the open-ended write-in seems quite feasible and agreeable to doctors. The open-ended version will be more costly, since it requires office coding in one or more dimensions, but the choice would seem to depend upon which type of information will be most useful to NAMCS users. “With respect to ease of recording and quality of the data, we would recommend using write-ins rather than the checklist. Although the latter may seem simple at first glance, the checklist has several difficulties. It takes time for the physician to acquaint himself with the numerous categories; many of these categories may be entirely irrele- vant to his practice and thus prove confusing and distracting; he must use his judgment to deter- mine in which category to place the particular drugs he prescribes and the judgment may be difficult because the category may be too general or too restrictive. Uncertainty about the coding of the therapeutic effect desired is compounded when a nurse or other assistant completes all or part of the patient record instead of the doctor. Using the open-ended version of the form, the assistant could simply copy the name of the drug from the medical record; using the precoded version, the assistant might not be able to decide correctly the effect desired by the physician. “We had thought that the write-in method would be more burdensome to the doctor, but as we have seen the majority of those who used this version preferred it as quicker and simpler than coding onto the list of desired effects. We had also thought that the write-in method might produce fewer mentions, since the precoded list could be expected to jog the physician’s memory, 5 and since the task of writing out the name of a drug might be expected to inhibit mention. Instead, we found an average of 1.8 medications written in per patient, compared to an average of only 1.6 categories checked. (It may be, of course, that physicians sometimes ordered two drugs to achieve one type of effect.) “The more specific and objective nature of the brand or generic name information appears to provide another advantage over the vaguer and more subjective checking of desired effects. The name of the prescribed drug is or should be unambiguous and will normally appear on the physician's record, where it is accessible to his assis- tant or to the doctor himself if he fills out the pa- tient record at a later date. “It is also subject to clarification or checking by the interviewer in the course of her edit of the form, If a drug is unfamiliar or recorded illegibly, she can verify the entry with the physician’s nurse or administrative assistant. The checklist of ef- fects is practically immune to verification or edit. “The reporting of nonprescription medication caused some problems and the need for this information could well be reviewed. Such medica- tions are not listed on the record in many cases and recommendations for their use may have been given informally on past occasions, as well as in the course of the reported visit. Accuracy and completeness of recall are unknown. Almost half of our pretest physicians were not sure they had reported all of these, whereas three-quarters were confident they had fully reported all prescription drugs ordered or provided.” Concurrent with the planning and conduct of the pretest, NAMCS staff contacted 40 health-care authorities, asking which version of the Patient Record they preferred—write-in of drug names or checklist of therapeutic effects—and inviting their comments. Among those contacted were: members of the NAMCS technical advisory panel, representatives of the 18 professional organizations that endorsed NAMCS, interested educators, and officials within the National Center for Health Statistics. Thirty-five of the contacts replied. Of these, 25 favored the write-in version, 5 the checklist, and 5 expressed no preference. Evidence from this informal elicitation of opinion, along with the recommendation from the Feasibility Study, pointed up the richer potential of useful information implicit in the write-in version, confirming NAMCS planners growing determination to use this approach. (The finished format of the 1980 Patient Record, closely modeled on Version B of the pretest, appears as figure 4.) However, espe- cially among authorities sophisticated in coding prob- lems, the caution was frequently expressed that the larger the number of drug dimensions explored, the more complex and expensive the classification and coding process would certainly become, thus, re- quiring a very specific forecasting of the uses to which the drug data would be put. ASSURANCE OF CONFIDENTALITY=AIl information which would permit identification of an individual, a practice, or an establishment will be held confidential, will be used only by persons engaged in and for the purposes of the survey and will not be disclosed or re- leased to other persons or used for any other purpose. Guava Ea a toy DNo699924 Office of Health Resserch, Statistics, and Technology National Center for Mesith Statistics LOG cord name and elow. For the #5, also com- TIME OF VISIT ATIENTS E 1 . DATEOF VISIT Month Day Year PATIENT RECORD NATIONAL AMBULATORY MEDICAL CARE SURVEY o the right. 2, OATEOF 13, sex dd] Don Month Day Yeer 4 COLOR OR RACE 3 ETHNICITY . . 1 [Jwwire 1 HISPANIC 2[JsLack O Naan a [[]asianeaciFic 0 ISLANDER ® Over ANIC « [Jamenican iNoian/ ALASKAN NATIVE @. PATIENT'S COMPLAINTI(S), SYMPTOM(S), OR OTHER REASON(S) FOR THIS VISIT [In patient's own words] a. MOST IMPORTANT b. OTHER y MAJOR REASON FOR THIS VISIT [Check one] 1 [J acute rrosLem 2[] cHronic poBLEM, ROUTINE 3 [J cumonic prosLEm, FLAREUP 4 [TJrosT sunaeny/rosT INvuRY 5 [CJ NON-ILLNESS CARE (ROUTINE PRENATAL, GENERAL EXAM, WELL BABY, ETC.) 8 DIAGNOSTIC SERVICES THIS VISIT ® [Check all ordered or provided | 1 [none os [Joka 2 [Jumireo mstomviexam, » [] vision Test. 3 [Jaenenac Historv/exam, 10 [_] enooscory « [Jrar vest un] MENTAL STATUS s [Jeunica Las Test 12 [J OTHER (Speers) 6 O X-RAY 7 [swoop pressune cHeck 9. PHYSICIAN'S DIAGNOSES a, PRINCIPAL DIAGNOSIS/PROBLEM ASSOCIATED WITH ITEM 6a, b., OTHER SIGNIFICANT CURRENT DIAGNOSES 10, HAVE YOU SEEN * PATIENT BEFORE? ip 2[]no IF YES, FOR THE CONDITION IN ITEM Be? 1[Jves 2 Jno 111. MEDICATION THERAPY THIS VISIT a, FOR PRINCIPAL DIAGNOSES IN ITEM Sa. 1 2. 3 4, CINONE [Using brand or generic names, record all new and continued medications ordered, injected, administered, or otherwise provided at this vist. Include immunizing and desensitising agents) D. FOR ALL OTHER REASONS. 1. 2 3. 12. NON-MEDICATION THERAPY 13. WAS PATIENT 14 DISPOSITION THIS VISIT 18 DURATION [Check all services ordered or provided this visit] REFERRED ® [Check all that apply | ® OF THIS . FOR THIS VISIT VISIT 8y ANOTHER 1 [Jno FoLLow-up PLANNED [Time actually 1 Od NONE 6 OJ DIET COUNSELING PHYSICIAN? 2 [Jretuan AT SPECIFIED TIME Sent i physician 2 []pHvsioTHERAPY [0] FAMILY/SOCIAL 3 [rerun i NeeoeD, PAN. COUNSELING 3[ J orice surgery . O s [J meoicat counseLing [ves o [[] reLEPHONE FOLLOW-UP PLANNED + [JramiLy panning 8 []ReFERRED TO OTHER PHYSICIAN s [J esvcHoTmenapy/ *Jormenstrin EY CHOTHERARY TUNING 20 ¢ [[JRETURNED TO REFERRING PHYSICIAN 7 [JaomiT To HospiTAL ; Minutes o [John (speciry) PHS-8108-D (9/79) OMB No. 68-R1498 Instruction on patient record jacket 11. MEDICATION THERAPY THIS VISIT: a. FOR PRINCIPAL DX IN ITEM 9.a.: List all prescrip- tion and non-prescription drugs and medications ordered or provided at this visit. Include drugs pre- scribed at a previous visit if the patient was in- structed, at this visit, to continue the medication. b. FOR ALL OTHER REASONS: Apply instructions in Item 11.a. for all reasons other than the principal diagnosis. Figure 4. Patient Record and Instruction on Patient Record jacket Drug identification and description A central issue, then, is how to identify and describe drugs in order to characterize prescribing in office-based practice. The question is a complex one because substantial variability exists in the detail with which drugs are commonly identified. In order to dispense a drug to a patient, there must be an instruction by the physician and a decision by the pharmacist to select a unique drug entity. For example, the medication received by a patient may be described as follows: AMPICILLIN 250 mg Capsule 1 QID 100 4 4 4 so 4 Drug Name Drug Strength Dosage Form Dosage Regimen Number of Doses 1m 2) 3) “@ (5) These five units of information can be expanded by adding the price (or cost) of the medication, pack- age size used by the pharmacist, manufacturer, or other information. Quite obviously the task of re- cording drug prescribing activities can become ex- tremely complex if the intent is to reflect all decision- making that occurs in the prescribing process. Another complicating factor is that, as the depth of information increases, so does the breadth of infor- mation. In the above example five basic and several supplemental units of information were identified that add information to an AMPICILLIN drug order. At this very explicit level of description there are over 50,000 unique drug entities available in the marketplace. That is, an AMPICILLIN prescription for a capsule would be considered different from an oral suspension, and so forth. If each drug order were to be described in complete detail the resource re- quirements would be unmanageable. The need to balance cost and utility considerations quickly leads to a parsimonious approach in data collection. Because complete capture of prescribing informa- tion is not feasible, there is a need to establish deci- sion rules for defining minimum essential informa- tion. Three guidelines were established: 1. The data base should reflect, insofar as possible, the prescriber’s primary intent in ordering drug therapy. 2. Physical characteristics of the product (e.g., dos- age form, strength, package size) are of minimal importance. 3. Identification procedures must be restricted to the information commonly included by physi- cians in ordering drugs. These three guidelines dictated that only the “core identifier” (e.g., AMPICILLIN) could be used in data collection. The core identifier is the prescrib- er’s designation of the pharmaceutical agent(s) used for cure, treatment, diagnosis, or prevention of disease. Information concerning dosage form, strength, and other medication characteristics is ex- traneous to this primary intent. Further, the physi- cian commonly omits physical product and dosage regimen information when recording drug orders. The data collection objective then becomes one of ex- tracting the maximum amount of useful information from the core identifier. A valuable unit of information available from the core identifier is simply the frequency with which drugs are used in office-based practice. Since drug therapy is the most frequently employed treatment modality in medical care there is a need to under- stand the rates of drug use. For example, the average number of prescriptions per office visit, and charac- teristics of drug users (e.g., age, sex, race) versus nonusers are measures that would be useful in planning public or private health services programs. The NAMCS drug data base would be a valuable tool if it captured the number of drug orders for a known and representative population of prescribers and patients. A second major information need has to do with not only how many, but what general types of drugs are used in office-based practice. Private insurance programs, public funding agencies, hospitals, Health Maintenance Organizations, and other organizations need to assure that patients have adequate access to comprehensive pharmaceutical care. Only by knowing what drugs are prescribed for what purposes—in gen- eral therapeutic categories, not specific diseases—can this need be met. The NAMCS drug data base can assist by documenting the different types of drugs (e.g., antibiotics) as well as different products (e.g., AMPICILLIN) prescribed in office-based practice. A third need is to assist in the clarification and evaluation of public policy issues in pharmaceutical care. Extensive discussion has occurred during the past decade about costs and merits of brand name prescription drugs versus a (usually) less costly gen- eric drug. If the physician prescribes a drug by its “family” or generic (e.g., AMPICILLIN) name the pharmacist can select any manufacturer’s product approved by the Food and Drug Administration. If the physician provides a specific manufacturer’s prod- uct (e.g., AMCILL) the pharmacist may or may not be able to dispense a different manufacturer’s prod- uct, depending upon State law, regulation, and professional standards of practice. In order to meas- ure the potential impact of drug substitution, it is necessary to characterize prescribing according to brand or generic designation. Drugs are subdivided into two major classifica- tions—prescription (legend) and nonprescription (over the counter, or OTC) drugs. Prescription drug use represents the judgment of physicians concerning a person’s medical care needs. Over-the-counter drug use represents a much greater reliance on self-care by the consumer. In many cases there are also significant cost differences between prescription and nonpre- scription drugs used for the same purpose, e.g., ASPIRIN or DARVON used for analgesic effect. Further, since some over-the-counter drugs are ex- tremely important in medical care (e.g., ASPIRIN for fever and pain, INSULIN for diabetes, antacids for peptic ulcer) this information can be useful in plan- ning health education programs. In order to under- stand and support the role of self-care in drug therapy, the NAMCS drug data base should distinguish between prescription and nonprescription drugs. An extremely important issue in health and social policy is the use of medications having significant po- tential for addiction and/or habituation. Potential for diversion into illicit channels, need for adequate regu- latory control, and risk of adverse treatment events all argue for complete documentation of narcotics and related substances. Public policy now places regulatory control of these substances in the Depart- ment of Justice, specifically the Drug Enforcement Agency (DEA). The DEA has placed drugs into five categories of control, depending upon therapeutic use and potential for addiction or habituation. Classifica- tions and their criteria are: Schedule I. The drug or other substance has a high potential for abuse, it has no currently ac- cepted medical use in treatment in the United States, and there is a lack of accepted safety for use under medical supervision. Schedule II. The drug or other substance has a high potential for abuse, it has a currently ac- cepted medical use in the United States, and abuse may lead to severe psychological or physi- cal dependence. Schedule III. The drug or other substance has a potential for abuse less than drugs in Schedules I and II, it has a currently accepted medical use in the United States, and abuse may lead to moder- ate or low physical or high psychological depend- ence. Schedule IV. The drug or other substance has a low potential for abuse relative to drugs or other substances in Schedule III, it has a currently ac- cepted medical use in the United States, and abuse of the drug may lead to limited physical or psychological dependence relative to drugs in Schedule III. Schedule V. The drug or other substance has a low potential for abuse relative to drugs or other substances in Schedule IV, it has a currently ac- cepted medical use in the United States, and abuse of the drug may lead to limited physical or psychological dependence relative to drugs in Schedule IV. Any attempt to describe prescribing patterns of office-based physicians must be able to classify drugs according to these properties. Finally, an issue of longstanding debate in drug policy concerns the use of drugs in fixed combina- tions, as opposed to single ingredient entities. As an example, a common approach in treating hyperten- sion is to prescribe a diuretic agent (e.g., HYDRO- CHLOROTHIAZIDE 50 mg tablets) and an antihy- pertensive agent (e.g., RESERPINE 0.125 mg tablets). The physician can do this by ordering two separate prescriptions, or by ordering a fixed combination of the two products (e.g., HYDROPRES-50). The latter approach is usually more costly and offers less flexi- bility in dosage adjustment, however, it offers greater potential convenience to the patient. The extent to which fixed combination products are used is an im- portant consideration in planning drug benefits in prepaid health plans. Also, this information is useful in the development of reimbursement policies for publicly sponsored programs. Data collection efforts should therefore be able to infer single or combina- tion ingredient status, and the specific ingredient(s) (e.g., HYDROCHLOROTHIAZIDE and RESERPINE) from core identifiers (e.g., HYDROPRES-50) used by physicians to order drugs. To summarize, the preceding discussion has indi- cated that the “core identifiers” employed by physi- 9 cians in ordering drugs must be used to provide addi- tion status, Federal control schedule, and single or tional data. Essential units of information to be combination drug entities. The following example obtained from the core identifier include frequency may be helpful in interpreting a typical medication of separate medication entry, therapeutic category, entry (see figure 4, item 11): generic or brand name, prescription or nonprescrip- 11 Medication Therapy This Visit None O [Using brand or generic names, record all new and continued medications ordered, injected, administered, or otherwise provided at this visit. Include immunizing and desensitizing agents. ] a. For Principal Diagnosis in Item 9a b. For All Other Reasons 1. EMPIRIN #3 1. 2. AMPICILLIN 2 3 3 4 4. This core identifier can be extrapolated to yield the following information: Number of Separate Medication Entries 2 Detailed Information for Entry 1. (EMPIRIN #3) Therapeutic category Analgesic Generic or brand Name Brand name Prescription or nonprescription status Prescription Federal control status III Single or combination ingredient Combination Specific ingredient(s) CODEINE PHOSPHATE ASPIRIN Detailed Information for Entry 2. (AMPICILLIN) Therapeutic category Antibiotic Generic or brand Name Generic Prescription or nonprescription status Prescription Federal control status None Single or combination ingredient Single Specific ingredient(s) AMPICILLIN This example shows how two core identifiers (AMPI- based physicians’ practices. What is needed next is a CILLIN and EMPIRIN #3) can be expanded to pro- complete coding system that records and classifies vide detailed information on drugs used in office- core identifiers into the desired units of information. 10 Classifying and coding device The task that now confronted NAMCS planners was that of finding or—in whole or in part—creating a system by which the drugs and their dimensions could be most readily, efficiently, and economically identified, coded, and copied into the machine read- able format already in use for the other variables of the NAMCS reporting system. The problems of creating—manually and from zero base—such a multidimensional coding device were too formidable to consider seriously. No single source book contains all the drug dimensions, let alone appropriate codes for each. At a minimum, two books are needed, for example, the current editions of the American Drug Index? and the Drug Topics Red Book.b The information in these sources would— with most dimensions—require the creation of unique coding schemes. Furthermore, coding from multiple sources by manual search and selection requires ad- vanced pharmaceutical expertise, costs dearly in time and effort, and is dogged by the possibility of human error. Consider the sheer volume of drug information to be found and coded for NAMCS 1980. NAMCS planners estimated that sampled physicians would make roughly 54,000 drug entries (the actual figure for 1980 was 51,372). If, along with identifying and coding each specific entry, NAMCS coders were re- quired to code six additional drug dimensions by manual search and find, between 350,000 and 400,000 coding decisions might be involved—a stag- gering prospect. Clearly, the best solution to these complex prob- lems lay in finding a computerized source, one that would provide—in machine readable and printable form—an exhaustive inventory of the drugs antici- pated to be prescribed in office-based ambulatory care, along with precoded classification formats for aBillups, Norman F. and Billups, Shirley M.: The American Drug Index 1979; J. P. Lippincott Company, Philadelphia 1979. bDrug Topics Redbook 1979; Medical Economics Company (Litton), Oradell, N.J. 1979. each of the drug dimensions that NAMCS wished to explore. With these demands in mind, NAMCS planners considered and selected the Drug Product Informa- tion File (DPIF), a computer-processable data base of information on more than 30,000 commercially avail- able drug products. Developed and maintained under the auspices of the American Society of Hospital Pharmacists, the DPIF is continually updated to add new products when they are marketed and to with- draw products when they are no longer available. Drug products are described in a fixed-field format in which 68 fields are used to record a broad range of drug information, including information on all the drug dimensions desired for NAMCS needs. The fields applicable to NAMCS usage are shown in table A and they are underlined. (Note: All DPIF materials are re- produced with the permission of the American Society of Hospital Pharmacists.) NAMCS planners now proceeded to construct their own computerized drug file by adopting or adapting relevant fields from the DPIF. The result was the NAMCS Drug File; its format appears as table B. Its data fields are described as follows. Field #01—National Drug Code.—The National Drug Code (NDC) is a 10-character identification code for a drug product and package. It is maintained by the U.S. Food and Drug Administration. One of two linkage codes in the NAMCS Drug File, it is bor- rowed without modification from the DPIF to permit possible linkages with drug classification systems other than the DPIF. ¢ Field #02—Brand Name.—The Brand Name (BRAND NAME) is the name under which the drug product is marketed and may or may not be a trade- mark. Note that this field also contains generic names if the products are marketed using generic names. The spelling of the brand name is complete in most instances. If the name exceeds 40 characters, abbrevi- ations may be used. The DPIF for June 1980 (the time of its acquisi- 1 Table A. DPIF record format (Record length is 480 characters) Field Field name (column heading) Fi i — Fre 01 National Drug Code (N, NPROD, NP) 10 001 010 X 02 Brand Name (BRAND NAME) 40 011 050 X 03 Generic Name (GENERIC NAME) 40 051 090 X 04 Dosage Form (DOSAGE FORM) 25 091 115 A 05 Route of Administration/Use (ROUTE/USE) 10 116 125 A 06 Strength Number (STR/NUMBER) 1 126 136 9 07 Strength Unit (ST/UNIT) 7 137 143 X 08 Package Size (P-NO, PU) 8 144 151 X 09 Package Description (PKG/DESCRIPT) 12 1562 163 X 10 Labeler (LABELER) 9 164 172 A 11 AHFS Inclusion (F) 1 173 b 12 Unit Packaging (H) 1 174 X 13 Compendial Status (C) 1 175 b 14 Legal Status Code (L#) 2 176 177 9 15 Labeler Code (LBLR#) 5 178 182 9 16 Record Type Code (T) 1 183 9 17 Generic Name Code (GN#) 5 184 188 9 18 Dosage Form Code (D#) 2 189 190 9 19 Strength Code (S#) 2 191 192 9 20 Package Code (P#) 2 193 194 9 21 Brand Drug Product Package Code (BDPP#) 5 195 199 9 22 Route of Administration Code (R#) 2 200 201 9 23 Therapeutic Category Code (AHFS#) 6 202 207 9 24 Generic Drug Product Code (GDP#) 5 208 212 9 25 Blank 4 213 216 b 26 Brand Drug Product Code (BDP#) 6 217 222 9 27 Volume (VOL) 3 223 225 X 28 Ingredient Code (1#) 5 226 230 9 29 Ingredient Strength Number (ISN) 1 231 241 9 30 Ingredient Strength Unit (ISU) 3 242 244 A 31 Ingredient Code (1#) 5 245 249 9 32 Ingredient Strength Number (ISN) 1 250 260 9 33 Ingredient Strength Unit (ISU) 3 261 263 A 34 Ingredient Code (1#) 5 264 268 9 35 Ingredient Strength Number (ISN) 1 269 279 9 36 Ingredient Strength Unit (ISU) 3 280 282 A 37 Ingredient Code (1#) 5 283 287 9 38 Ingredient Strength Number (ISN) 1" 288 298 9 39 Ingredient Strength Unit (ISU) 3 299 301 A 40 Ingredient Code (1#) 5 302 306 9 41 Ingredient Strength Number (ISN) 11 307 317 9 42 Ingredient Strength Unit (ISU) 3 318 320 A 43 Ingredient Code (1#) 5 321 325 9 a4 Ingredient Strength Number (ISN) 1 326 336 9 45 Ingredient Strength Unit (ISU) 3 337 339 A 46 Ingredient Code (1#) 5 340 344 9 47 Ingredient Strength Number (ISN) 11 345 355 9 48 Ingredient Strength Unit (ISU) 3 356 358 A 49 Ingredient Code (1#) 5 359 363 9 50 Ingredient Strength Number (ISU) 1 364 374 9 51 Ingredient Strength Unit (ISU) 3 375 377 A 52 Ingredient Code (1#) 5 378 382 9 53 Ingredient Strength Number (ISN) 1 383 393 9 54 Ingredient Strength Unit (ISU) 3 394 396 A 55 Ingredient Code (1#) 5 397 401 9 56 Ingredient Strength Number (ISN) 1 402 412 9 57 Ingredient Strength Unit (ISU) 3 413 415 A 58 Ingredient Code (1#) 5 416 420 9 59 Ingredient Strength Number (ISN) 1 421 431 9 60 Ingredient Strength Unit (ISU) 3 432 434 A 61 Ingredient Code (1#) 5 435 439 9 62 Ingredient Strength Number (ISN) 1 440 450 9 63 Ingredient Strength Unit (ISU) 3 451 453 A 64 Secondary Package Size (P2-NO) 5 454 458 9 65 Secondary Package Description (P2-DESC) 10 459 468 X 66 Maximum Allowable Cost (MAC, E) 6 469 474 X 67 Replaced/Replacing Record Reference (REFER) 5 475 479 9 68 Update Flag 1 480 A *9 = Numeric, A = Alphabetic, X = Alphanumeric, b = blank. 12 Table B. NAMCS drug file format Field Field name Field First Final Field length position position class* 01 National Drug Code (NDC) 10 001 010 AN 02 Brand Name (BRAND NAME) 40 011 050 AN 03 Brand Drug Product Package Code (BDPP#) 5 051 055 N 04 Generic Name (GENERIC NAME) 40 056 095 AN 05 Generic Name Code (GN#) 5 096 100 N 06 Medication Code List Name (MCL NAME) 40 101 140 AN 07 Medication Code List Code (MCL#) 5 141 145 N 08 Entry Status Code (E#) 1 146 N 09 Prescription Status Code (P#) 1 147 N 10 Federal Controlled Substance Status Code (DEA#) 1 148 N 1 Composition Status Code (C#) 1 149 N 12 Therapeutic Category Code (AHFS#) 6 150 155 N 13 Ingredient Code (1#) 5 156 160 N 14 Ingredient Code (1#) 5 161 165 N 15 Ingredient Code (1#) 5 166 170 N 16 Ingredient Code (1#) 5 17 175 N 17 Ingredient Code (1#) 5 176 180 N *AN = Alphanumeric, N = Numeric. tion by NAMCS) contained nearly 35,000 records in the brand name field. To adapt the field for NAMCS needs, several operations were undertaken. First, a computer run was made that unduplicated the DPIF records by selecting a single record for those drug products with the same name but with multiple label- ers. This run reduced the number of brand name rec- ords to about 9,700. Second, this time by manual inspection and deletion, certain differentiating infor- mation on dosage strength was eliminated as irrele- vant to NAMCS needs. Thus for the following three DPIF records: APAP W/CODEINE APAP W/CODEINE 15 MG APAP W/CODEINE 30 MG only the record APAP W/CODEINE was retained, an operation that further reduced the NAMCS brand name field to about 6,800 records. Finally, the re- maining brand names were modified in most instances to agree with the nomenclature found in the “Index” of the looseleaf, two-volume publication of the American Hospital Formulary Service.© The brand names, as modified, then were used to form the core of the 1980 NAMCS Medication Code List (see the discussion of field #06). Field #3—Brand Drug Product Package Code.—In the DPIF, the Brand Drug Product Package Code (BDPP#) identifies a labeler’s package of a brand of a particular drug product. Each record in the DPIF has a unique BDPP#. One of two linkage codes in the NAMCS Drug File, the BDPP# is retained without modification to permit linkages back to the ap- plicable records in the original DPIF. CAmerican Hospital Formulary Service; American Society of Hospital Pharmacists, Inc., January 1980. Field #4—Generic Name.—The generic name (GENERIC NAME) is the (public, scientific, nonpro- prietary, established) name as assigned by the United States Pharmacopeia or other responsible authorities. The spelling of the generic name is complete in most instances. If the name exceeds 40 characters, abbrevi- ations are used. In the NAMCS Drug File, the records in the DPIF generic name field have been modified to agree in nomenclature with the generic forms used in the current volumes of the American Hospital Formulary Service. Thus the DPIF “BROMPHENIR- AMINE MALEATE” becomes the NAMCS ‘“BROM- PHENIRAMINE.” A further modification—and resultant overall re- duction—was then undertaken, which aimed at selec- tively replacing certain product-specific generics in the DPIF by the more functional, utilization-oriented configurations needed by NAMCS. For example, the following DPIF generics IODINE (ETHIODIZED OIL) IODIDES TINCTURE IODINE TINCTURE IODINE (POLAXAMER-IODINE) IODINE (IODOPHOR) SODIUM IODIDE IODINE (PROVIDONE-IODINE) were grouped for NAMCS Needs into a single generic category “IODINE TOPICAL PREPARATIONS.” The final result of these adaptations was the NAMCS Inventory of Generic Names, which appears in appendix II. Selected names from this inventory were added to the 1980 Medication Code List to insure a more complete representation of generic- entry choices. (See the discussion of field #06.) Note: Drugs containing more than one active ingredient generally have “COMBINATION PROD- UCT” as the GENERIC NAME. Some common 13 combination drugs have been assigned a generic name; e.g., APC/PROPOXYPHENE. Field #05—Generic Name Code.—A Generic Name Code (GN#), created for and unique to the NAMCS Drug File, is assigned to each generic name in field #04. The following codes are used: GN# 50005 — 56300 generic names in field #04 50000 generic name undetermined See also appendix II and the discussion of ingredient codes to follow. Field #06—Medication Code List Name.—The Medication Code List Name (MCL NAME) is the name of a drug as it appears on the NAMCS Medica- tion Code List, an alphabetized inventory of single- source and multiple-source drugs for use in coding the entries on the NAMCS Patient Records (see appendix III). The core constituents of the list are the brand names in field #02. Selected generic names from field #04 have been added to insure a more complete representation of generic-entry choices. The spelling of the MCL name is complete in most instances. If the name exceeds a fixed field of 40 characters, abbreviations are used. Field #07—Medication Code List Code— A Med- ication Code List Code (MCL#), unique to the NAMCS Drug File, is assigned to each MCL name in field #06. The serial order of the MCL codes parallels the alphabetic order of the MCL names (see appen- dix V). Field #08—Entry Status Code.—The Entry Status Code (E#) is unique to NAMCS, denoting the nature of the entry that the physician makes on the Patient Record. Every item on the Medication Code List is assigned an entry status. The following codes are used: E# Entry Status 1 Generic name—(About 17 percent of the coding choices on the 1980 MCL are generic (multiple-source) entries.) 2 Brand (nongeneric) name—The use of “brand” to define an entry status is restricted to nongeneric (single-source) entries. (About 82 percent of the cod- ing choices are brand-name entries.) 3 Therapeutic effect—This code is used for the 95 coding choices where the phy- sician did not specify a brand or gen- eric medication, stating only the therapeutic effect desired; e.g., laxa- tive, analgesic, etc. 4 Other—This code applies to entries which, though identifiable as brand or 14 generic names or as therapeutic ef- fects, are not in the MCL. 5 Illegible Field #09—Prescription Status Code.—The Pre- scription Status Code (P#) is derived from the DPIF Legal Status Code (L#), which is used to indicate the Federal legal classification of drug products. The fol- lowing codes are used by NAMCS: P# Prescription Status 1 Prescription (Rx) drug—(About 60 per- cent of the coding choices on the 1980 MCL are prescription drugs.) 2 Nonprescription (OTC) drug—(These drugs represent about 36 percent of the coding choices.) 3 Undetermined Field #10—Federal Controlled Substance Status Code.—The Federal Controlled Substance Status Code or Drug Enforcement Administration Code (DEA#) is derived from the DPIF Legal Status Code. It denotes the degree of potential abuse and Federal control of a drug. The following codes are used by NAMCS: DEA# Federal Control Schedules 1 Schedule I (LSD, HEROIN, MARI- JUANA —research only) 2 Schedule II (MORPHINE, DEMEROL, AMPHETAMINES —most abused) 3 Schedule III (APC/CODEINE, etc.—less abused) 4 Schedule IV (VALIUM, etc.—potential abuse) 5 Schedule V (Controlled sale by phar- macy only—Rx required—or State re- stricted) 6 No control 7 Undetermined Controlled drugs comprise about 8 percent of the coding choices on the 1980 MCL. Field #I11—Composition Status Code.—The Composition Status Code (C#) is derived from the DPIF Record Type Code (T#). It is used to distin- guish between single- and multiple-entity drugs. The following codes are used by NAMCS: C# Composition Status 1 Single-entity drug—The drug contains only one active ingredient. (About 51 percent of the coding choices on the 1980 MCL are single-entity drugs.) 2 Combination drug—The drug contains more than one active ingredient. (About 48 percent of the MCL coding choices are combination drugs.) 3 Undetermined 5 Multivitamin—A combination drug given separate identity for purposes of anal- ysis; there are 453 multivitamin cod- ing choices in the MCL. Field #12—Therapeutic Category Code.—The Therapeutic Category Code (AHFS#) is borrowed without modification from the DPIF. It identifies the pharmacologic-therapeutic category of a drug, accord- ing to the classification system of the American Hos- pital Formulary Service (AHFS). The AHFS classifi- cation is reproduced as appendix IV. Note: An AHFS classification number is conven- tionally printed with a colon between the second and third digits and a period between the fourth and fifth digits (e.g., 20:12.12). The fifth and sixth digits are omitted if they are zeros. On the NAMCS Drug File, the colon and period are not included and all digits are carried (e.g., 8:04 becomes 080400). The following codes appear on the NAMCS Drug File: AHFS# 040000 - 960000 Therapeutic category 970000 Undetermined For many drugs, especially for combination drugs, more than one therapeutic category is possible. On the NAMCS Drug File, only one AHFS# appears for each drug entry. The selection of this single AHFS# was a judgment of the DPIF staff. Field #13-17—Ingredients of Combination Drugs.—For combination drugs, the NAMCS Drug File uses ingredient codes adapted from the DPIF to identify the active, generic ingredients of the drugs. The Ingredient Code (I#) is the five-character Generic Name Code (GN#) described for field #05. Fields are available for a maximum of five ingredients; this represents a truncation of the DPIF, in which 12 in- gredient code fields are allowed. An estimated 84 per- cent of all active ingredients are identified in the five fields retained by NAMCS. The following codes are used: I# Ingredients 50005 - 56300 Generic names in field #04, NAMCS Drug File 50000 Generic ingredient undeter- mined Note: The NAMCS Drug File does not retain the DPIF ingredient codes for generic components that cannot also function as a single-entity drugs. Thus if a generic substance is identified by the DPIF as having “ingredient-status” only, e.g., ALPHA AMYLASE, it is coded by NAMCS as I# ‘“Undeter- mined.” 15 Coding process The coding of the NAMCS drug information be- gan in August 1980, 7 months into the 1980 survey year. The coding volume was roughly predictable. NAMCS findings in past years suggested that the sam- ple of office visits—expressed in completed Patient Records—would probably approximate a raw number of 50,000, and that about 60 percent of these visits (30,000) would result in some form of drug therapy. For every visit that entailed drug therapy, pretest findings indicated that about 1.8 drugs would be ordered or provided. This led to a rough prediction of 54,000 drug entries (or “mentions”) to be coded. (The final raw total was 51,372.) The coding task was performed by the National Opinion Research Center (NORC), contractor for all collection and coding operations of NAMCS. The contractor agreed to pro- vide double, independent coding of 100 percent of the first half (27,000) of the raw drug entries; 50 per- cent of the next 14,000 entries; and 25 percent of the remainder. The balance was to be single-coded by those coders who had proved to be the most profi- cient in the coding of the drug entries. Coders were supplied with the NAMCS Medication Code List (MCL), reproduced as appendix III, along with the Coding Procedures for Medication Entries, repro- duced as appendix V. In the early weeks of coding, the coauthors gave onsite instruction to the NORC coding staff in the use of these instruments. At the NORC level, the job of the coder was to apply one of three coding choices to every drug entry on the Patient Record; namely: 1. The entry was legible and on the MCL, in which case it was given the appropriate code. 2. The entry was illegible to the coder and assigned the code #99999. 3. The entry was classifiable as “other”; that is, it was legible to the coder but not found on the MCL, in which case it was assigned the code #99980. All disagreements between independent coders, as well as all illegibles and all ‘“‘others” were for- 16 warded to the NAMCS staff for adjudication.d See appendix V for a detailed description of the coding and adjudicating functions. As the coding progressed through 1980 and early 1981, there was a perceptible decline in the overall proportion of adjudicable entries; i.e., disagreements, illegibles, and ‘“‘others,” from about 8 percent for the first 5,000 entries to about 3.5 percent for the final 5,000. Interestingly, illegibles remained constant at between 1.5 and 2 percent throughout the entire coding operation; it was the diminishing need to resort to the “other” category that contributed most to the overall decline in adjudicables. In great part, this diminishing recourse to the coding choice “other” was due to an ongoing exam- ination of the ‘“‘other” entries, leading to continuing additions to the MCL in order to create a more realistic representation of coding choices. Examples: NAMCS planners had not anticipated that some physicians would write in only the thera- peutic effect (e.g., allergy relief, laxative, anal- gesic, etc.) instead of the specific drug name. Yet it became evident in the early weeks of coding that this was indeed happening to a minor but dis- turbing degree (with about 3 percent of entries). Thus a representative battery of therapeutic- effect choices was devised and added to the MCL, care being taken to make the choices compatible with the therapeutic categories in appendix IV. Every time a hitherto unanticipated generic appeared on the Patient Records, it was added to the MCL in order to insure a better representation of generic choices. Chief among the ‘other’ entries were the new brand drugs that appeared on the market subse- quent to June 1980. Though these new products were known to NAMCS by monthly updates of the Drug Product Information File, they num- dThe adjudicating function was performed for NAMCS by Dr. William H. Campbell. bered in the hundreds and no attempt was made to add all of them in periodic, massive updates of the MCL. After all, NAMCS is primarily utiliza- tion oriented and does not require the strict topicality needed in studies of market trends. Thus only those new drugs were added that began to appear with substantial frequency on the Pa- tient Records (examples: ZOMAX, TRANXENE, MECLOMEN). Whenever a new MCL choice was added, a back search and correction was made of all previously coded Patient Records. Thus, as has been seen, the need for adjudication was steadily decreased by increasing the range and representation of the MCL. The formal process of adjudication further acted to clarify doubtful areas by reconciling all disagreements between coders, and by reducing the proportion of “others” to 1.1 per- cent and illegibles to 0.2 percent. NOTE: Not least among the problems that the adjudicator helped to solve were those that involved variant spellings and nonstandard abbreviations of items that were already on the MCL, problems some- times complicated by problems of illegibility. The variant spellings probably occurred chiefly at those visits where the physician dictated the drug entries to an assistant, apparently not an uncommon arrange- ment. For this kind of problem the list of ‘“Look- Alikes or Sound-Alikes,” reproduced in appendix V, was especially helpful to coders. The write-in of non- standard abbreviations created special problems of interpretation. PENICILLIN, for example, appeared in the following variations: Porp P-cillin P cill Pcin Pen Interestingly, the use of many nonstandard abbre- viations probably owed its origin to a kind of short- hand communication—an informal system of shared understandings—that had developed over a long period between prescribing physician and dispensing pharmacist. The adjudicator—who was experienced, as a practicing pharmacist—was well fitted to solve this problem. 17 Evaluation and plan for analysis The final Medication Code List for 1980 con- tained 7,227 coding choices, of which 2,532 were actually used by the coders. These 2,532 drugs are underlined in the List that appears as appendix V. Table C shows the extent to which key drug dimen- sions were provided for among the 7,227 coding choices and the extent they were actually used in the coding operation. In the 1980 NAMCS, it was believed that the addition of a drug item on the questionnaire might negatively affect the physician response rate. How- ever, this proved not to be so because the response rate of about 77 percent was higher than in 1978 and 1979 (72.8 and 71.8 percent, respectively). None of the 530-odd refusals to participate in the 1980 NAMCS could be attributed to the presence of the drug item. On the contrary, the presence of the drug item seemed to improve the completeness and relevance of the response to other items on the Patient Record; e.g., diagnostic procedures, nonmedication therapy, and both principal and other-listed diagnoses. In analyzing the drug findings of the 1980 NAMCS, the NAMCS staff will use the degree of specificity that best describes the subject, being always alert to the effect on precision produced by the size of an estimate. To be most flexible, the ana- lyst should be able to operate at successive levels of specificity, each level should be made up of homoge- neous groupings that would allow for the cumulation of data from lower levels, and result in estimates of ever-increasing size. The NAMCS drug data provide for this kind of flexibility by permitting grouping and analysis at any of six levels. The inverted ladders, shown below, graphically illustrate these six levels of specificity. Level 6 \ Therapeutic category (2-digit level) / Level 5 \ Therapeutic category (4-digit level) / Level 4 \ Therapeutic category (6-digit level) / Level 3 \ Generic family / Level 2 Entry status; prescription status; control level; composition status Level 1 Specific drug product 18 Example: the drug PENBRITIN Anti-infectives (AHFS# 080000) Level 6 \ Level 5 \ Level 4 \ Level 3 \ Antibiotics (AHFS# 081200) Penicillins (AHFS# 081216) AMPICILLIN / Level 2 Brand name entry; prescription drug; no DEA controls; single- entity drug Level 1 PENBRITIN In figure 5 the perspective of analysis is broad- ened beyond the drug dimensions to include the other dimensions of office-based care surveyed by NAMCS. The rich mosaic of potential contrasts is obvious. These contrasts will be explored and made avail- able to the public in three forms: 1. Response to specific data requests should be possible by early 1982. 2. A series of formal publications should begin to appear in early 1982. 3. A tape of 1980 NAMCS findings should become available in mid-1982. Inquiries may be addressed to: Hugo Koch Ambulatory Care Statistics Branch National Center for Health Statistics Center Building, Room 2-43 Prince George Center 3700 East-West Highway Hyattsville, Maryland 20782 Phone: (301) 436-7132 Table C. Coding choices provided and used, MCL 1980 MCL MCL coding coding Percent Drug dimension choices choices of use provided used A B B/A Total ChOICES . . . vt i i tt tt ee ee ee ee ee ee ee ee ee ee ee ee ee 7.227 2,632 36.4 Entry status GeNEFIC BNEIY . . . oi i i tt ee ee ee ee eee ee ee ee eee ee eee 1,201 528 44.0 Brand (nongeneriCl entry « « = + « v4 25 YE 5 4 5 LB i WI BOE SE FE SA Wo wy Es 5,929 2,044 345 Therapeutic effect . o vs s x +o w15 v9 5.4 5% x Fw gs HE wm 8 3 0% 9 @ 9% s 98% 3 95 47 49.5 Prescription status Prescription (RX) drug . . . . «© ct i i i te ee ee ee ee eee 4,335 1,786 41.2 Nonprescription (OTC) drug. . . «tv tt i te ee ee ee ee ee eee 2,591 743 28.7 Federal controlled substance status SchedulB lI'dIUG + 2 sh ac BF MLB 9B EAR AS BAe Ba 8S Main om vase wo 95 43 45.3 Schedule Ml drug +.» « «sv s9 0 s 0 om 5 5H 33 Fw #3 68 0 6® 2P EH BE TEE 198 87 43.9 Schedule IV drug . . . . . i ee ee ee ee ee ee eee ee ee eee 148 ral 48.0 Schedule Vdrig +5355 555 i # 30 8m 3 BFF AB aM A ms ® s mes So vem 152 51 33.6 Totalcontrolled 5 ss vv v3 3 vs Ws Wow UF FR FF HT AF FOE EH BEETS 8 & 8 593 252 42.5 Composition status Single-entity drug . . . . LL i ee ee ee ee ee ee ee ee eee 3,651 1,487 40.7 Combination div = + a» 23 Fs BE 3 FER FW 5s 08 518 50 ao do tome wa 3,479 1,075 30.9 19 oz Other dimensions of office-based care DRUG THERAPY For principal diagnosis, all other reasons, or both Physician Information Professional identity (M.D.orD.O.) . . . .. Specialty . . ................... Solo or multiple-member practice. . . . . . . PRYSICIaN 800 +. coun vans « ss mmm me Geographic region of practice . . . ...... Patient and visit information Patientage. . . ................. PotiBntSOX . vo wmmmims » 5 5 + sr 2m mms Patientethnicity . ............... Visit status (new patient, old patient with problem or old patient with old problem . ................... Referral stats. . « « vw ct « 2s + spo mn Symptom (principal or other-listed) | EBT Diagnosis (principal or other-listed). . . . . . Nature of problem (acute or chronic) . . . . Diagnostic procedures . . . .......... Non-medication therapy . . . . ........ Disposition. . . . . .........000.... Duration of visit (patient-physician COMA) .. ov vv vusnnmsnsssaman Visit frequency Dru All at visits All drug mentions Specific drug Generic name or names (for combination drugs) Entry status: generic or brand Prescription status: Rx or OTC Federal control status Composition status: single-entity or combination Therapeutic category 1 Drug visits: Those visits where one or more drugs is ordered or provided. Figure 5. Analysis plan: Potential contrasts Appendixes Contents I. Response to precoded questions on evaluation INTEIVIEW . . . uu ii iii it tt it tt te tt ttt ie tenet eens 22 IH. INVeNTOrY OF GBNBIICIIBITIBE «vw mrs =» mo wre: 0 ln nem wim 0k fm mmo cam os dm 0 0 Tm wm Br a? 1 25 111. Medication Code List, NAMGCS 1080 . . . . .. tii iii itt i ttt te tt ttt ttt tet ttt t ett e ete e eens 32 IV. Coding procedures for medication entries, NAMCS 1980 . . . .. .. iii iii titi titi ttt tet e teens 73 Attachment 1. “Common Abbreviations Used in Medical Orders’ . . ...... citi titi ttt enenennnens 80 Attachment 2. “Caution! 1,000 Drugs Whose Names Look-Alike or Sound-Alike"” .......... cir nnn. 84 Attachment 3. Drug Form | (Coder A) «oo cii titi iit tt ttt t ett tet e tsetse seein enn 87 Attachment 4. Drug Form | (Coder B) . .. oc iii iii t tt tit tit ett iets t sitet ieee 88 Attachment 5. Drug Form I {AJJUGIGAION) 2. cs cums a nmmus va ammass sus omens owsnes stnsenassoss 89 V. American Hospital Formulary Service classification system and therapeutic category codes .................... 90 21 Appendix |. Response to precoded questions on evaluation interview Q. 1: Doctor, as you know, we are especially interested on this pretest in your reactions to the medication items—that is, (Item 15-16 on A form, Item 12 on B form). Was this information very easy for you to fill out, was it occasionally difficult, or did you have real prob- lems in recording the medication therapy you ordered or provided? NUMBER OF CASES IN: “A” Sample “B” Sample TOTAL (Precoded) (Open-ended) Very easy 10 15 25 Occasionally difficult 7 2 9 Real problems — 1 1 TOTALS 17 18 35 Q. 2: Do you think you included all prescription drugs that you ordered or provided for these patients, or might you have missed some? NUMBER OF CASES IN: “A” Sample “B” Sample TOTAL (Precoded) (Open-ended) Included all 14 13 3 Might have missed some 3 5 8 TOTALS 17 18 3 22 Q. 3: How about nonprescription drugs—Are you pretty sure you recorded all of these, or might you not have reported all of them? Reported all Might not have Not applicable (Volunteered if no such drugs ordered or provided) TOTALS NUMBER OF CASES IN: “A” Sample “B” Sample TOTAL (Precoded) (Open-ended) 9 10 19 6 8 14 2 - 2 17 18 35 Q. 4: The form asked you to report both “new and continuing” medication. Do you think you always included old medication that you wanted the patient to continue, even though it might not have been related to this visit? Yes No Don’t know TOTALS NUMBER OF CASES IN: “A” Sample “B” Sample TOTAL (Precoded) (Open-ended) 13 15 28 2 2 4 2 1 17 18 35 Q. 5: How confident are you that you included all immunizing or desensitizing agents among the medications that you ordered or provided—Are you fairly sure you always reported, or do you think you might some- times overlook these? Always reported Sometimes overlooked Not applicable (Volunteered if no immunizations ordered or provided) TOTALS NUMBER OF CASES IN: “A” Sample “B” Sample TOTAL (Precoded) (Open-ended) 15 13 28 - 2 2 2 3 5 17 18 35 23 Q. 7: Would you rather use the write-in version of the medication items, or would you rather use the check-off 24 version? Prefer write-in Prefer checklist No preference TOTALS NUMBER OF CASES IN: “A” Sample “B” Sample TOTAL (Precoded) (Open-ended) 4 11 15 9 6 15 4 1 5 17 18 35 az 50000 50005 50010 50015 50020 50025 50030 50035 50040 50045 50050 50055 50060 50065 50070 50075 50080 50085 50090 50100 50105 50108 50110 50120 50125 50130 50135 50140 50145 50150 50155 50160 50165 50170 50175 50180 50185 50190 50195 50200 50205 50210 50220 50225 50235 50240 50245 50250 50255 50260 50265 50270 50270 50275 50280 50285 50290 50295 50300 50305 UNDETERMINED ACETAMINOPHEN ACETAMINOPHEN & CODEINE ACETANIL ID ACETAZOLAMIDE ACETIC ACID ACETOHEXAMIDE ACETONE ACETOPHENAZINE ACETRIZOATE SODIUM ACETYLCARBROMAL ACETYLCHOLINE ACETYLCYSTEINE ACONITE ACRI FLAVINE ACRISORCIN ADENOSINE ALBUMIN HUMAN ALCOHOL ALKAVERVIR ALLOPURINOL ALOIN BELLADONNA CASC & PODOPHY ALPHAPRODINE ALSEROXYLON ALUMINUM ALUMINUM ACETATE ALUMINUM AMMONIUM SULFATE ALUMINUM CHLORIDE ALUMINUM HYDROXIDE ALUMINUM NICOTINATE ALUMINUM PHOSPHATE ALUMI NUM SUBACETATE AMANTADINE AMBE NONIUM AMC I NONIDE AMIKACIN AMINO ACIDS AMINOACETIC ACID AMINOBENZOIC ACID AMINOCAPROIC ACID AMINOHIPPURATE SODIUM AMINOPHYLLINE AMINOSALICYLIC ACID AMITRIPTYLINE AMMONIA SPIRIT AROMATIC AMMONIATED MERCURY AMMONIUM CHLORIDE AMOBARBITAL AMODIAQUINE AMOXICILLIN AMPHETAMINE AMPHOTERICIN AMPHOTERICIN B AMPICILLIN AMYL NITRITE ANILERIDINE ANISE OIL ANISINDIONE ANI SOTROPINE ANTAZOLINE 50310 50315 50320 50325 50330 50335 50340 50345 50350 50360 50365 50370 50375 50380 50400 50405 50410 50415 50420 50430 50435 50440 50445 50450 50455 50460 50465 50475 50480 50485 50490 50495 50500 50505 50515 50520 50525 50530 50535 50540 50545 50550 50555 50558 50560 50565 50570 50575 50580 50585 50590 50595 50600 50605 50610 50615 50620 50625 50630 50635 INVENTORY OF GENERIC NAMES ANTHRALIN ANTICOAGULANT CITRATE DEXTROSE ANTIHEMOPHILIC FACTOR HUMAN ANTIPYRINE ANTIRABIES SERUM ANTISEPTIC SOLUTION APC APC W/CODEINE APC W/PROPOXYPHENE APOMORPHINE APROBARBITAL ARGININE ARNICA TINCTURE ASAFETIDA TINCTURE ASPARAGINASE ASPERGILLUS ASPIRIN ASPIRIN & CODEINE ATROPINE ATTAPULGITE AUROTHIOGLUCOSE AZATADINE AZATHIOPRINE BACITRACIN BACLOFEN BAL SAM BANDAGE BARBITAL BARIUM SULFATE BCG VACCINE BEC LOME THA SONE BELLADONNA BENDROFLUMETHIAZIDE BENOXINATE BENZALKONIUM CHLORIDE BENZETHONIUM CHLORIDE BENZ IN BENZ OCA INE BENZOIC & SALICYLIC ACID BENZOIN BENZONATATE BENZOYL PEROXIDE BENZPHE TAMINE BENZQUINAMIDE BENZTHIAZ IDE BENZ TROP INE BENZYL ALCOHOL BENZYL BENZOATE BETA-CAROTENE BETAMETHA SONE BETAZOLE BETHANECHOL BICHLCRACETIC ACID BILAZO REAGENT BILE ACIDS BIGFLAVONOIDS BIOTIN BIPERIDEN BISACODYL BISMUTH ANTI-DIARRHEA AGENTS 50640 50650 50655 50660 50665 50670 50675 50680 50685 50690 50695 50700 50705 50710 50715 50720 50725 50730 50735 50740 50745 50750 50755 50760 50770 50775 50780 50785 50800 50835 50840 50845 50855 50860 50865 50870 50875 50880 50885 50890 50895 50900 50905 50908 50910 50915 50920 50925 50930 50935 50940 50945 50950 50950 50955 50960 50965 50975 50980 50985 BISMUTH SALICYLATE BISMUTH TRIBRCMOPHENATE BLEOMYCIN BORIC ACID BRETYL IUM BRILLIANT GREEN BROMELAINS BROMOCRIPTINE BROMODIPHENHYDRAMINE BROMPHENIRAMINE BROWN MIXTURE BUCHU JUNIPER & POT ACETATE BUCL IZ INE BUPIVACAINE BUSULFAN BUTABARBITAL BUTACA INE BUTAMBEN BUTAPERAZINE BUTORPHANOL CAFFEINE CAFFEINE & SODIUM BENZOATE CALAMINE CALCITONIN CALCIUM REPLACEMENT AGENTS CALCIUM ACETATE CALCIUM BROMIDE CALCIUM CARBONATE CALCIUM HYDROXIDE CALLICREIN CALUST ERONE CAMPHOR CANDICIDIN CANTHARIDIN CAPREOMYCIN CARAMEL CARBACHOL CARBAMAZEPINE CARBARSONE CARBAZOCHROME CARBENICILLIN CARBINOXAMINE CARBOL-FUCHS IN CARBON DIOXIDE CARBON TETRACHLORIDE CARDAMOM CARISOPRODOL CARMUSTINE CARPHENAZ INE CASCARA CASTOR OIL CEFACLOR CEFADROXIL CEFADROXL CEFAMANDOLE CEFAZCLIN CEFOXITIN CELLULOSE CEPHALEXIN CEPHALOGLYCIN sowieu 211auab Jo Auojuanu| ‘|| xipuaddy 9c 50990 50995 51000 51 005 51010 51015 51020 51025 51030 51035 51040 51 045 51050 51 055 51060 51070 51075 51080 51 085 51090 51095 51100 51105 51110 51115 51120 51125 51130 51135 51145 51150 51155 51160 51165 51170 51175 51180 51185 51190 51195 51200 51205 51210 51215 51220 51225 51230 51235 51240 51245 51250 51255 51260 51265 51270 51275 51280 51285 51290 51295 CEPHALORIDINE CEPHALOTHIN CEPHAPIRIN CEPHRADINE CERIUM OXALATE CETALKONIUM CHLORIDE CETYL ALCOHOL CHARCOAL CHERRY SYRUP CHLOPHEDIANOL CHLORAL HYDRATE CHLORAMBUCIL CHLORAMPHENICOL CHLORDIAZEPOX IDE CHLORHEXIDINE CHLOROALLYLHEXAMINIUM CHLORIDE CHLOROBUTANOL CHLOROFORM CHLOROPHYLL CHLOROPROCAINE CHLOROQUINE CHLOROTHIAZIDE CHLOROTHYMOL CHLOROTRIANISENE CHLOROXINE CHLOROXYLENOL CHLORPHENESIN CHLORPHENIRAMINE CHLORPHE NOXAMINE CHLORPHENTERMINE CHLORPROMAZINE CHLORPROPAMI DE CHLORPROTHIXENE CHLORTETRACYCL INE CHLORTHALIDONE CHLORZOXAZONE CHOLERA VACCINE CHOLESTEROL CHOLESTYRAMINE CHOLINE CHOLINE SALICYLATE CHORIONIC GONADOTROPIN CHROMIC OXIDE CHYMOTRYPSIN CIMETIDINE CINNAMON OIL CISPLATIN CITRIC ACID CITRONELLA OIL CLEMASTINE CLIDINIUM CLINDAMYCIN CLOFIBRATE CLOMIPHENE CLONAZEPAM CLONIDINE CLORAZEPATE CLORTERMINE CLOTRIMAZOLE CLOVE OIL 51300 51305 51310 51315 51320 51325 51330 51335 51340 51345 51350 51355 51360 51365 51375 51380 51380 51385 51390 51395 51400 51405 51410 51415 51425 51430 51435 51440 51445 51450 51460 51465 51470 51475 51480 51485 51490 51495 51500 51505 51510 51515 51520 51530 51535 51540 51545 51550 51555 51560 51565 51570 51575 51585 51590 51595 51600 51605 51610 51615 INVENTORY OF GENERIC NAMES CLOXACILLIN COAL TAR COCAINE COCCIDIOIDIN COCILLANA COCOA BUTTER COCONUT OIL COD & HALIBUT LIVER OIL CODE INE COLCHIC INE COLD CREAM COLESTIPOL COLISTIN COLLAGEN DERIVATIVE COLLODION COMBINATION PRODUCT TR IMETHOPRIM CONGO RED CONTACT LENS SOLUTION CORN OIL CORTICOTROPIN CORTISONE COSYNTROPIN COTTONSEED OIL CRESOL CROMOLYN SODIUM CROTAMITON CRYPT TENAMINE CUPRIC SULFATE CUPRIC SULFATE REAGENT CYCLACILLIN CYCLANDELATE CYCLOBENZAPRINE CYCLOMETHYCAINE CYCLOPENTOLATE CYCLOPHOSPHAMIDE CYCLOSER INE CYCLOTHIAZIDE CYCLIZINE CYCR IMINE CYPROHEPTADINE CYSTEINE CYTARABINE DACARBAZ INE DACTINOMYCIN DANAZOL DANTHRON DANTROLENE DAP SONE DAUNOCRUBICIN DEANOL DECAME THONI UM DEFEROXAMINE DEHYDROCHOLIC ACID DEMECARIUM DEMECLOCYCLINE DE SERPIDINE DESIPRAMINE DESLANOSIDE DE SMOPRE SSIN 51620 51625 51630 51635 51640 51645 51660 51665 51670 51675 51685 51695 51700 51705 51710 51715 51720 51725 51730 51735 51740 51745 51750 51755 51760 51765 51770 51775 51780 51785 51790 51795 51800 51805 51810 51815 51820 51825 51830 51835 51840 51845 51850 51855 51860 51865 51870 51875 51885 51890 51890 51900 51910 51915 51920 51925 51930 51935 51940 51945 DESONIDE DESOXIMETASONE DESOXYCORTICOSTERONE DEXAMETHASONE DEXCHLORPHENIRAMINE DEXPANTHENOL DEXTRANOMER DEXTROAMPHETAMINE DEXTROMETHORPHAN DEXTROSE DEXTROTHYROXINE DIAZEPAM DIAZOX IDE DIBUCAINE DICHLOROTETRAFLUOROETHANE DICHLORPHENAMIDE DICLOXACILLIN DICUMAROCL DICYCLOMINE DIENESTROL DIETARY SUPPLEMENT DIETHYLCARBAMAZINE DIETHYLPROPION DIETHYLSTILBESTROL DIFLORASONE DIGALLOYL TRIOLEATE DIGITALIS DIGITOXIN DIGOXIN DIHYDROERGOTAMINE DIHYDROTACHYSTEROL DIHYDROXYALUMINUM AM INOACETATE DIHYDROXYALUMINUM SOD CARB DIISOBUTYLPHENOXYPOL YE THOXYETHAN DIMENHYDRINATE DIMERCAPROL DIMETHINDENE DIMETHISOQUIN DIMETHYL SULFOXIDE DIMETHYLAMINOBENZALDEHYDE DINOPROSTONE DIOXYL INE DIPH TET TOXOIDS PERTUSSIS DIPHEMANIL DIPHENHYDRAMINE DIPHENIDOL DIPHENOXYLATE & ATROPINE DIPHENOXYLATE & ATROPINE DIPHENYLPYRAL INE DIPHTERIA TETANUS TOXIODS DIPHTHERIA TETANUS TOXOIDS DIPHTHERIA ANTITOXIN DIPHTHERIA TOXOID DIPYRIDAMOLE DISOPYRAMIDE DISULFIRAM DOBUTAMINE DOCUSATE DODECAETHYLENEGL YCOL DOPAMINE Lz 51950 51955 51960 51965 51970 51980 51 985 51990 51 995 52000 52 005 52010 52015 52020 52025 52030 52035 52040 52045 52050 52055 52060 52 065 52070 52075 52080 52085 52090 52 095 52100 52105 52110 52115 52120 52125 52130 52130 52135 52140 52145 52150 52155 52160 52165 52170 52175 52180 52185 52190 52195 52200 52210 52215 52220 52225 52270 52273 52275 52280 52290 DOXA PRAM DOXEPIN DOXORUBICIN DOXYCYCLINE DOXYLAMINE DROMOS TANOLONE DROPERIDOL DYCL ONINE DYDR OGESTERONE DYPHYLLINE ECHOTHIOPHATE EDETATE CALCIUM DISODIUM EDETATE DISODIUM EDROPHONIUM EMETINE ENF LURANE EPHEDRINE EPINEPHRINE ERGOCALCIFEROL ERGONOVI NE ERGOTAMINE ERYTHRITYL TETRANITRATE ERYTHROMYCIN ESTRADIOL ESTROGENS ESTRONE ETHACRYNIC ACID ETHAMBUTOL ETHAVERINE ETHCHLORVYNOL ETHER ETHI NAMATE ETHI ONAMIDE ETHOHE PT AZINE ETHOPROPAZINE ETHOSUXAMIDE ETHOSUXIMIDE ETHOTOIN ETHOXAZENE ETHOXZOL AMI DE ETHYL ACETATE ETHYL CHLORIDE ETHYLESTRENOL ETHYLMORPHINE ETHYLNOREPINEPHRINE ETIDOCAINE ETIDOCAINE & EPINEPHRINE ETIDRONATE DISODIUM EUCALYPTOL EUGENOL EVANS BLUE FAST GREEN FCF FENF LURAMINE FENOPROFEN FENTANYL FIBRINOLYSIN FICRINAL W/CODEINE FLAVOXATE FLOXURIDINE FLUCYTOSINE 52295 52300 52305 52310 52315 52320 52325 52325 52330 52335 52340 52345 52350 52355 52360 52365 52370 52375 52380 52385 52390 52395 52400 52405 52410 52415 52420 52425 52430 52435 52440 52445 52450 52455 52460 52465 52470 52475 52480 52485 52490 52495 52500 52505 52510 52520 52525 52530 52535 52540 52545 52550 52555 52560 52565 52575 52580 52585 52590 52595 INVENTORY. OF GENERIC NAMES FLUDROCORTI SONE FLUME THA SONE FLUOC INOLONE FLUOC INONIDE FLUORESCEIN FLUOROME THOLONE FLUOROURACIL FLUOUROURACIL FL UD XYME STERONE FLUPHENAZ INE FLUPREDNISOL ONE FLURANDRENOLIDE FLURAZEPAM FOLIC ACID FORMALDEHYDE FORMIC ACID FRUCTOSE FULLER'S EARTH FURAZOLIDONE FUROSEM IDE GALLAMINE GELATIN GENTAMICIN GENTIAN VIOLET GINSENG GITALIN GLUCAGON GLUCOSE GLUCOSE ENZYMATIC TEST GLUTAMIC ACID GLUTARALDEHYDE GLUTETHIMIDE GLYCERIN GLYCOPYRROLATE GLYCOPYRROLATE W/PHENCBARBITAL GLYCYRRHIZA GOLD SODIUM THIGOMALATE GOLD SODIUM THIOSULFATE GRISEOFULVIN GUAIACOL GUAIFENESIN GUANETHIDINE GUANID INE HALAZONE HALC INONIDE HALOPER IDOL HALOPROGIN HALOTHANE HAMAMELIS WATER HEPARIN HEPATITIS B IMMUNE GLOBULIN HESPERIDIN HETASTARCH & SODIUM CHLORIDE HE XACHLOROPHENE HE XAFLUORENIUM HE XE STROL HEXOBARBITAL HEXOCYCLIUM HEXYLCA INE HEXYLRE SORCINOL 52600 52605 52610 52615 52618 52620 52625 52630 52635 52640 52645 52650 52655 52660 52665 52670 52675 52680 52685 52690 52695 52700 52705 52710 52715 52720 52730 52735 52740 52745 52750 52755 52765 52770 52775 52780 52785 52790 52795 52800 52805 52810 52815 52820 52820 52825 52830 52835 52845 52850 52855 52860 52865 52870 52875 52880 52885 52888 52890 52895 HISTAMINE HISTIDINE HISTOPLASMIN HOMATROPINE HONEY HOUSE DUST ALLERGENIC EXTRACT HYALURONIDASE HYDRAL AZINE HYDRIODIC ACID HYDROCHLORIC ACID HYDROCHLOROTHIAZIDE HYDROCODONE HYDROCORT ISONE HYDROFLUMETHIAZIDE HYDROGEN PEROXIDE HYDROMORPHONE HYDROQUINONE HYDROXOCOBAL AMIN HYDROXYAMPHETAMINE HYDROXYCHL OROQU INE HYDROXYPROGESTERONE HYDROXYPROPYL METHYLCELLULOSE HYDROXYSTILBAMIDINE HYDROXYUREA HYDROXYZINE HYOSCYAMINE IBUPROFEN I EHTHAMMOL IDOXURIDINE IMIPRAMINE IMMUNE GLOBULIN INDIGOINDISULFONATE INDOCYANINE GREEN INDOMETHACIN INFANT FORMULA INFLUENZA VIRUS VACCINE INOSITOL INSULIN INUL IN INVERT SUGAR IOCARMATE MEGLUMINE IOCETAMIC ACID IODAMIDE MEGLUMINE IODINE TOPICAL PREPARATIONS TOPICAL IODINE PREPARATIONS IODINATED GLYCEROL IODINE SOLUTIONS IODIPAMIDE MEGLUMINE IODOCHLORHYDROXYQUIN 1000QU INCL IOPANOIC ACID IOPHENDYLATE IOTHALAMATE IPECAC IPECAC AND OPIUM IPODATE CALCIUM IRON PREPARATIONS IRON QUININE & STRYCHNINE IRON BILE SALTS ISOCARBOXAZID 8C 52900 52 905 52910 52915 52920 52925 52930 52935 52945 52950 52 955 52965 52970 52975 52980 52985 52990 52995 53000 53 005 53010 53015 53020 53025 53030 53035 53040 53 045 53050 53055 53060 53065 53070 53 085 53090 53095 53100 53105 53110 53115 53120 53125 53135 53140 53145 53150 $3155 53160 53165 53170 53175 53180 53185 53190 53195 53220 53235 53240 53245 53250 I SOE THARINE ISOFLUROPHATE I SOL EUCINE ISONIAZID I SOPROPAMIDE ISOPROPYL ALCOHOL ISOPROTERENOL ISOSORBIDE I SOXSUPRINE KANAMYCIN KAOLIN KARAYA GUM KETAMINE LACTASE LACTIC ACID LACTOBACILLUS ACIDOPHILUS LACTOSE LACTULOSE LANATOSIDE C LAURYL SULFOACETATE LAVENDER OIL LEAD ACETATE LECITHIN LEMON OIL LEUCINE LEUCOVORIN LEVALLORPHAN LEVODE SOXY EPHEDRINE LEVODOPA LEVORPHANOL LEVOPROPOXYPHENE LEVOTHYROXINE LIDOCAINE LIME SOLUTION SULFURATED LINCOMYCIN LIND ANE LINOLENIC ACID LINSEED OIL LIOTHYRONINE LIOTRIX LITHIUM LIVER DERIVATIVE LOMUSTINE LOPERAMIDE LORAZEPAM LOXAPINE LYPRESSIN LYSINE MAFENI DE ‘MAGALDRATE MAGNESIUM ANTACIDS MAGNESIUM GLUCONATE MAGNESIUM & SODIUM CITRATES MAGNESIUM CHLORIDE MAGNESIUM CATHARTICS MAGNESIUM SALICYLATE MALT SOUP EXTRACT MANG ANESE GLUCONATE MANNITOL MAZINDOL 53255 53260 53265 53270 53275 53280 53285 53290 53295 53298 53300 53315 53320 53325 53330 53335 53340 53345 53350 53355 53360 53365 53370 53375 53380 53380 53385 53390 53395 53400 53405 53410 53415 53420 53425 53430 53435 53440 53445 53450 53455 53460 53465 53470 53475 53485 53490 53495 53500 53505 53510 53515 53520 53525 53530 53535 53540 53545 53550 53555 INVENTORY OF GENERIC NAMES MEASLES VIRUS VACCINE MEBENDAZOLE MECAMYLA MINE MECHLORETHAMINE MECLIZINE MEDROXYPRDGESTERONE MEDRYSONE MEFENAMIC ACID MEGESTROL MEGLUMINE MELPHALAN MENINGOCOCCAL VACCINE MENOTROP INS MENTHOL MEPENZOLATE MEPERIDINE MEPHENE SIN MEPHENTERMINE MEPHENYTOIN MEPHOBARBITAL MEPIVACAINE MEPREDNI SONE MEPROBAMATE MERBROMIN MERCAPTOMERIN MERCAPTOPURINE MERCAPTOPURINE MERCOCRESOLS MERCURIC CHLORIDE MERCURIC CYANIDE MERCUR IC ICDIDE MERCURIC OXIDE MERCURIC SULFIDE MERCUROPHYLLINE MERCUROUS CHLORIDE MERSALYL & THEOPHYLLINE MESORIDAZINE METOCLOPRAMIDE METOPROLOL METAPRO TERENOL METARAMINOL METAXALONE METHACHOLINE METHACYCLINE METHADONE METHAMPHETAMINE METHANDR IOL ME THANDROSTENOLONE METHANTHELINE METHAPYRILENE METHAQUALONE METHARBITAL METHAZOLAMIDE METHDILAZINE METHENAMINE METHICILLIN METHIMAZOLE METHIODAL SODIUM METHION INE METHIXENE 53560 53565 53570 53575 53580 53585 53590 53595 53600 53605 53610 53615 53620 53625 53630 53635 53640 53645 53650 53655 53660 53670 53675 53680 53685 53690 53695 53700 53705 53710 53715 53720 53725 53730 53735 53740 53745 53750 53755 53760 53765 53770 53775 53780 53800 53810 53815 53820 53825 53835 53840 53845 53855 53860 53865 53870 53875 53880 53885 53890 METHOCARBAMOL METHOHEX ITAL METHOTREXATE METHOTRIMEPRAZ INE METHOXAMINE METHOXSALEN METHOXY FLURANE METHOXYPHENAMINE METHSCOPOL AMINE METHSUXIMIDE METHYCLOTHIAZIDE METHYL ALCOHOL METHYL SALICYLATE METHYLBENZETHONIUM CHLORIDE METHYLCELLULOSE METHYLDOPA METHYLENE BLUE METHYLERGONOVINE METHYL PARABEN METHYLPHENIDATE METHYL PREDNISCLONE METHYLTESTOSTERONE METHYPRYLON METHYSERGIDE METOCUR INE METOLA ZONE METRIZAMIDE METRONIDAZOLE METYRAPONE METYROS INE MICONAZOLE MINERAL OIL MINOCYCLINE MINOXIDIL MITHRAMYCIN MITOMYCIN MITOTANE MOLINDONE MONOBENZONE MORPHINE MORPHINE & ATROPINE MORRHUATE SODIUM MOUTHWASH MULTIVITAMINS GENERAL MULTIVITAMINS PRENATAL MUMPS IMMUNE GLOBULIN MUMPS SKIN-TEST ANTIGEN MUMPS VIRUS VACCINE MUSTARD OIL MYRRH NADOL OL NAFCILLIN NALBUPHINE NALIDIXIC ACID NALOXONE NANDR OL ONE NAPHAZOL INE NAPROXEN NATAMYC IN NEGATOL 62 53895 53900 53 905 53915 53920 53925 53930 53935 53940 53945 53950 53955 53960 53965 53970 53975 53980 53985 53990 53995 54000 54005 54010 54015 54020 54025 54030 54035 54 045 54050 54055 54060 54065 54070 54075 54080 54085 54090 54095 54100 54105 54110 54115 54120 54130 54135 54140 54145 54150 54155 54160 54165 54170 54175 54180 54190 54195 54200 54205 54210 NEOMYC IN NEOSTIGMINE NIACIN NI ACINAMIDE NICOTINYL ALCOHOL NIKE THAMIDE NITROFURANTOIN NITROFURAZONE NITR CGEN NITROGLYCERIN NITROUS OXIDE NONOXYNOL NOREPINEPHRINE NORE THINDRONE NORGESTREL NORTRIPTYLINE NOSCAPINE NOVOBIOCIN NYLIDRIN NYSTATIN OATMEAL OCTYL DIMETHYL PABA OCTYL METHOXYCINNAMATE OINTMENT HYDROPHILIC OLEANDOMYCIN OLEIC ACID OLIVE OIL OPIUM ORANGE OIL ORPHENADRINE ORTHOTOLIDINE REAGENT OUABAIN OX BILE EXTRACT OXACILLIN OXALIC ACID OXANDROLONE OXAZEP AM OXOLINIC ACID OXYPHENCY CLIMINE OXTRIPHYLLINE OXYBUTYNIN OXYCHLOROS ENE OXYGEN OXYMETAZOLINE OXYMETHOLONE OXYMOR PHONE OXYPHENBUTAZONE OXYP HE NOMI UM OXYTETRACYCL INE OXYTOCIN PANCREATIN PANCRELIPASE PANCURONIUM PANTOTHENIC ACID PAPAIN PAPAVERINE PARALDEHYDE PARAMETHADIONE PARAMETHASONE PARATHYROID 54215 54220 54225 54230 54235 54240 54245 54250 54270 542175 54280 54290 54295 54300 54305 54310 54320 54325 54330 54345 54350 54355 54360 54365 54370 54375 54380 54385 54390 54395 54400 54405 54410 54415 54420 54425 54430 54435 54440 54445 54450 54455 54460 54465 54470 54475 54480 54485 54490 54495 54500 54505 54510 54515 54520 54525 54530 54535 54540 54545 INVENTORY OF GENERIC NAMES PAREGORIC PARGYLINE PAROMOMYCIN PATCHOULI OIL PEANUT OIL PEMOLINE PENICILLAMINE PENICILLIN PENICILLOYL PENTAERYTHRITOL PENTAGASTRIN PENTAZOCINE PENTOBARBITAL PENTYLENETETRAZOL PEPPERMINT PEPSIN PERPHENAZ INE PERTUSSIS IMMUNE GLOBULIN PETROLATUM PHENACEMIDE PHENACETIN PHENAPHEN W/CODE INE PHENAPHTHAZ INE PHENAZOPYRIDINE PHENDIMETRAZINE PHENELZ INE PHENIND IONE PHENIRAMINE PHENMETRAZINE PHENOBARBITAL PHENOL PHENOLPHTHALEIN PHENOL SULFONPHTHALEIN PHENOXYBENZAMI NE PHENPROCOUMON PHENSUXIMIDE PHENTERMINE PHENTOLAMINE PHENYL SALICYLATE PHENYLALANINE PHENYLBUTAZONE PHENYLE PHRINE PHENYLMERCURIC NITRATE PHENYLPROPANOLAMINE PHENYTOIN PHOSPHOMOLYBDATE REAGENT PHOSPHORIC ACID PHTHALYLSULFATHIAZOLE PHYSOSTIGMINE PHYTONADIONE PILOCARPINE PIPERACETAZINE PIPERAZINE PIPERAZINE ESTRONE PIPEROCAINE PIPOBROMAN PITUITARY POSTERIOR PLACEBO PLAGUE VACCINE PLANTAGO SEED 54550 54555 54560 54565 54570 54575 54585 54590 54595 54605 54610 54615 54620 54625 54630 54640 54645 54650 54655 54700 54705 54710 54715 54720 54725 54730 54735 54740 54745 54750 54755 54760 54765 54770 54775 54785 54790 54795 54800 54805 54810 54815 54825 54830 54835 54840 54845 54850 54860 54865 54870 54875 54880 54885 54890 54895 54900 54905 54910 54915 PLASMA PROTEIN FRACTION PNEUMOCOCCAL VACCINE PCDOPHYLLUM POISON IVY EXTRACT POLAXAMER POLIG VACCINE POLLEN ANTIGEN POLYETHYLENE GLYCOL POLYMIXIN B POLYTHIAZIDE POLYVINYL ALCOHOL POTASH SULFURATED POTASSIUM ALKALINIZING AGENTS POTASSIUM AMINOBENZOATE POTASSIUM ARSENITE POTASSIUM BITARTRATE POTASSIUM BROMIDE POTASSIUM CARBONATE POTASSIUM REPLACEMENT SOLUTIONS POTASSIUM GUATIACOLSULFONATE POTASSIUM HYDROXIDE POTASSIUM IODIDE POTASSIUM NITRATE POTASS IUM OXYQUINOLINE SULFATE POTASSIUM PERCHLORATE POTASSIUM PERMANGANATE POTASSIUM ACIDIFYING AGENTS POTASSIUM SODIUM TARTRATE POTASSIUM THIOCYANATE PRALIDOX IME PRAMOX INE PRAZEPAM PRAZOSIN PREDNISOLONE PREDNISONE PRILOCAINE PRIMAQUINE PRIMIDONE PROBENECID PROBUCOL PROCAINAMIDE PROCAINE PROCARBAZINE PROCHLORPERAZINE PROCYCLIDINE PROFLAVINE PROGESTERONE PROMAZINE PROMETHAZ INE PROPRANOLOL PROPANTHEL INE PROPARACAINE PROPIOLACT ONE PROPIOMAZINE PROPOXYPHENE PROPYLENE GLYCOL PROPYLHEXEDRINE PROPYLPARABEN PROPYLTHIOURACIL PROTAMINE SULFATE 0g 54920 54930 54935 54940 54945 54950 54960 54965 54975 54980 54985 54990 54995 55000 55005 55015 55020 55025 55030 55035 55040 55045 55050 55055 55060 55063 55065 55070 55075 55080 55085 55095 55105 55110 55115 55120 55125 55130 55135 55140 55145 55150 55160 55165 55170 55180 55185 55190 55195 55200 55210 55215 55220 55225 55230 55235 55240 55245 55250 55255 PROTEIN HYDROLYSATE PROTEOLYTIC ENZYME PROT IRELIN PROTOKYLOL PROTOVERATRINE PROTRI PTYLINE PSEUDOEPHEDRINE PSYLLIUM PYRANTEL PYRAZINAMIDE PYRIDOSTIGMINE PYRI DOXINE PYRILAMINE PYRIMETHAMINE PYRI THIONE PYROGALLOL PYROXYLIN PYRVINIUM QUINACRINE QUINESTROL QUINETHAZONE QUINIDINE QUININE RABIES IMMUNE GLOBULIN RABIES VACCINE RACEPHEDRINE RASP BERRY SYRUP RAUWOLFIA RESC INNAMINE RESERPINE RESORCINOL RHUBARB & SODA RIBOFLAVIN RIFAMPIN RINGER'S LACTATED ROSA GALLICA EXTRACT ROSE WATER RUBELLA VIRUS VACCINE RUTIN SACCHARIN SAFF LOWER OIL SALICYLAMIDE SALICYLIC ACID SALSALATE SCARLET RED SCOPOL AMINE SECOBARBITAL SECRETIN SELENIUM SULFIDE SENNA SESAME OIL SILICIC ACID SILVER IODIDE SILVER NITRATE SILVER PROTEIN SILVER. SULFADIAZINE SIMETHICONE SINCALIDE SITOSTEROLS SMALLPOX VACCINE 55260 55265 55270 55275 55280 55285 55290 55295 55300 55305 55310 55315 55320 55325 55330 55335 55340 55345 55355 55360 55365 55370 55380 55385 55390 55400 55405 55410 55415 55420 55425 55430 55435 55440 55445 55450 55455 55455 55460 55465 55470 55475 55480 55485 55490 55495 55500 55505 55510 55515 55520 55525 55530 55535 55540 55545 55550 55555 55560 55565 INVENTORY OF GENERIC NAMES ANTIVENIN SNAKE BITE SOAP SODIUM ACETATE SODIUM BENZOATE SODIUM BICARBONATE SODIUM BISULFATE SODIUM BISULFITE SODIUM BORATE SODIUM BROMIDE SODIUM CACODYLATE SODIUM CARBONATE SODIUM CHLORIDE SODIUM CITRATE SODIUM DICHROMATE SODIUM FLUORIDE SODIUM GLUTAMATE SODIUM HYDROXIDE SODIUM HYPOCHLORITE SODIUM LACTATE SODIUM LAURYL SULFATE SODIUM NITRITE SODIUM NITROPRUSSIDE SODIUM PERBORATE SODIUM PHOSPHATE SODIUM POLYSTYRENE SULFONATE SODIUM SUCCINATE SODIUM SULFATE SODIUM SULFITE SODIUM TETRADECYL SULFATE SODIUM THIOSALICYLATE SODIUM THIOSULFATE SOMATROPIN SORBITOL SOYBEAN OIL SPEARMINT OIL SPECTINOMYCIN ANTIVENIN SPIDER BITE ANTIVENIN SPIDER-BITE SPIRONOLACTONE STANNOUS FLUORIDE STANOZOLOL STAPHYLOCOCCUS TOXOID STARCH STEARIC ACID STEARYL ALCOHOL STREP TOKINASE STREPTOMYCIN STRYCHNINE SUCCINYLCHOLINE SUCCINYLSULFATHIAZOLE SUCROSE SUL FACETAMIDE SULFACHLORPYRIDAZINE SULFACYTINE SW FADIAZ INE SULFAMETER SULFAMETHIZOLE SUL FAMETHOXAZ OLE SULFAMETHOXYPYRIDAZINE SULFANILAMIDE 55570 55575 55580 55585 55585 55590 55595 55600 55605 55610 55613 55615 55615 55620 55630 55635 55640 55645 55650 55655 55665 55670 55675 55680 55685 55690 55695 55700 55705 55710 55715 55720 55725 55730 55739 55740 55745 55750 55755 55760 55765 55770 55775 55780 55785 55790 55795 55800 55805 55810 55815 55820 55825 55830 55835 55840 55845 55850 55870 55875 SULFAPYRIDINE SULFASALAZINE SULFATHIAZOLE SULFINPYRAZOLE SULFINPYRAZONE SULFISOXAZOLE SULFOBROMOPHTHALEIN SULFOSALICYLIC ACID SULFOXONE SODIUM SULFUR SULFUR (CATHARTIC) SULINDAC SULINOAC SULISOBENZONE SUTILAINS SYRUP TALBUTAL TALC TAMOXIFEN TANNIC ACID TARTARIC ACID TERBUTALINE TERP IN HYDRATE TERPIN HYDRATE & CODEINE TERPIN HYDRATE & DM TESTOLACTONE TESTOSTERONE TETANUS ANT ITOXIN TETANUS IMMUNE GLOBULIN TETANUS TOXOID TETRACAINE TETRACHLORSALICYLANILIDE TETRACYCL INE TETRAHYDROZOL INE THEOBROMINE THEOBROMINE MAGNESIUM OLEATE THEOPHYLLINE THIABENDAZOLE THIAMINE THIAMYLAL THIETHYLPERAZINE THIMEROSAL THIOGUANINE THIOPENTAL THIORIDAZINE THIOTEPA THIOTHIXENE THIPHENAMIL THREONINE THROMB IN THYMOL THYROGLOBULIN THYROID THYROTROPIN TICARCILLIN TICRYNAFEN TIMOLOL TOBRAMYCIN TOLAZAMIDE TOLAZOL INE 1 55880 55885 55890 55895 55900 55905 55910 55915 55925 55930 55935 55940 55945 55950 55955 55960 55965 55965 55970 55975 55985 55990 55995 55998 55998 56 000 56000 56005 56010 56015 56020 56025 56030 56040 56045 56050 56060 56065 56075 56080 56085 56090 56095 56100 56105 56110 56115 56120 56125 56130 56135 56140 56145 56150 56155 56160 56160 56165 56170 56175 TOLBUTAMIDE TOLMETIN TOLNAFTATE TRAGACANTH TRANYLCYPROMINE TRETINOIN TRIACETIN TRIAMCINOLONE TRIAMTERENE TRICHLORMETHIAZIDE TRICHLOROACETIC ACID TRICHLOROETHY LENE TRICLOFOS TRIDIHEXETHYL TRIETHANOLAMINE TRIFLUOPERAZINE TRIFLUOPERAZ INE TRIFLUPROMAZ INE TRIGLYCERIDES TRIHEXYPHENIDYL TRIMEPRAZINE TRIMETHADIONE TRIMETHAPHAN TRIMETHOPRIM W/SULFAMETHOXAZOLE TRIMETHOPRIM W/SULFASOXAZOLE TRIMETHOBENZ AMIDE TRIMIPRAMINE TRIMIPRAMINE TRIOXSALEN TRIPELENNAMI NE TRIPROLIDINE TRISULFAPYRIMIDINES TROL AMINE TROMETHAMINE TROPICAMIDE TRYPTOPHAN TUAMINCHEPTANE TUBERCUL IN TUBOCURARINE TURPENTINE TYBAMATE TYLOXAPOL TYPHOID VACCINE TYPHUS VACCINE TYROPANOATE SODIUM UNDECYLENIC ACID URACIL UREA URETHAN UROKINASE VALERIAN VALINE VALPROIC ACID VANCOMYCIN VANI LLIN LYPRESSIN VASOPRESSIN VERATRUM VIRIDE VIDARABINE VINBLASTINE 56180 56185 56190 56192 56193 56194 56195 56198 56205 56210 56220 56225 56230 56235 56240 56245 56250 56255 56260 56265 56270 56285 93905 INVENTORY OF GENERIC NAMES VINCRISTINE VITAMIN A VITAMIN A €D VITAMIN B-12 VITAMIN C VITAMIN D VITAMIN E VITAMIN K WARFARIN WATER STERILE WHITE LOTION WHITE PINE SYRUP WILD CHERRY SYRUP WINE XYLOMETAZOL INE XYLOSE YEAST YELLOW FEVER VACCINE YOHIMBENE ZINC TOPICAL AGENTS ZOLAMINE ZINC SULFATE NIACIN ce 00005 A & D VITAMIN 00010 A AND D 00015 A.A.S. 00020 A.C. A. 00025 A.D.C. VITAMIN DROPS 00030 A.P.Ce 00035 A.P.C. BUFFERED 00040 A.P.C. COMPOUND 00045 A.P.C. NO. 2 00050 A.P.C. NO. 3 00055 A.P.C. NO. & 00060 A.P.C. W/BUTALBITAL 00065 A.P.C. W/CODEI 00070 A.P.C. W/DEMEROL 00075 A.P.C. W/GELSEMIUM 00080 A.P.C. W/GELSEMIUM & PB 00085 A.P.C. W/MEPERIDINE HCL 00090 A.P.C. W/PHENOBARB 00095 A.P.l. 00100 A.S.A. 00105 A.S.A. & CODEINE 00110 A.S.A. COMPOUND 00115 A.S.A. NO. 2 00120 A.Se.A. NO. 3 00125 A.S.A. NO. 4 00130 A-CAINE 00135 A-CILLIN 00140 A-FIL 00145 A-HYDROCORT 00150 A-M-T 00155 A-METHAPRED 00160 A-POXIDE 00165 A-10 D-5-W 00170 A-200 PYRINATE 00175 A-5 D-5-W 00180 ABBOKINASE 00185 ABCDG 00190 ABDEC 00195 ABDOL 00200 ABDOL W/ VITAMIN C 00205 ABUK 00210 ACCELERASE 00215 ACCELERASE-PB 00220 ACCURBRON 00225 ACD 00230 ACEDOVAL 00235 ACEDYNE WAFER 00240 ACEPHEN 00245 ACETA 00250 ACETA W/ CODE INE 00255 ACETAGESIC 00260 ACETAMINOPHEN 00265 ACETAMINOPHEN NO. 2 00270 ACETAMINOPHEN NO. 3 00275 ACETAMINOPHEN NO. 4 00280 ACETAMINOPHEN W/ CODE INE 00285 ACETANILID 00290 ACETATED RINGER'S 00295 ACETAZOLAMIDE 00300 ACETEST MEDICATION CODE LIST, NAMCS 1980 00305 ACETIC ACID 00310 ACETIC ACID GLACIAL 00315 ACETOHIST 00320 ACETONE 00325 ACETOSPAN 00330 ACETYCOL 00335 ACETYLCHOLINE 00340 ACHROMYCIN 00345 ACHROMYCIN V 00350 ACHROSTATIN V 00353 ACIDIFYING AGENT 00355 ACI-JEL 00360 ACID MANTLE 00365 ACID PH 00370 ACID-EZE 00375 ACID-EZE EC 00380 ACIDULATED PHOSPHATE FLUORIDE 00385 ACIDULIN 00390 ACNAVEEN BAR 00395 ACNE 00400 ACNE-AID 00405 ACNE-DOME 00410 ACNEDERM 00415 ACNQ 00420 ACNOMEAD 00425 ACNOMEL 00430 ACODA ZEM 00435 ACON 00440 ACONITE 00445 ACOTUS 00450 ACRIFLAVINE 00455 ACTH 00460 _ACTHAR 00465 ACTICORT 00470 ACTIDIL 00475 ACTIFED 00480 ACTIFED-C 00485 ACTOL 00490 AD-CEBRIN W/FLUORIDE DROPS 00495 ADABEE 00500 ADABEE W/MINERALS 00505 ADAPETTES 00510 ADAPIN 00515 ADAPT 00520 ADEECON 00525 ADEFLOR 00530 ADENO 00535 ADENOSINE 00540 ADIPEX 00545 ADOPHYLLIN ELIXIR 00550 ADPHEN 00555 ADRENAL CORTEX 00560 ADRENALIN 00563 ADRENERGIC AGENT 00564 ADRENERGIC BLOCKING AGENT 00565 ADRENOSEM 00570 ADRIAMYCIN 00575 ADROYD 00580 ADRUCIL 00583 ADSORBENT 00585 00590 00595 00600 00605 00610 00615 00620 00625 00630 00635 00640 00645 00650 00655 00660 00665 00670 00675 00680 00685 00690 00695 00700 00705 00710 00715 00720 00725 00730 00735 00740 00745 00750 00755 00760 00765 00770 00775 00780 00785 00790 00795 00800 00805 00810 00815 00820 00825 00830 00835 00840 00845 ADSORBOCARP INE ADSCRBONAC ADSORBOTEAR AEROLATE AEROLONE AEROSEB-DEX AEROSEB-HC AEROS POR IN AFKO EAR DROPS AFKO-LUBE AFKO-LUBE LAX AFKO-LUBE SYRUP AFRIN AFRINOL REPETAB AFTATE AGORAL AHBID-12 AIR-TABS GG AIRET AIRET GG AKINETON AKNE DRYING LOTION AKNE ORAL AKRINOL AL-AY AL-0-MAG AL-R ALADRINE ALAMAG ALAMINE ALBALON ALBALON-A ALBAMEAD ALBAMYCIN ALBUCON ALBUCONN ALBUMINAR ALBUMISOL ALBUSPAN ALBUTEIN ALCAINE ALCOHOL ALCOHOL ABSOLUTE ALCOHOL ISOPROPYL ALCOHOL RUBBING ALCON ALCON-EFRIN ALCOPHOBIN ALDACTAZIDE ALDACT ONE ALDDCLOR ALDOCLOR-250 ALDOMET 00850 ALDORIL 00855 00860 00865 00870 00875 00880 ALERBUF ALERMINE ALEVAIRE ALGIGEL AL ISED ALKA-SELTZER 086L SOINVN ‘1sI7 8p0oD uonesipa\ “lil Xipueddy €€ 00885 ALKA-SELTZER PLUS 00890 ALKA-SELTZER WITHOUT ASPIRIN 00895 ALKA-2 00900 ALKALINE AROMATIC 00903 ALKALINI ZING AGENT 00905 ALKALOL 00910 ALKARAU 00915 ALKERAN 00920 ALKETS 00925 ALLBEE 00930 ALLBEE C 800 PLUS IRON 00935 ALLBEE W/C 00940 ALLBEE-T 00945 ALLERBEN 00950 ALLERBID 00955 ALLERCREME 00960 ALLEREST 00965 ALLEREST NASAL SPRAY 00970 ALLEREST TIME CAPSULE 00975 ALLERFRIN 00980 ALLERGY RELIEF OR SHOTS 00980 ANTIGEN INJECTION 00985 ALLERMINE 00990 ALLERNADE T.D. IMPROVED 00995 ALLERPHED 01000 ALLERPHED C EXPECTORANT 01005 ALLERPHED SYRUP 01010 ALLERPROP T.D. 01015 ALLERSONE 01020 ALLERSPAN 01025 ALLERSTAT 01030 ALLOPURINOL 01035 ALLYLGESIC W/ ERGOT AMINE 01040 ALMA-MAG NO. 4 01045 ALMEBEX PLUS B-12 01050 ALMERET 1000 01055 ALMEZYME 01060 ALMOCARPINE 01065 ALMOPHEN 01070 ALMORA 01075 ALO TUSS IMPROVED 01080 ALOPHEN PILL 01085 ALPHA CHYMAR 01090 ALPHA-KERI 01095 ALPHA-RUVITE 01100 ALPHADERM 01105 ALPHADROL 01110 ALPHALIN 01115 ALPHALIN GELSEAL 01120 ALPHAMUL 01125 ALPHAPRODINE 01130 ALPHAREDISOL 01135 ALPHATOCOPHEROL 01140 ALPHOSYL HC 01145 LTERN 01150 ALTO-PRED 01155 ALU-CAP 01160 ALU-MAG 01165 ALU-TAB 01170 ALUDINE MEDICATION CODE LIST, NAMCS 1980 01175 01180 01185 01190 01195 01200 01205 01210 01215 01220 01225 01230 01235 01240 01245 01250 01255 01260 01265 01270 01275 01280 01285 01290 01295 01300 ALUDROX ALUM ACETATE SOLUTIGN (BUROW"S) ALUM AMMONIUM LUMP ALUM AMMONIUM POWDER ALUM ALUM POWDER ALUMADR INE ALUMAG ALUMINETT ALUMINUM ALUMINUM HYDROXIDE ALUM HYDROX-MAG TRISILICATE ALUMINUM MAGNESIUM HYDROXIDE ALUMINUM PASTE ALUM SUBACETATE SOLUTION ALUM-MAG HYDROX W/SIMETHICONE ALUPENT ALURATE ELIXIR ALUSCOP ALUSIL ALUWETS WET DRESSING AM ZYME AM-PHED AMANTADINE AMARIL D AMA VIL 01305 AMBENY 01310 01315 01320 AMBODRYL AMCILL AMCORT 01325 01330 01335 01340 01345 01350 01355 01360 01365 01370 01375 01380 01385 01390 01395 01400 01405 01410 01415 AMEN AMERI EZP AMERICA INE AMER ICA INE DROPS AMERICAINE FIRST AID SPRAY AMER ICAINE LUBRICANT AMERIPEN AMER TAN AME SEC AMICAL FORMULA NO. 1 AMICAR AMID-SAL AMIDE V.C. VAGINAL AMIDE V.S. VAGINAL INSERT AMIDOXINE AMIGEN AMIKACIN AMIKIN AMIN-AID 1420 AMINO-CERV 01425 01430 01435 01440 01445 01450 01455 01460 01465 01470 AMINOACETIC ACID AMINODUR DURA-TAB AMINOCYNE AMINODYNE ELIXIR AMINOHIFPURATE AMINOPHYLLIN AMINOPHYLLINE AMINOPHYLLINE & AMYTAL AMINOPHYLL INE COMPOUND AMINOPHYLL INE EPHED AMOBARB 01475 AMINOPHYLL INE PHENOBARBITAL 01480 AMINOSALICYLATE SODIUM 01485 AMINOSALICYLIC ACID ENSEAL 01490 AMINOSOL 01495 AMINOSYN 01500 AMINPHYLLINE 01505 AMIPAQUE 01510 AMIRON 01515 AMITID 01520 AMITONE 01525 AMITRIL 01530 AMITRIPTYLINE 01535 AMITRIPTYLINE W/PERPHENAZINE 01540 AMIZINE 01545 AMLAX 01550 AMMENS 01555 AMMONIA AROMATIC ASPIROL 01560 AMMONIA AROMATIC SPIRIT 01565 AMMONIATED MERCURY 01570 AMMONIUM CHLORIDE 01575 AMMONIUM CHLORIDE RED 01580 AMNESTROGEN 01585 AMO-FED 01590 AMOBARB SECOBARB 01595 AMOBARBITAL 01600 AMOBARBITAL-EPHEDRINE 01605 AMODRINE 01610 AMOGEL PG 01615 AMOLIN 01620 AMONIDRIN 01625 AMOSTAT 01630 AMOXICILLIN 01635 AMOXICILLIN TRIHYDRATE 01640 AMOXIL 01645 AMPHED 01650 AMPHENE 01655 AMPHENOL 01660 AMPHETAMINE 01665 AMPHICOL 01670 AMPHOJEL 01675 AMPHOJEL W/MINERAL OIL 01678 AMPHOTERICIN 01680 AMPI-CO 01685 AMPICILLIN 01690 AMPICILLIN TRIHYDRATE 01695 AMQUIN KCL 10% SUGAR-FREE 01700 AMQUINTUSSIN 01705 AMQUINTUSSIN DM 01710 AMSED 01715 AMTET 01720 AMYL NITRITE 01723 AMYLOIDOSIS 01725 AMYTAL 01730 ANA EMERGENCY INSECT STING KIT 01735 ANABOL 01738 ANABOLIC AGENT 01740 ANABOLIN 01745 ANABOLIN LA-100 01750 ANACEL 01755 ANACIN ve 01760 ANADROL 01765 ANAIDS 01770 ANALBALM 01775 ANALGESIC 01780 ANALGESIC BALM 01785 ANALGESIC COMPOUND 01790 ANALGESIC EMULSION 01795 ANALGESIC LINIMENT 01800 ANALGESINE 01805 ANALGEST INE 01810 ANALGETS WAFER 01815 ANALONE 01820 ANAMINE 01825 ANANASE 01830 ANAPHEN 01835 ANAPHYLATIC SHOCK DRUG KIT 01838 ANAPROX 01840 ANASPAZ 01845 ANASPAZ-PB 01850 ANATUSS 01855 ANAVAR 01860 ANBESOL 01865 ANCEF 01870 ANCOBON 01875 AND-EST SUSPENSION 01880 ANDOIN 01885 ANDRIOL 01890 ANDRO CYP 01895 ANDRO-FEM 01900 ANDROID-G 01905 ANDROID-HCG 01910 ANDROID-T 01915 ANDROID-10 01920 ANDROID-25 01925 ANDROID-5 BUCCAL 01930 ANDROCLAN AQUEOUS 01935 ANDROLONE 01940 ANDROLONE D 100 01945 ANDROLONE D SO 01950 ANDRONAQ-50 01955 ANDRONATE 01960 ANDRYL 200 01965 ANDURACAINE 01970 ANECAL 01975 ANECTINE 01980 ANESTACON 01983 ANESTHETIC 01985 ANESTRO 01990 ANEXSIA W/CODEINE 01995 ANEXSIA 02000 ANG-O-SPAN 02005 ANGIDIL 02010 ANGIO-CONRAY 02015 ANHYDRON 02020 ANISE OIL 02025 ANISOTROPINE 02030 ANISOTROPINE W/PHENOBARB 02035 ANOCAINE 02040 ANODYNOS FORTE 02045 ANODYNOS-DHC MEDICATION CODE LIST, NAMCS 1980 02050 02053 02055 02060 02065 02070 02075 02080 02085 02088 02090 02095 02100 02105 02110 02113 02115 02120 02125 02130 02135 02140 02145 02150 02155 02157 02158 02160 02165 02170 02175 02180 02185 02190 02195 02200 02205 02210 02215 02220 02225 02230 02235 02235 02240 02240 02245 02250 02255 02260 02265 02270 02275 02280 02285 02290 02295 02300 02305 02310 ANOQUAN ANOREXIC AGENT ANOVO ANSEMCO ANSEMCO NO. 8 ANSPOR ANTABUSE ANTACID ANTACID #6 ANTACID AND ADSORBENT ANTAGONATE ANTALGESIC ANTAR ANTAR 11 ANTEPAR ANTHELMINTIC AGENT ANTHRA-DERM ANTHRAL IN ANTIACID ANTI TEN ANTI-ITCH ANTI-NAUSEA SUPPRETTE ANTI-THERM ANTI-TUSS ANTI-TUSS DM ANTIANEMIA AGENT ANTIBIOTIC AGENT ANTIBIOPTO ANTICOAGULANT ANTIHEMOPHILIC FACTOR HUMAN ANTILIRIUM ANTIMINTH ANTIPHLOGISTINE ANTIPRESS ANTIPYR INE ANTIRABIES SERUM ANTISEPTIC MOUTHWASH ANTISEP NO 3 R ST J PERRY ANTISEPTIC SOLUTION ANTISPAS ANTISPASMODIC ANTISPASMODIC COMPOUND ANTIVENIN SPIDER BITE SPIDER BITE ANTIVENIN ANTIVENIN SNAKE BITE SNAKE BITE ANTIVENIN ANTIVENIN MICRURUS FULVIUS ANTIVERT ANTORA ANTORA-B T.D. ANTRENYL ANTRIN ANTROCOL ANTUI TRIN "Ss" ANTURANE 2290 ANUGESIC ANUJECT ANULAN ANUPHEN ANUSOL 02315 02320 02325 02330 02335 02340 02345 02350 02355 02360 02365 02370 02375 02380 02385 02390 02395 02400 ANUSOL -HC APAC APACOMP APADON APAP APAP W/CODEINE APAP W/CODEINE ELIXIR AP AP W/PHENYLPROPANOLAMINE APATATE APHCO HEMORRHCIDAL APHONALS APLISOL APLITEST APOMORPHINE APPET-AID APPET-IRON APRESAZIDE APRESODEX 02405 APRESOLINE 02410 APRESOLINE-ESIDRIX 02415 APRISAC 02420 APTROL 02425 AQUA TABS 02430 AQUA-DUCE 02435 AQUA-TON 02440 AQUA-TON-S 02445 AQUACARE 02450 AQUACARE/HP 02455 AQUACHLORAL SUPPRETTE 02460 AQUALIN SUPPRETTE 02465 AQUAMEPHYTON 02470 AQUAPHOR 02475 AQUAPRES 02480 AQUASERP 02485 AQUASOL A 02490 AQUASOL BODY LOTION 02495 AQUASOL E 02500 AQUASTAT 02505 AQUATAG 02510 AQUATENSEN 02515 AQUEX 02520 ARALEN 02525 ARALEN PHOSPHATE 02530 ARALEN W/PRIMAQUINE 02535 ARAMINE 02540 ARCO-CEE 02545 ARCO-LASE 02550 ARFONAD 02555 ARGININE 02560 ARGYROL S.S. 02565 ARIDOSE 02570 ARISTO-PAK 02575 ARISTOCORT 02580 ARISTOCCRT A 02585 ARISTOCORT FORTE 02590 ARISTOCORT HP 02595 ARISTOCORT INTRALESIONAL 02600 ARISTOCORT LP 02605 ARISTOCORT R 02610 ARISTOGEL Ge ARI STOSPAN 02615 02620 ARITHMIN 02625 ARLIDIN 02630 ARNICA 02635 ARC-PEPSIN 02640 AROMATIC SPIRITS OF AMMONIA 02645 ARTANE 02650 ARTHRALGEN 02655 ARTHRIN 02660 ARTHRITIS PAIN FORMULA 2665 ARTHROLATE 02670 ARTHROPAN 02675 ARTRA SKIN TONE 02680 AS-CA-PHEN 02685 ASAFETIDA 02690 ASBRON G 02695 ASCODEEN-30 02700 ASCOR-B-SOL W/D-5-W 02705 ASCORBIC ACID 02710 ASCORBIC ACID SYRUP 02715 ASCORBICAP 02720 ASCORVITE 02725 ASCRIPTIN 02730 ASCRIPTIN NO. 2 02735 ASCRIPTIN NO. 3 02740 ASCRIPTIN W/ CODE INE 02745 ASELLACRIN 02750 ASMA SYRUP 02755 ASMA-LIEF 02760 ASMAC 02765 ASMACOL 02770 ASMADIL UNICELLE 02775 ASMALIX ELIXIR 02780 ASMINOREL IMPROVED 02785 ASMINYL 02790 ASPERGUM 02795 ASPHAC-G 02800 ASPIRBAR 02805 ASPIRIN 02810 ASPIRIN COMPOUND 02815 ASPIRIN COMPOUND #2 02820 ASPIRIN COMPOUND #3 02825 ASPIRIN COMPOUND W/CODE INE 02830 ASPIRIN W/PHENOBARBITAL 02835 ASPIRIN-PHENACET IN-CAFFEINE 02840 ASPIRIN-SECOBARBITAL SODIUM 02845 ASPROJECT 02850 ASTHMACON 02855 ASTRING-0-SOL 02860 ASTRINGENT 02865 ASULAM VAGINAL 02870 ATABRINE 02875 ATARAX 02880 ATARAXOID 02885 ATHEMOL 02890 ATHEMOL-N 02895 ATHROMBIN-K 02900 ATIVAN 02905 ATOKA 02910 ATRIDINE MEDICATION CODE LISTs NAMCS 1980 02915 02920 02925 02930 02935 02940 02945 02950 02955 02960 02965 02970 02975 02980 02985 02990 02995 03000 03005 03010 03015 03020 03023 03025 03030 03035 03040 ATROBYL ATROCHOL IN ATROMID-S ATROPHYSINE ATROPINE & PHENOBARBITAL ATROPINE ATROPINE MURO ATROPINE SULFATE ATROPINE SULF & MEPERIDINE ATROPINE SULFATE HT ATROPINE SULFATE MUROCOLL ATROPINE SULFATE 21 GA ATROPINE SULFATE 22 GA ATROPISOL ATTENUVAX AURAL ACUTE DRCPS AURALGAN AURASOL AUREOMYCIN AUREOMYCIN STERILE AUR INOL AUROTO DROPS AUTONOMIC AGENT AVALGESIC LOTION AVAZYME AVAZYME-100 AVC 03045 03050 03055 03060 03065 03070 03075 03080 03085 03090 03095 03100 03105 03110 03115 03120 03125 03130 03135 03140 03145 03150 03155 03160 03165 03170 03175 03180 03185 03190 03195 03200 03205 AVC/DIENESTROL AVEENO AVEENO LOTION AVEENO OILATED AVEEND-BAR AVENTYL HCL AVITENE AVLOSULFON AW-CILLIN AVP-NATAL AVP-NATAL-FA AVSUL VAGINAL AXON AXOTAL AYR SALINE MIST AZAPEN AZ ENE AZMA-AID AZO GANTANOL AZO GANTRISIN AZO METHALATE AZO-MANDELAMINE AZO-MED AZO-SOXAZOLE AZO-STANDARD AZO-SULFISOCON AZO-SULFISOXAZOLE AZO-SULFIZIN AZO-SULFSTAT AZ0-100 AZODINE AZOLATE AZOLID 03210 AZOSTIX 03215 AZOSUL 03220 AZOTREX 03225 AZULFIDINE 03230 A5-D-5-W 03235 B & A HYGIENIC 03240 B & C VITAMIN 03245 B & 0 SUPPRETTE 03250 B COMPLEX 03255 B COMPLEX #100 03260 B COMPLEX B-12 W/C 03265 B COMPLEX FORTE 03270 B COMPLEX HIGH POTENCY 03275 B COMPLEX W/ASCORBIC ACID & B-12 03280 B COMP W/ASCORB ACID & B-12-200 03285 B COMPLEX W/B-12 03290 B COMPLEX W/B-12 VIT C & LIVER 03295 B COMPLEX W/VITAMIN C 03300 B OF 03305 B VITAMINS W/C 03310 B.F.I. 03315 B.P.E. 03320 Be.P.P.-LEMMON 03325 B-A 03330 B-C-BID 03335 B-C—E & ZINC 03340 B-COMPLEX LIVER & IRON 03345 B-COMPLEX W/B-12 & VITAMIN C 03350 B-NUTRON 03355 B-12 03360 B-12 THIAMINE 03365 BABY COUGH SYRUP 03370 BABY MINS 03375 BABY OINTMENT 03380 BABY POWDER JOHNSON'S 03385 BAC-NEO-POLY 03390 BACARATE 03395 BACID 03400 BACIGUENT 03405 BACIMYCIN 03410 BACITRACIN 03415 BACITRACIN-NEOMYCIN-POLYMYXIN 03420 BACITRACIN-POLYMYXIN 03423 BACLOFEN 03425 BACTOCILL 03430 BACTRIM 03435 BACTRIM DS 03440 BAFIL 03445 BAKER'S BEST HAIR LOTION 03450 BAKER'S INFANT FORMULA 03455 BAKER'S READY-4 BOUILLON 03460 BAKER'S READY-4 COLA WATER 03465 BAKER'S READY—-4 GELATIN WATER 03470 BAKER'S READY-4 TEA 03475 BAL IN OIL 03480 BALNEOL LOTION 03485 BALNETAR 03490 BALSAM PERU NF 9 BALTRO 03500 BAMATE 9€ 03505 BAMO 03510 BANALG LINIMENT 03515 BANCAP 03520 BANCAP W/CODEINE 03525 BANESIN 03530 BANGESIC LINIMENT 03535 BANIGESIC 03540 BANTHINE 03545 BARBATOSE #2 03550 BARBATRO NO. 2 03555 BARBELOID 03560 BARBIDONNA 03565 BARBIPIL 03570 BARBITA 03575 BARBITAL 03580 BARISOL DROPS 03585 BARIUM ENEMA PREP KIT 03590 BARI UM 03595 BARODENSE 03600 BAROLOID 03605 BAROPHEN ELIXIR 03610 BAROPLEX ELIXIR 03615 BAROSPERSE 03620 BAROTRAST 03625 BARSEB 03630 BARTRATE 03635 BASALJEL 03640 BASIC DROPS 03645 BASIS SOAP 03650 BASIS SOAP W/ GLYCERIN 03655 BASIS SOAP W/SULFUR 10% 03660 BAUMODYNE 03665 BAXIMIN 03670 BAXINETS 03675 BC 03680 BC-LONG 03685 BCG VACCINE 03690 BE-CE-PLEX FORTE 03695 BECAUSE 03700 BECLOMETHASONE 03705 BECLOVENT 03710 BECOPLEX FORTIFIED 03715 BECOTIN 03720 BECOTIN W VITAMIN C 03725 BECOTIN-T 03730 BEECEEPLEX 03735 BEEF IRON AND WINE 03740 BEELITH 03743 BEE STING ANTIVENIN 03745 BEESIX 03750 BEEZE W/ VITAMIN C 03755 BEJECTAL 03760 BEJECTAL W/ VITAMIN C 03765 BEJEX 03770 BELAP 03775 BELEXAL 03780 BELFER 03785 BELL ADENAL 03790 BELL ADENAL-S 03795 BELLADOL ELIXIR MEDICATION CODE LIST, NAMCS 1980 03800 03805 03810 03815 03820 03825 03830 03835 03840 03845 03850 03855 03860 03865 03870 03875 03880 03885 03890 03895 03900 03905 03910 BENAI 03915 03920 03925 03930 03935 03940 03945 03950 03955 03960 03965 03970 03975 BELLAD ONNA BELLADONNA ALKALOIDS W/PB BELLAD BELLAD BELLAD ONNA EXTRACT ONNA FORMULA ONNA FORM W/HYDROCORT BELLADONNA W/PHENOBARBITAL BELLAF BELLAF BELLAL BELLER EDROL A-H OL INE PHEN GAL BELLERGAL-S * BELLERGOTAL BELLKA TAL BELLO-PHEN BELLOPHEN BELPHE BEMEX BEMINA BEM INA BEMINA N NO. 1 L L FORTE W/VITAMIN C L FORM W/IRON & LIVER BEN GAY BENADR YL BENADRYL W/EPHEDRINE SULFATE BENADYNE DROPS BENAHI BENASE BENDEC ST PT VAGINAL GEL TIN BENDOPA BENDYL BENEDI ATE CT'S QUALITATIVE SOLUTION BENEGYN BENEMI D BEN ISONE BENOJECT BENDQUIN BENOXI NATE 03980 BENOXYL BENSULFOID BENSULFOID LOTION BENSULFOID POWDER 03985 03990 03995 04000 04005 BENTYL BENTYL W/PHENOBARBITAL 04010 BENYL N DM COUGH SYRUP 04015 04020 04025 04030 04035 04040 04045 04050 04055 04060 04065 04070 04075 04080 04085 04090 04095 BENYL BENZAC BENZAC N_ SYRUP W GEL BENZAGEL BENZAL KONIUM BENZEDREX INHALER BENZEDR INE BENZO BENZOC MENTH AINE BENZOCOL BENZODENT BENZOI BENZ OI N N COMPOUND BENZOMEAD PLUS THROAT DISC BENZOYL BENZTHIAZ IDE BENZYL 04100 04105 04110 04115 04120 04125 04130 04135 04140 04145 04150 04155 04160 04165 04170 04175 04180 04185 04190 04195 04200 04205 04210 04215 04220 04225 04230 04235 04240 04245 04250 04255 04260 04265 04270 04275 04280 04285 04290 04295 04300 04305 04310 04315 04320 04325 04330 04335 04340 04345 04350 04355 04360 04363 04365 04368 04368 04370 04375 04380 BEROCCA BEROCCA-C BEROCCA—C 500 BERUBIGEN BETADINE BETADINE PERINEAL WASH BETADINE VAGINAL DOUCHE BETADINE VAGINAL GEL BETALIN COMPLEX BETALIN COMPLEX ELIXIR BETALIN COMPLEX F.C. BETALIN COMPOUND BETALIN BETALIN 12 CRYSTALLINE BETAMETHAS ONE BETAPAR BETAP EN-VK BETA PRONE BETHANEC HOL BETULINE LINIMENT BEWON ELIXIR BEXIBEE BEXOMAL-C BEXOPHENE BIAVAX II BICARBONATE OF SODA BICHLORACETIC ACID BICILLIN BICILLIN C-R BICILLIN LONG-ACTING BICITRA BICNU BIFED 20 BILAMIDE BILE ACIDS MIXED BILE SALTS BILEZYME BILOGEN BILOPAQUE SODIUM BILRON BINEX-C BIO-DES BIO-SORB BIOFLAVONOIDS BIOLAX BIOPAR FORTE BIOSONE BIOTHESIN BIOTIC-0 BIOTIN BI0T RES BIOZYME BIPECTOL WAFER BIPER IDEN BIPHETAMINE BIRTH CONTROL MEDICATION CONTRACEPTIVE AGENT BISACODYL BISACODYL PATIENT PACK BISCO LAX nun 4 04385 BISCODYL 04390 BISCOLAN 04395 BISCOLAN HC 04400 BISILAD 04405 BISMAPEC 04410 BISMUTH 4415 BISMUTH PAR 04420 BISMUTH SUBCARBONATE 04425 BISMUTH SUBGALLATE 04430 BISMUTH SUBGALL W/COD LIV OIL 04435 BISMUTH SUBGALL W/SHK LIV OIL 04440 BISMUTH SUBSALICYLATE 04445 BISMUTH VIOLET 04450 BISMUTH VIOL-SALCYL & BENZ ACDS 04455 BITRATE 04460 BLACK & WHITE BLEACHING 04465 BLANK (PLACEBO) HT 04470 BLEACHING PEROXIDE 20 VOLUME 04475 BLENOXANE 04478 BLEOMYCIN 04480 BLEPH 04485 BLEPHAMIDE 04490 BLINK-N-CLEAN 04495 BLINX 04500 BLISTAID COLD SORE LOTION 04505 BLISTEX 04510 BLISTIK BALM 04513 BLOOD FORMATION AND COAGULATION 04515 BLUBORO 04520 BLUE GEL 04525 BLUE-GRAY 04530 BLUESTONE 04535 BMC-TEST MECONIUM 04540 BNC STANDARD 04545 BNC SUPER 0455Q BO-—CAR-AL 04555 BOIL N SOAK 04560 BON-A-DAY 04565 BON-A-DAY W/ IRON 04570 BON-A-DAY W/MINERALS 04575 BONATE 04580 BONINE 04585 BONTRIL _PDM 04590 BORAX 04595 BORIC ACID 04600 BOROFAX 04605 BOULTON'S SOLUTION 04610 BOWDRIN 04615 BOWSTEROL (ANTI-RUST DIS INF) 04620 BOWTUSSIN D.M. SYRUP 04625 BOWTUSSIN SYRUP 04630 BPN 04635 BRASIVOL 04640 BREMIL READY-TO-FEED 04645 BREOKINASE 04650 BRETHINE 04655 BRETYLOL 04660 BREVICON 04665 BREVITAL 04670 BRICANYL MEDICATION CODE LIST, NAMCS 1980 04675 BRILLIANT GREEN 04680 BRISTACYCLINE 04685 BRISTAMYCIN 04690 BRO-LAC 04695 BRO-TANE EXPECTORANT 04700 BRO-TAPP 04705 BROCON 04710 BROHEMBIONE 04715 BROM-CORT EXPECTORANT #1 04720 BROM-CORT-DC EXPEC TORANT #2 04725 BROM-CORTAPP 04730 BROM-PHENIRAMINE 04735 BROMALATE D.C. EXPECTORANT 04740 BROMALATE 04745 BROMALIX ELIXIR 04750 BROMAMINE 04755 BROMANATE DC EXPEC TORANT 04760 BROMANATE 04765 BROMANATE EXPECTORANT 04770 BROMANYL EXPECTORANT 04775 BROMATANE 04780 BROMATANE DC EXPECTORANT 04785 BROMATANE EXPECTORANT 04790 BROMATAPP 04795 BROMEPAPH 04800 BROMEPHEN ELIXIR 04805 BROMO-SELTZER 04808 BROMOCRIPTINE 04810 BROMOPHEN TeDe. 04815 BROMPHEN DC EXP W/CODEINE 04820 BROMPHEN 04825 BROMPHENIRAMINE COMPOUND ELIXIR 04830 BROMPHENIRAMINE COMP EXP 04835 BROMPHENIRAMINE COMPOUND S.A. 04840 BROMPHENIRAMINE EXP SYRUP 04845 BROMPHENIRAMINE PHENIRAMINE MALEATE ELIXIR 04855 BROMPHENIRAMINE W/COD DC EXP 04860 BROMPHENTAPP T.D. 04865 BRONCHLOFORM SYRUP 04870 BRONCHOBID 04875 BRONDECON 04880 BRONDELATE 04885 BRONITIN 04890 BRONITIN MIST 04895 BRONKAID 04900 BRONKAID MIST 04905 BRONKEPHRINE 04910 BRONKODYL 04915 BRONKOLIXIR 04920 BRONKOMETER 04925 BRONKOSOL 04930 BRONKOTABS 04935 BROPHENTAPP 04940 BROWN MIXTURE 04945 BROWN MIXTURE AMMONI ATED 04950 BSS 04955 BU-LAX 04960 BU-LAX PLUS 04965 BUCHU ELIXIR 9 04975 04980 04985 04990 04995 -S SOF BUF ACNE CLEANSING BAR BUFF FULL STRENGTH INCERT BUFF HALF STRENGTH INCERT BU FF-A BUFF-A-COMP 05000 BU FFADYNE 05005 BUFFERED SOLUTUON ISOTONIC FF 05015 BUFFERIN ARTHRITIS STRENGTH 05020 BUFFEX 05025 BUMINATE 05030 BUMINTEST 05035 BUPIVACAINE 05040 BUREN (IMPROVED) 05045 BURO-SOL ANTISEPTIC 05050 BUROW'S OINTMENT 05055 BUROW'S SOLUTION 05060 BUSULFAN 05065 BUTA-KAY ELIXIR 05070 BUTABARBITAL 05075 BUTABARBITAL-BELLADONNA ELIXIR 05080 BUTABELL-HMB 05085 BUTAL 05090 BUTALAN ELIXIR 05095 BUTALBITAL 05100 BUTALBITAL W/A.P.C. 05105 BUTALIX 05110 BUTAZOLIDIN 05115 BUTAZOLIDIN ALKA 05120 BUTESIN PICRATE 05125 BUTIBEL 05130 05135 05140 05145 05150 BUTIBEL-ZYME BUTICAPS BUTIGETIC BUTISOL BUTSECO 05155 BUTYN 05160 05165 05170 05175 05180 05185 05190 05195 05200 05205 05210 05215 05220 05225 05230 05235 05240 05245 05250 C&T C TUSSIN EXPECTORANT C.DeM. EXPECTORANT C.V.P. C-B VONE C-FLAVONOIDS C-LONG GRANUCAP C-PLEX C-RON C-RON FA C-RON FORTE C-STIX C-TABS C-100 PLUS CABADON M CAFACETIN CAFECON CAFENAMINE CAFERGOT 05255 CAFERGOT P-B 05260 05265 CAFETRATE CAFFEINE 8¢ 05270 05275 05280 05285 05290 05295 05300 05305 CAFFEINE & SODIUM BENZOATE CAFFEINE ALKALOID CAFFEINE CITRATED CAL PRENAL CAL PRENAL RX CAL-C-BATE CAL-NOR CAL-20 05310 CALADRYL 05315 05320 05325 05330 05335 CALAMATUM CALAMINE CALAMINE COMPOUND PASTE CALAMINE LINIMENT CALAMINE LOTION 05340 05345 05350 05355 05360 05365 05370 05375 05380 05385 05390 05393 05395 05400 05405 05410 05415 05420 05425 05430 05435 05440 CALAMINE LOTION W/ PHENOL CALAMINE OINTMENT CALCEE CALCET CALCICAPS CALCICAPS W/ IRON CALCIDRINE CALCIFEROL CALCIHAB CALCILAC CALCIMAR CALCITONIN CALCIUM ACETATE CALCIUM BROMIDE CALCIUM CARBONATE CALCIUM CARBONATE W/ATROPINE CALCIUM CALCIUM DISODIUM VERSENATE CALCIUM GLUCEPTATE CALCIUM GLUCONATE CALCIUM GLUCONATE W/VITAMIN D CALCIUM 10DI ZED 05445 05450 05455 05460 05465 05470 05475 05480 05485 05490 05495 05500 05505 05510 05515 05520 05525 05530 05533 05535 05540 05545 CALCIUM LACTATE CALCIUM PANTOTHENATE CALCIUM STEARATE NF CALC IUM-AMINO CALCIUM-D CALCIUM, PHOSPHATE & VITAMIN D CALCIWAFERS CALDECORT CALDESENE CALEATE CALFOS D CALGLYCINE CALINATE-FA CALMOL 4 CALOCARB CALOMEL CALOXOL LOTION CALPHOSAN CALORIC AGENT CALSCORBATE CAL SUX APHEN CALURIN 05550 CAMA 05555 CAMALOX MEDICATION CODE LIST, NAMCS 1980 05560 CAMOQUIN 05565 CAMPHO-PHENIQUE 05570 CAMPHOR 05575 CAMPHOR & SOAP LINIMENT 05580 CAMPHORATED OIt 05585 CANDEPTIN 05590 CANDEX 05595 CANTHARIDIN 05600 CANTIL 05605 CANTIL W/PHENOBARBITAL 05610 CANTRI 05615 CANZ 05620 CAPADE 05625 CAPASTAT 05630 CAPEX 05635 CAPITAL 05640 CAPITAL W/CODE INE 05645 CAPITROL 05650 CAPRON 05655 CAPSOLIN 05660 CAQUIN 05665 CARAMEL 05670 CARBACEL 05675 CARBACHOL 05680 CARBAMAZEPINE 05685 CARBARSONE 05690 CARBENICILLIN 05695 CARBOCAINE 05700 CARBOL-FUCHSIN 05705 CARBOLIC ACID SOAP 05710 CARBON DIOXIDE 05715 CARBON TETRACHLORIDE 05720 CARBONIS DETERGENS USP 05725 CARBOWAX 400 05730 CARBRITAL 05735 CARBROGESIC 05740 CARDABID 05745 CARDAMOM COMPOUND 05750 CARDEC-DM 05755 CARDEC-S SYRUP 05758 CARDIAC AGENT 05760 CARDILATE 05765 CARDILATE-P 05770 CARDIO-GREEN 05775 CARD-PULMON RESUS DRUG KIT 05780 CARDIOGRAFIN 05785 CARDIOCQUIN 05788 CARDIOVASCULAR AGENT 05790 CARDRASE 05795 CARDUI 05800 CARI-TAB SOFTAB 05805 CARIPEPTIC LIQUID 05810 CARISOPRODOL 05815 CARISOPRODOL COMPOUND 05820 CARMOL 05825 CARMOL-HC 05830 CAROID 05835 CARQCID AND BILE SALTS 05840 CARTER®S LITTLE PILLS 05845 CARTRAX 05850 05855 05860 05865 05870 05875 05880 05885 05890 05893 05895 05900 05903 05905 05910 05915 05920 05925 05930 05935 05940 05945 05950 05955 05960 05965 05970 05975 05980 05985 05990 CAS-EVAC CASAFRU CASCARA CASCARA COMPOUND CASCARA SAGRADA CASEC CASTADERM CASTOR OIL CASYLL IUM CEFADROX IL CATAPRES CATARASE CATHARTIC AGENT CAUSALIN CE-B ZINC CE-vI-SOL CEBEFORTIS CEBENASE CEBETINIC CEBEX CEBO-CAPS CEBRAL CEBRALAN CECLOR CECON DROPS CEDILANID CEDILANID-D CEENU CEETOLAN CEFADYL E FOL 05995 FAZOL IN 05995 06000 06005 06010 06015 06020 06023 06025 06030 06035 06040 06045 06050 06055 06060 06065 06070 06075 06080 06085 06090 06095 06100 06105 06110 06115 06120 06125 EFOZOL IN ELBENIN ELESTONE PHOSPHATE ELESTONE SOLUSPAN CELLOTHYL CELL STIMULANT AND PROLIFERANT CELOID CELONT IN CEN-E CENA K 10% SUGAR-FREE CENAC CENAFED CENAHIST CENAID CENALAX CENALENE CENALONE CENOCORT CENOLATE CENT ET-250 CENTRAX CENTRUM CEC-TWO CEPACOL CEPACOL ANESTHETIC TROCHE CEPASTAT CEPHALEXIN 6€ 06127 06128 06130 06133 06135 06140 06145 06150 06155 06160 06165 06170 06175 06180 061 85 06190 06195 06200 062 05 CEPHALORIDINE CEPHALOSPORINS CEPHALOTHIN CEPHRADINE CEPHULAC CEREBID CEREBRO-NICIN CERESPAN CERIUM CEROSE CERUBIDINE CERUMENEX CERYLIN CETACAINE CETACORT CETAMIDE CETANE CETANE W/PRESERVATIVE CETANE-NO PRESERVATIVE 06210 CETAPHIL 06215 CETAPRED 06220 CETRO-CIROSE 06225 06230 06235 06245 06250 06255 06260 06265 06270 06275 CETYL ALCOHOL CETYLCIDE CEVALIN CEVI-BID CEVITA CEVITA KAYKAP CHAP STICK CHARCOAL CHARCOCAPS CHARCODATE CHARCOTABS 06280 _CHARDONNA-2 06285 06290 06295 CHEALAMIDE CHEK-STIX CHEL -I RON 06300 CHERACOL 06305 06310 06315 06320 CHERACOL SYRUP CHERALIN CHERALIN W/CODEINE CHERRALEX 06325 CHERRALEX W/CODE INE 06330 06335 06340 06345 06350 06355 06360 06365 06370 06375 06380 06385 06390 06395 CHERRI-B CHERRY SYRUP CHEW E CHEW VITE CHEW-TABS CHEW-VITE CHEXIT TIMED-RELEASE CHIGGEREX CHIGGERTOX CHLOR MAL SYRUP CHLOR-HAB CHLOR-PHED CHLOR-PHENT ERMINE CHLOR-RES 06400 CHLOR-TRIMETON 06410 HLOR-TR CHLOR-TRIMETON EXPECTORANT MEDICATION CODE LIST, NAMCS 1980 06415 CHLOR-TRIMETON REPETAB 06420 CHLOR-TRIMETON SYRUP 06425 CHLOR-100 06430 CHLORAFED 06435 CHLORAFED TIMECELLE 06440 CHLORAL HYDRATE 06445 CHLORAL-METHYLOL 06450 CHLORAMATE 06455 CHLORAMBUCIL 06460 CHLORAMEAD 06465 CHLORAMPHENICOL 06470 CHLORASEPTIC 06475 CHLORASEPTIC CHILDREN'S 06480 CHLORASEPTIC DM 06485 CHLORASEPTIC GEL 06490 CHLORDIAZACHEL 06495 CHLORDIAZEPOXIDE 06500 CHLORDIAZEPOXIDE W/CLIDIN BROM 06505 CHLORESIUM 06510 CHLOROBUTANOL 06515 CHLOROBUTANOL HYDROUS 06520 CHLOROFON-F 06525 CHLOROFORM 06530 CHLORDHIST SPRAY 06535 CHLOROMYCETIN 06540 CHLOROMYCETIN OTIC 06545 CHLOROMYCETIN PALMITATE 06550 CHLOROMYCETIN SODIUM SUCCINATE 06555 CHL OROMYCETIN-HYDROC ORTISONE 06560 CHLOROMYXIN 06565 CHLOROPHYLL 06570 CHLOROPTIC 06575 CHLOROPTIC-P 06580 CHLOROQUINE 06585 CHLOROSERPINE 06590 CHLOROTHIAZIDE 06595 CHLOROTHIAZIDE W/RESERPINE 06600 CHLOROTHYMOL 06605 CHORPHENIRAMINE 06610 CHLORPHENIRAMINE 2 MG W/SPC 06615 CHLORPRO 06620 CHLORPROMAZINE 06625 CHLORPROPAMIDE 06630 CHLORSPAN-12 06635 CHLORTAB 06640 CHLORTAB-8 06645 CHLORTHALIDONE 06650 CHLORULAN 06655 CHLORZINE 06660 CHLORZONE FORTE 06665 CHLORZOXAZONE W/APAP 06670 CHO-FREE 06675 CHOLAN 06680 CHOLAN HMB, 06685 CHOLAN V 06690 CHOLEBRINE 06695 CHOLEDYL 06700 CHOLERA VACCINE 06705 CHOLESTEROL 06710 CHOLESTYRAMINE RESIN DRIED 06715 CHOLEX EMULSION 06720 CHOLINE 06725 CHOLINE CHLORIDE 06730 CHOLINE SALICYLATE 06733 CHOLINERGIC AGENT 06734 CHOLINERGIC BLOCKING AGENT 06735 CHOLINOL 06740 CHOLOGRAFIN 06745 CHOLOXIN 06750 CHOOZ 06755 CHOREX 06760 CHORIONIC G 06765 CHORIONIC GONADOTROPIN 06765 GONADATROPIN 06765 HCG 06770 CHORIONIC SPECIAL DILUENT 06775 CHRISTODYNE DHC 06780 CHROMAGEN 06785 CHROMAGEN-D 06790 CHROMELIN COMPLEXION BLENDER 06795 CHROMIC OXIDE 06800 CHRONULAC SYRUP 06805 CHYMORAL 06810 CIDEX 06815 CIMETIDINE 06820 CIN-QUIN 06825 CINNAMON OIL USP 06830 CINNASIL 06835 CIN 06840 CIRAMINE 06845 CIRCANOL SUBLINGUAL 06850 CIRIN 06855 CITANEST 06860 CITRATE OF MAGNESIA 06865 CITRIC ACID 06870 CITROCARBONATE GRANULAR 06875 CITRONELLA OIL 06880 CITROVAL 06885 CLEAN-N-SOAK 06890 CLEANING & SOAKING SOL BARNES 06895 CLEAR EYES 06900 CLEARASIL 06905 CLEOCIN 06910 CLEOCIN PHOSPHATE 06915 CLERA 06920 CLINDAMYCIN 06925 CLINDAMYCIN (PHOSPHATE) 06930 CL INDEX 06935 CLINORIL 06940 CLIPOXIDE 06945 CLISTIN 06950 CLISTIN EXPECTORANT 06955 CLISTIN R-A 06960 CLISTIN-D 06965 CLOCREAM 06970 CLOFIBRATE 06975 CLOMID 06980 CLONAZEPAM 06985 CLONIDINE 06990 CLONOPIN ov 06993 CLORAZEPATE 06995 07000 07005 07010 07015 07020 07025 07030 07035 07040 07045 07050 07055 07060 07065 07070 07075 07080 07085 07090 07095 07100 07105 07110 CLORAZINE CLORPACTIN CLOVE OIL CLOXACILLIN CLOXAPEN CLUSIvVOL CLUSIVOL SYRUP CLYSODRAST CM W/PAREGORIC CO TINIC CO-APAP CO-GEL CO-LU-GEL M-T CO-PYRONIL CO-SALT CO-XAN SYRUP COAL TAR COASTALDYNE COASTALGESIC COBADOCE FORTE COBALASINE COBALIN COBALPHAMEAD COCAINE 07115 COCCIDIOIDIN SKIN TEST 07120 07125 07130 07135 07140 07145 07150 07155 07160 07165 07170 07175 07180 COCILLANA COCILLANA COMPOUND SYRUP COCI LLANA COMP SYR W/ CODEINE COCO-QUININE COCOA BUTTER COCONUT OIL COD LIVER OIL CODALAN CODANOL CODAP CODASA 30 MG CODE HIST CODE INE 07185 CODE INE PHOSPHATE 07190 CODE INE SULFATE 07195 07200 07205 07210 07215 07220 07225 07230 07235 07240 07245 07250 07255 07260 07265 07 07 bs 7270 1275 7280 07285 CODI MAL CODIMAL DH SYRUP CODIMAL DM SYRUP CODI MAL EXPECTORANT CODI MAL L.A. CODIMAL PH SYRUP CODITRATE CODI VAL CODONE CODR OX OMIN CODY LAX COGENT IN COHIDRATE COL-VI-NOL COLACE COLBENEMID COLCHICINE COLCHICUM COLD CAPSULE MEDICATION CODE LIST, NAMCS 1980 07290 07295 07300 07305 07310 07315 07320 07325 07330 07335 07340 07345 07350 07355 07360 07365 07370 07375 07380 07385 07390 07395 07400 07405 07410 07415 07420 07425 07430 07435 07440 07445 07450 07455 07460 07465 COLD CREAM COLD SORE LOTION COLD TABLET COLD TABLETS COLDATE COLESTID COLLAGENASE ABC COLLODION COLLYRIUM COLLYRIUM W/EPHEDRINE COLOCTYL COLOGEL COLONAID COLREX COLREX COMPOUND COLREX COMPOUND ELIXIR COLREX DECONGE STANT COLREX EXPECTORANT SYRUP COLREX SYRUP COLREX TROCHE COLSALIDE COLSPAN COLTAB COL Y-MYC] COLY=-MYC] COLY-MYCIN COM-VI PLUS COMBEX COMBID COMBIPRES COMBISTIX COMEBACK COMFOLAX COMFOLAX-PLUS COMFORTINE wn S PEDIATRIC COMHIST 07470 07475 07480 07485 07490 07495 07500 07505 07510 07515 07520 07525 07530 07535 07540 07545 07550 07555 07560 07565 07570 07575 07580 07585 COMPAZINE COMPOBARB COMTREX COMYCIN CONACETOL CONAR CONCEP TROL CONEX CONEX D.A. CONEX PLUS CONEX W/CODE INE CONGESPIRIN SYRUP CONGO RED CONHIST CONJEC CONRAY CONSOTUSS CONSTIBAN CONTAC CONTAC JR. CONTAC NASAL MIST CONTACTISOL CONT IQUE CONTIQUE CLEANING + SOAKING 07590 07595 07600 07605 07610 07615 07620 07625 07630 07635 07640 07645 07650 07655 07660 07665 07670 07675 07680 07685 07690 07695 07700 07705 07710 07715 07720 07725 07730 07735 07740 07745 07750 07755 07760 07770 07775 CONT IQUE SOAKING CONT IQUE WETTING CONTROFLEX SYRUP CONVERSPAZ CONERSPAZ IMPROVED CONVERZYME COPAVIN COPPERTONE COPROBATE COR-TAR-QUIN CORACIN CORALSONE CORAMINE CORAVAL CORDRAN CORDRAN-N CORGARD CORGONJECT-5 CORICIDIN CORICIDIN "D" CORICIDIN COUGH SYRUP CORICIDIN MIST CORIFORTE CORILIN CORIMIST CORLUTIN L.A. 250 CORMONE COROVAS TYMCAP CORRECTIVE MIXTURE CORRECTIVE MIXTURE W/PAREGORIC CORRECTOL CORT-DOME CORT-QUIN CORTAID CORTEF CORTENEMA CORT ICAINE 07778 CORTICOTROPIN 07780 07785 07790 07795 07800 07805 07810 07815 07820 07825 07830 07835 07840 07845 07850 J7855 07860 07865 07870 07875 07880 07885 CORT IFOAM CORT IGEL CORT IGES IC DROPS CORT ISONE CORT ISPORIN CORTONE CORTRIL CORTROPHIN CORTROSYN CCRYBAN-D CORYBAN-D COUGH SYRUP CORY ZA CORYZA BRENGLE CORYZAID CORY ZT IME CORZONE SYRUP COSANYL COSFA COSMEGEN COTAZYM COTAZYM-B COTON SYRUP Ly 07890 COTROL-D 07895 COTROPIC GEL 0 07905 COTUSSIS SYRUP 07910 COTYLENOL 07915 OUGH SILENCER 07920 COUGH SYRUP 07925 COUGHI NE_SYRUP 07930 COUMADIN 07935 COUNTERPAIN RUB 07940 COVANAMINE 07945 COVANGESIC 07950 COVERLET 07955 COVIEONE 07960 CPA 07965 CPD HAEMO-PK W/IN-LINE PIL TUB NDL 07970 CPD HAEMO-PK W/STD DONOR TUBE 07975 CPD HAEMO-PK W/7ML IN-LINE PIL TUB 07980 CP2 07985 CREAM OF TARTAR 07990 CREAMALIN 07995 CREO-TERPIN COMPOUND 08000 CREOSOTE NF XII 08005 CRES CORMON 08010 CRESCL COMPOUND 08015 CROMOLYN 08020 CRUEX 08025 CRYSTI-12 GEL 08030 CRYSTICILLIN 08035 CRYSTIMIN 08040 CRYSTODIGIN 08045 CUPREX 08050 CUPRIC SULFATE 08055 CUPRIC SULF ANHYD REAG POWD 08060 CUPRIMINE 08065 CURRETAB 08070 CYAMINE 08075 CYANIDE ANTIDOTE PACKAGE 08080 CYANO 08085 CYANOCOB 08090 CYANOCOBALAMIN 08095 CYANOCOBALAMIN NEO-VADRIN 08100 CYANOCOBALAMIN REPOSITORY 08105 CYANOJECT 08110 CYANTIN 08113 CYCLACILLIN 08115 CYCLAINE 08120 CYCLANDELATE 08125 CYCLANFOR 08130 CYCLAPEN 08135 CYCLOCORT 08140 CYCLOGYL 08145 CYCLOMYDRIL 08150 CYCLOPAR 08153 CYCLOPHOSPHAMIDE 08155 CYLOPENTOLATE 08160 CYCLOSPASMOL 08165 CYDEL 08170 CYLANA SYRUP 08175 CVYLERT MEDICATION CODE LIST, NAMCS 1980 08180 CYPROHEPTADINE 08185 CYSTEINE 08190 CYSTEX 08195 CYSTO-CONRAY 08200 CYSTOGRAFIN 08205 CYSTOSPAZ 08210 CYSTOSPAZ-SR 08215 CYTAL 08220 CYTELLIN 08225 CYTOFERIN 08230 CYTOMEL 08235 CYTOMINE 08240 CYTOSAR 08245 CYTOXAN 08250 CYVASO 08255 CZO LOTION 08260 C3 08265 D CAINE 08270 D.H.E. 45 08275 De.P.X.lL. 08285 Do.S.S. W/CASANTHRAN 08290 D-FEDA 08295 D-S-S 08300 D-S-S PLUS 08305 D-SEB GEL 08310 D-SINUS 08315 D-10 08320 D-2.5 08325 D-20-W 08330 D-3.3-1/3S 08335 D-5 08340 D-50 08345 DACRIOSE 08350 DAILY MULTIPLE VITAMIN 08355 DAILY MULTIPLE VITAMIN W/IRON 08360 DAILY VITAMIN FLAVORED SYRUP 08365 DAINITE 08370 DALIBOUR 08375 DALIDOME 08380 DALKON 08385 DALLERGY 08390 DALMANE 08395 DALOLONE 08400 DANAZOL 08405 DANEX SHAMPOO 08410 DANOCRINE 08415 DANTHRON 08420 DANTRIUM 08425 DANTROLENE 08430 DAPA 08435 DAPASE 08440 DAPSONE 08445 DARANIDE 08450 DARAPRIM 08455 DARBID 08460 DAR ICON 08465 DARICON PB 08470 DARVOCET-N 08475 DARVON 08480 DARVON COMPOUND 08485 DARVON W/A.S.A. 8490 DARVON-N 849 DAR ~ A A 08500 DASIKON 08 DASIN 08510 DATRIL 08515 DAXOLIN 08520 DAYALETS 08525 DAYCARE 08530 DBH B-12 08535 DDAvVP 08540 DDS COMPOUND 08545 DE CAL 08550 DEANER 08555 DEAPRIL-ST 08560 DEBRISAN 0856 DEBROX 08570 DECA-DURABOL IN 08575 DECADERM 08580 DECADROL 08585 DECADRON 08590 DECADRON ELIXIR 859 DECADRON PH HA 600 NP A XY 08605 DECADRON-LA 08610 DECAJECT 08615 DECAPRYN 08620 DECAPRYN SYRUP 08625 DECASPRAY 08630 DECAVITAMIN 08635 DECHOLIN 08640 DECLOMYCIN 08645 DECLOSTATIN 08650 DECOBEL LANACAP 08655 DECOHIST SYRUP 08660 DECONAMINE 08665 DECONEX 08670 DECONGESTANT 08675 DECONGESTANT AT 08680 DECONGESTANT ELIXIR 08685 DECONGESTANT EXPECTORANT 08690 DECONGESTCAPS 08695 DECONGEX 08700 DEGEST 2 08705 DEHIST 08710 DEHYDROCHOLATE 08715 DEHYDROCHOLIC ACID 08720 DELADUMONE 08725 DELALUTIN 08730 DELATESTRYL 08735 DELAXIN 08740 DELCID 08745 DELESTROGEN 08750 DELFEN 08755 DELLADEC 08760 DELTA-CORTEF 08765 DELTALIN 08770 DELTASONE 08775 DEMAZIN cv 08780 08785 08790 08795 08800 08805 08810 08815 08820 08825 08830 08835 08840 08845 08850 08855 08860 08865 08870 08875 08880 DEMEROL DEMEROL SYRUP DEMEROL-APAP DEMI-REGROTON DEMS ER DEMULEN DEMULEN-28 DENCORUB DENDRID DENTA-FL DENTAVITE DEPAKENE DEPEN DEPI NAR DEPLETITE-25 DEPO PRED DEPO-ESTRADIOL DEPO-MEDROL DEPG-PROVERA DEPO-TESTADIOL DEPO-TESTOST ERONE 08885 08890 08895 08900 08905 08910 08915 08920 08925 08930 08935 08940 08945 08950 08955 08960 08965 08970 08975 08980 08985 08990 08995 09000 09005 09010 09015 09020 09025 09030 09035 09040 09045 09050 09055 09060 09065 09070 09075 DEPOGEN DEPOPRED DEPOTEST DEPOTESTOGEN DE PROL DERFULE DERIFIL DERMA MEDICONE DERMA MEDICONE-HC DERMA PACK DERMA PH LOT ION DERMA SOAP DERMA-COVER DERMACOAT SPRAY DERMACORT DERMAL RUB DERM AR EX DERMASSAGE DERMAVAL DERMOGEN DERMOLATE DERMOLATUM DERMOL IN DERMOPLAST DERMOV AN DESENEX DESFERAL DESIPRAMINE DESITIN DESO-CREME DESOXYCORTICOSTERONE DESOXYN DE SQUAM-X DESQUAM-X WASH DETOXIL DETUSSIN DEX-SALT DEXACEN DEXAMETHASONE MEDICATION CODE LIST, NAMCS 1980 09080 09085 09090 09095 09100 09105 09110 09115 09120 09125 09130 09135 09145 09150 09155 09160 09165 09170 09175 09180 09185 09190 09195 09200 09205 09210 09215 09220 09225 09230 09235 09240 09245 09250 09255 09260 09265 09270 09275 09280 09285 09290 09295 09300 09305 09310 09315 09320 09325 09330 09335 09340 09345 09350 09355 09360 09365 09370 09375 09380 DEXAMETHASONE ACETATE DEXAMETHASONE ELIXIR DEXAMETHASONE SODIUM DEXAMETHASONE SODIUM PHOSPHATE DE XAMPE X DEXAMYL DEXAPED DEXA SONE DE XEDRINE DEXON DEXONE DEXPANTHENOL DE XSONE DEXSONE II DEXSONE L.A. DEXTRAN DE XTRO-TUSSIN SYRUP DEXTROAMPHE TAMI NE DEXTROSE DEXTROSE W/ELECTROLYTE DEXTROSTIX DEZONE DI PHEN DI-CAL MEAD DI-CALCIUM PHOSPHATE W/D DI-EST (MODIFIED) DI-FERRIN DI-GEL DI-GENIK DI-ISOPACIN DI-PRED DI-SPAZ DIA-QUEL DIABINE SE DIABISMUL DIACIN DIACTION DIAFEN DIAHIST ELIXIR DIALIXIR DIALOG DIALOSE DIAL UME DIAMOND ANTISEPTICS DIAMOX DIAMULSIN IMPROVED DIANABOL DIANEAL DIAPA KARE DIAPARENE DIAPID DIAQUA DIASAL GRANULE DIASONE SODIUM ENTERAB DIASPORAL DIASTIX DIATRIZOATE DIAZEPAM DIBENT DIBENT PB 09385 DIBENZYLINE 09390 DIBUCAINE 09395 DICAL-D 09400 DICAL-D W/ IRON 09405 DICALCIUM 09410 DICALCIUM PHOSPHATE 09415 DICALCIUM PHOS W/VIT D 09420 DICALCIUM PHOS W/VIT D & IRON 09425 DICARBOSIL 09430 DICEN 09433 DICLOXACILLIN 09435 DICODID 09440 DICORT 09445 DICUMAROL 09450 DICURIN 09455 DICYCLOMINE 09460 DICYCLOMINE wW/PB 09465 DIDREX 09470 DIDRONEL 09475 DIENESTROL 09480 DIETAC 09485 DIETAC DROPS 09490 DIETHYL-PROPION 09495 DIETHYLPROPION 09500 DIETHYLSTILBESTROL 09505 DIGESTA-LAX 09510 DIGESTALIN 09515 DIGESTANT 09520 DIGESTOZYME 09525 DIGIFORTIS 09530 DIGIGLUSIN 09535 DIGITALIS 09540 DIGITOXIN 09545 DIGOXIN 09550 DIHISTINE 09555 DIHISTINE ELIXIR 09560 DIHISTINE EXPECTORANT 09565 DIHORMOGEN 09570 DIHYCON 09575 DIHYDROTACHYSTEROL 09580 DIIODOHYDROXYQUIN 09585 DILANTIN 09590 DILANTIN W/PHENOBARBITAL 09595 DILAUDID COUGH SYRUP 09600 DILAUDID 09605 DILAX 09610 DILIN 09615 DILOCAINE 09620 DILOCOL 09625 DILOR 09630 DILOR-G 09635 DIMACOL 09640 DIMENHYDRINATE 09645 DIMENHYDRINATE 22 GA 09650 DIMENTABS 09655 DIMERAY 09660 DIMETANE 09665 DIMETANE ELIXIR 09670 DIMETANE EXPECTORANT 09675 DIMETANE EXPECTORANT-DC ev 09680 DIMETANE EXTENTAB 09685 DIMETANE-TEN 09690 DIME TAPP 09695 DIMINDOL 09700 DIO SUL 09705 Dp1ocTALOSE 09710 DIOCTO 09715 DIOCTO PLUS 09720 DIOCTO SYRUP 09725 DIOCTO-C SYRUP 09730 DIOCTYL SOD SULFOSUCC W/CASANTH 09735 DIGCTYL SODIUM SULFOSUCCINATE 09740 DIOCTYL W/CASANTHRANOL 09745 DIOCTYL W/DANTHRON 09750 DIOGEZE 09755 DIOLAX 09760 DIOMEDICONE 09765 DIONEX 09770 DIONIN 09775 DIOSTATE D 09780 DIOGSUCCIN 09785 DIOSUCCIN C SYRUP 09790 DIOSUCCIN SYRUP 09795 _ DIGTHANE 09800 DIOTHRON 09805 DIOVAL 09810 DIPAV 09815 DIPH TET TOXIODS Pi R 09820 DPT 09825 DIPHEN-EX 09830 DIPHENALLIN COUGH SYRUP 09835 DIPHENATOL 09840 DIPHENHYDRAMINE COMP EXP 09845 DIPHENHYDRAMINE COUGH SYRUP 09850 DIPHENHYDRAMINE 09855 DIPHENHYDRAMINE HCL COUGH SYR 09860 DIPHENHYDRAMINE HCL ELIXIR 09865 DIPHENHYDRAMINE HCL EXP 09870 DIPHENHYDRAMINE HCL SYRUP 09875 DIPHENHYDRAMINE HCL 22 GA 988 DIPH ATE ATROP_SULF 09885 DIPHENYLAN SODIUM 09895 DIPHTHERIA ANTITOXIN 09900 DIPHTHERIA TETANUS TOXOIDS 09905 DIPHTHERIA TOXOID ADULT 09910 DIPHYLETS 09915 DIPROSONE 09920 DIPYRIDAMOLE 09925 DISALCID 09930 DISANTHROL 09935 DI SI PAL 09940 DISMISS DISPOSABLE 09945 DISOLAN 09950 DISONATE 09955 DISOPHROL 09960 DI SOPLEX 09965 DISOTATE 09970 DISPATABS MEDICATION CODE LIST, NAMCS 1980 09975 DISULFIRAM 09980 DITAN 09985 DITATE 09990 DIPH TET TOXOIDS PERTUSSIS 09995 DITROPAN 10000 DIUCARDIN 10005 DIULD 10010 DIUMEAD 10015 DIUPRES 10020 DIURETIC 10025 DIURIL 10030 DIUTENSEN 10035 DIZMISS 10040 DM-PLUS COUGH SYRUP 10045 DOAK OIL 10050 DOBELL'S SOLUTION 10055 DOBUTREX 10060 DOCA ACETATE 10065 DOCAPHEN 10070 DOCTATE 10075 DOCTATE-P 10080 DOCTIENT 10085 DOCTIENT W/HYDROC ORTISONE 10090 DODEX 10095 DOKTORS NASAL MIST 10100 DOKTORS NOSE DROPS 10105 DOLACET 10110 DOLANEX ELIXIR 10115 DOLENE 10120 DOLENE COMPOUND-65 10125 DOLICAINE 10130 DOLOPHINE 10135 DOLOR PLUS 10140 DOME-PASTE BANDAGE 10145 DOMEBORO 10150 DOMEBORO OTIC 10155 DOMEFORM-HC 0.5% 10160 DOMERINE MEDICATED SHAMPOO 10165 DOMMANATE 10170 DOMOL 10175 DONABARB 10180 DONATUSSIN 10185 DONATUSSIN SYRUP 10190 DONNA-PHENAL ELIXIR 10195 DONNA-SED ELIXIR 10200 DONNAGEL 10205 DONNAGEL-PG 10210 DONNATAL 10215 DONNAZYME 10220 DONPHEN 10225 DOPAMINE 10230 DOPAR 10235 DOPRAM 10240 DORBANE 10245 DORBANTYL 10250 DORCOL PEDIATRIC COUGH SYRUP 10255 DORIDEN 10260 DORME 10265 DORMTABS 10270 DORSACA INE 10275 10280 10285 10290 10295 10300 10305 10310 10315 10320 10325 10330 10335 10340 10345 10350 10355 10360 10365 10370 10375 10380 10385 10390 10395 10400 10405 10410 10415 10420 10425 10430 10435 10440 10445 10450 10455 10460 10465 10470 10475 10480 10485 10490 10495 10500 10505 10510 10515 10520 10525 10530 DOSS 300 DOUBLE K DOVACET DOVER'S POWDER DOVERAM PLUS DOVERIN DOVERLYN DOVOSAL DOW- ISONIAZID DOXAN DOXEPIN DOXIDAN DOXINATE DOXY C DOXY 6 DOXYCHEL DOXYCYCLINE DOXYLAMINE SUCC & PYRIDOXINE DRAKE'S (DR) COUGH SYRUP DRAL SERP DRALZINE DRAMAMINE DRAMAMINE SUPPOSICONE DRAMIL IN DRAMOCEN DRAMOJECT DRAWING PASTE DREST GEL DRI-PHED DRI-PHED EXPECTORANT W/CODE INE DRI-PHED SYRUP DRINOPHEN DRINUS DRISDOL DR ISTAMEAD DRISTAMEAD LONG ACTING MIST DRISTAN DRISTAN COUGH FORMULA DRISTAN INHALER DRISTAN NASAL MIST DRISTAN ROOM VAPORIZER DRISTAN VAPOR SPRAY DRIXORAL DRIZE DROLBAN DROPER IDOL DROT IC DROXOMIN DROXOVITE DRYVAX DSS DSS W/CASANTHRANOL 10535 DTIC 10540 10545 10550 10555 10560 10565 10570 DT IC DU-MIN DUA-PRED DUADAC IN DUAL WET DUAL EX DUBBALAX DULARIN SYRUP 10575 10580 10585 10590 10595 10600 10605 10610 10615 10620 10625 10630 10635 10640 10645 10650 10655 10660 10665 10670 10675 10680 10685 10690 10695 10700 10705 0710 DULC OL AX DULCOLAX BOWEL PREP KIT DUD CITRUS BIOFLAVONOIDS DUO CYP DUO HIST DUO PRED DUC-C. VePe DUC-DE ZONE DUO-K DUC-MEDIHALER DUOB ARBI TAL DUCFILM DUOGEN DUOHAL ER DUOGJECT DUOLUBE DUOSCORB DUOSOL DUOTRATE DUOVAL-P.A. DUOVENT DUPHALAC SYRUP DUPHASTON DUPHRENE DURA ESTRIN DURABOLIN DURACILLIN DURADYNE 10715 DURADYNE D.H.C. 10720 10725 10730 10735 10740 10745 10750 10755 DURAGEN DURAGESIC DURALONE DURALUGEN DURALUTIN DURANEST DURANEST HCL W/EPINEPHRINE DURAQUIN 10760 DURATEARS 10765 10770 10775 10780 10785 10790 10795 10800 080! 10810 10815 10820 10825 10830 10835 10840 10845 10850 10855 10860 10865 10870 DURATEST DURATESTRIN DURATET 500 DURATHESIA DURATION DROPS DURATION SPRAY DURE ZE DURI CEF UVOI ov DYAZ IDE DYCILL DYCLONE DYLL INE DYMELOR DYMENAT Be DYNAPEN DYNOSAL DYPAP ELIXIR DYPHYLLINE DYRENIUM DYTEST MEDICATION CODE LIST, NAMCS 1980 10875 10880 10885 10890 10895 10900 10905 10910 10915 E.E. Se E.T.H. COMPOUND E-BIOTIC E-CARP INE E-FEROL E-IONATE P.A. E-MYCIN E-PILO E-R-0 10920 E-VITES 10925 10930 10935 10940 10945 10950 10955 10960 10965 10970 10975 10980 10985 10990 10995 11000 11005 11010 11015 11020 11025 11030 11035 11040 11045 11050 11055 11060 EAR DROPS EAR DROPS FORMULA #2 EAR SOL #5 W/HYDROCORTI SONE EAR-DRY EAROCOL ECEE ECHODIDE ECLIPSE ECONOCHLOR ECONOPRED ECOTRIN ECTASULE MINUS EDECRIN EFED EFEDRON EFFACOL EFFERSYLLIUM EFODINE EFR ICEL EFRICON EXPECTORANT EFUDEX EKKO EKRISED EL-DA-MINT ELA SE ELASE-CHLOROMYCETIN ELASTOMULL ELASTOPATCH 11065 ELAVIL 11070 11075 11080 11085 11090 1095 ELDADRYL COUGH SYRUP EL DEC ELDECORT ELDEFED ELDER COMPOUND 65 ELDERCAPS 11100 ELDERTONIC ELIXIR 1105 ELDOCORT 2.5% 11110 11115 11120 11125 11130 11135 11140 11145 ELDODRAM ELDOPAQUE ELDOQUIN ELECTRODE PASTE ELECTROLYTE ELI XICON EL IXOPHYLLIN EL IXOPHYLLIN S.R. 11150 11155 11160 11165 11170 EL I XOPHYLLIN-KI ELSPAR ELZ YME EMAGR IN EMAGRIN IMPROVED 11175 EMBRON 11180 EMETE-CON 11183 EMETIC AGENT 11185 EMET INE 11190 EMETROL 11195 EMFABID 11200 EMFASEEM 11205 EMKO FOAM 11210 EMPIRIN 11215 EMPIRIN COMPOUND 11220 EMPIRIN COMPOUND #1 11225 EMPIRIN COMPOUND #2 11230 EMPIRIN COMPOUND #3 11235 EMPIRIN COMPOUND #4 11240 EMPIRIN COMPOUND W/CODEINE 11245 EMPIRIN NO. #2 11250 EMPIRIN NO. #3 11255 EMPIRIN NO. #4 11260 EMPIRIN W/ CODE INE 11265 EMPRACET 11270 EMPRAZIL 11275 EMPRAZIL-C 11280 EMUL-0-BALM 11285 EMULAVE 11290 EMULSOIL 11295 EN-CEBRIN 11300 EN-CEBRIN F 11305 ENARAX 11310 ENCARE OVAL 11315 END-A-KOFF 11320 ENDECON 11325 ENDEP _ 11330 ENDOTUSSIN-NN 11335 ENDRATE 11340 ENDURON 11345 ENDURONYL 11350 ENDURONYL FORTE 11355 ENFAMIL 11360 ENFAMIL W/ IRON 11365 ENGRAN 11370 ENISYL 11375 ENO 11380 ENOVID 11385 ENOXA 11390 ENSURE 11395 ENTEX 11400 ENTOZYME 11405 ENTUSS 11410 ENUCLENE 11615 ENURETROL 11420 ENZACTIN 11425 ENZOBILE 11430 ENZYMATIC DIGESTANT 11433 ENZYME 11435 ENZYMET 11440 ENZYPAN 11445 EPHED-ORGANIDIN ELIXIR 11450 EPHEDRINE 11455 EPHEDRINE & AMYTAL 11460 EPHEDRINE & PHENOBARBITAL Sy 11465 EPHEDRINE & SECONAL SODIUM 11470 EPHEDRINE & SODIUM PB 11475 EPHEDRINE AND NEMBUTAL-25 11480 EPHEDRINE SULFATE 11485 EPHEDRINE SULF & AMOBARB 11490 EPHEDRINE SULFATE SYRUP 11495 EPHEDROL W/ CODEINE 11500 EPHEDSOL 11505 EPHENYLLIN 11510 EPI-CLEAR 11515 EPICORT LOTION 11520 EPIFOAM 11525 EPIFRIN 11530 EPINAL 11535 EPINEPHRICAINE 11540 EPINEPHRINE 11545 EPINEPHRINE MUROCOLL 11550 EPITRATE 11555 EPPY 11560 EPRAGEN 11565 EPROLIN GELSEAL 11570 EPSILAN-M 11575 EPSOM SALT 11580 EQUAGESIC 11585 EQUANIL 11590 EQUANITRATE 11595 EQUILET 11600 ERCA 11605 ERGOBEL 11610 ERGOCAFFEIN 11615 ERGOCALCIFEROL 11620 ERGOMAR 11625 ERGONOVINE 11630 ERGOPHENE 11635 ERGOSTAT 11640 ERGOTAMINE 11645 ERGOTRATE 11650 ERO FORTE 11655 ERYPAR 11660 ERYTHROCIN 11665 ERYTHROMYCIN 11670 ESCOT 11675 ESEMGESIC 11680 ESERINE 11685 ESERINE ALKALOID 11690 ESIDRIX 11695 ESIMIL 11700 ESKALITH 11705 ESKATROL 11710 ESOPHOTRAST CREAM 11715 ESTAR 11720 ESTATE 11725 ESTI NYL 11730 ESTIVIN 11735 ESTOMUL-M 11740 ESTRACE 11745 ESTRADIOL 11750 ESTRADURIN 11755 ESTRAGUARD 11760 ESTRATAB MEDICATION CODE LIST, NAMCS 1980 11765 ESTRATEST 11770 ESTRAVAL 11775 ESTRO CYP 11780 ESTROATE 11785 ESTROCON 11790 ESTROFEM 11795 ESTROFOL 11800 ESTROGEN 11805 ESTROJECT 11810 ESTROJECT-L.A. 11815 ESTRONE 11820 ESTRONE & POT ESTRONE SULFATE 11825 ESTRONE AQUEOUS 11830 ESTRONE SUSPENSION 11835 ESTRONE 5 11840 ESTRONOL AQUEOUS 11845 ESTROVIS 11850 ETHAMBUTOL 11855 ETHAMIDE 11860 ETHAQUIN 11865 ETHATAB 11870 ETHAVERINE HCL 11875 ETHER 11880 ETHINYL ESTRADIOL 11890 ETHIODOL 11895 ETHOBRAL 11898 ETHOSUXAMIDE 11900 ETHRANE 11905 ETHRIL 11910 ETHYL ACETATE 11915 ETHYL CHLORIDE 11920 _ETRAFON 11925 EUCALYPTOL 11930 EUCALYPTUS OIL NF 11935 EUCAPHEN 11940 EUCAPINE SYRUP 11945 EUCER IN 11950 EUGEL 11955 EUGENOL USP 11960 EWAX 11965 EUTHROID 11970 EUTONYL 11975 EUTRON 11978 EXPEC TORANT 11980 EVAC-Q-KIT 11985 EVAC-Q-KWIK 11990 EVAC-U-LAX 11995 EVAGEN 12000 EVANS BLUE 12005 EVERONE 12010 EVEX 12015 EX APAP 12020 EX-CALORIC WAFER 12025 EX-LAX 12030 EX-OBESE 12035 EXCEDRIN 12040 EXNA NA- 12050 EXSEL LOTION 12055 EXTENDRYL 12060 EXTRAL IN 12065 EXTRALIN B 12070 EXTRALIN F 12075 EXZIT 12078 EYE PREPARATION 12080 EYE-SED 12085 EYE-STREAM 12090 EZE-LAX 12095 El 12100 E2 12105 F & B CAPS 12110 F.C.A.H. 12115 F-B~C 12120 F-E-P 12125 FACTORATE 12130 FAMILY VITAMIN 12135 FAST GREEN FCF 12140 FASTIN 12145 FEBRIDOL 12150 FEBRINOL 12155 FEBRO-BAR 12160 FEDAHIST 12165 FEDAHIST EXPECTORANT 12170 FEDAHIST SYRUP 12175 FEDRAZIL 12180 FEDRELEN 12185 FEDRINAL 12190 FEEN-A-MINT 12195 FELLOWS COMPOUND SYRUP 12200 FEM-H 12205 FEMGUARD 12210 FEMICIN 12215 FEMININS 12220 FEMINONE 12225 FEMIRON 12230 FEMOGEN 12235 FENDOL 12240 FENDON 12245 FENOPROFEN 12250 FENYLHIST 12255 FEOSOL 12260 FEOSOL ELIXIR 12265 FEQOSOL PLUS 12270 FEOSOL SPANSULE 12275 FEQSTAT 12280 FEOSTIM 12285 FER-IN-SOL 12290 _FERANCEE 12295 FERATE C 12300 FERGON 12305 FERGON ELIXIR 12310 FERGON PLUS 12315 FERINATE 12320 FERMALOX 12325 FERNDEX 12330 FERNHIST 12335 FERNISOLONE 12340 FERO-FOLIC-500 2345 FERO-GRAD-50 12350 FERO-GRADUMET or 12355 FEROSORB C 12360 FEROTRAN 12365 FERRALET 12370 FERRALET PLUS 12375 FERRALYN LANACAP 12380 FERRIC AMMONIUM 12385 FERRIC AMMONIUM SULFATE 12390 FERRIC CHLORIDE 12395 FERRIC SUBSULFATE 12400 FERRITRINSIC 12405 FERRO PLUS 2410 FERRO 12415 FERRO-SEQUEL 2420 FERROBID 12425 FERROCTYL 12430 FERROLIP 12435 FERROLIP PLUS 12440 FERROLIP 50 MG SYRUP 12445 FERRONESE 12450 FERRONEX 12455 FERRONORD 12460 FERROSPAN 12465 FERROUS FUMARATE 12470 FERROUS FUMARATE W/DSS 12475 FERROUS G EL IXIR 12480 FERROUS GLUCONATE 12485 FERROUS PLUS 12490 FE S04 12490 FERROUS SULATE 12495 FERROUS SULF W/BREWER'S YEAST 12500 FERROUS SULFATE W/ THIAMINE 12505 FERYL 12510 FESTAL 12515 FESTALAN 12520 FEVER REDUCER ST JOSEPH 12525 FILIBON 12530 FILIBON F.A. 535 FILIBON FORTE 12540 FILIBON OT 12545 FIOGESIC 12550 FIORINAL 12555 FIORINAL NO. 1 12560 FIORINAL NO. 2 12565 FIORINAL NO. 3 12570 FIORINAL W/CODEINE 12575 FIRST AID SPRAY 12580 FIZRIN 12585 FLAGYL 12590 FLAVITAB 12595 FLAXEDIL 12600 FLEET BAGENEMA 12605 FLEET BARIUM ENEMA 12610 FLEET BAROBAG 12615 FLEET BISACODYL 12620 FLEET ENEMA 12625 FLEET THEOPHYLLINE 12630 FLEX CARE 12635 FLEXERIL 12640 FLEXICAL 12645 FLEXOJECT MEDICATION CODE LIST, NAMCS 1980 12650 12655 12660 12665 12670 12675 12680 12685 12690 12695 12700 12705 12710 12715 12720 12725 12730 12735 12740 12745 12750 12755 12760 FLEXON FLEXSOL FLO TAB FLOR INEF FLORONE FLOROPRYL FLOZENGES FLUAX TYPE A,B FLUOC TNOLONE FLUOGEN FLUOGEN-B MONOVALENT FLUONID FLUOR-I-STRIP FLUORESCEIN FLUORESC ITE FLUORESEPTIC FLUORI-METHANE SPRAY FLUORIDE FLUOR IDENT FLUORIGARD FL UOR INSE FLUORITAB FLUORODEX 12763 FLUOROMETHOLONE 12765 FLUOROPLEX 12770 FLUOROURACIL 12770 12775 12780 12785 12790 12795 12800 12805 12810 12815 12820 12825 12830 12835 12840 12845 12850 12855 12860 12865 12870 12875 12880 12885 12890 12895 12900 12905 12910 12915 12920 5 FU FLUOTHANE FLUPHENAZINE FLURA-DROPS FLUWRA-LOZ FLURA-PREN FLURA-TABLETS FLURA-VITE FLURAZEPAM FLURESS FLURO-ETHYL SPRAY FLUROSYN FLUZONE FLUZONE TRIVALENT TYPE A,B FLUZONE-CONNAUGHT TYPE A,B FML LIQUIFILM FO-MATE FOILLE FOLBESYN FOLIC ACID FOLLESTROL FOLLUTE IN FOL VITE FOL VRON FOMAC FOR-DYNE FORBAXIN FORCOLD SYRUP FORHISTAL FORMALDEHYDE FORMALIN 12925 FORMATRI 12930 12935 FORMIC ACID FORMULA #81 12940 12945 12950 12955 12960 FORMULA 2 FORMULA 44 COUGH MIXTURE FORMULA 44 DISC FORMULA 44-D COUGH MIXTURE FORMULA-T 12965 FORTAPLEX 12970 12975 12980 12985 12990 FORTESPAN FOSFREE FOSTEX FOSTEX BPO GEL FOSTEX MEDICATED CLEANSER 12995 FOSTRIL 13000 13005 13010 13015 13020 13025 13030 13035 13040 13045 13050 13055 13060 13065 13070 13075 13080 13083 13085 13090 13095 13100 13105 13110 13115 13118 FOWLER'S SOLUTION FREAMINE FREPP FRIGIDERM SPRAY FRUCTOSE FRUCTOSE 10-W FUDR FUL-GLO FULLER'S EARTH POWDER FULVICIN FUMARAL SPANCAP FUMASORB FUMATINIC FUMATRIN FORTE FUMERIN FUNDUSCEIN FUNGACETIN FUNGICIDE FUNG IZ ONE FUNGIZONE INTRAVENOUS FUNGIZONE LOTION FURACIN FURADANT IN FURALAN FURATOIN FUROS EM IDE 3120 FUROXON 13125 13130 13135 13140 13145 13150 13153 13155 13160 13165 13170 FYNEX G.B. PREP EMULSION GeB.Se. G-RECILLIN G-TUSSIN DM SYRUP G-TUSSIN SYRUP GALLBLADDER MEDICATION GAMAST AN GAMMA BENZENE HEXACHLORIDE GAMMAGEE GAMMAR 13175 GAMOPHEN 13180 13185 13190 13195 13200 13205 13208 13209 13210 GAMULIN RH GANATREX ELIXIR GANPHEN GANT ANCL GANTRISIN GARAMYCIN GARGLE GASTRIC AGENT GASTROENTERASE Ly 13215 GASTROGRAFIN 13220 GAVISCON 13225 GAYSAL 13230 GAYSAL S 13235 GBH 13240 GEE-GEE 13245 GEL-KAM 13250 GELATIN 13255 GELAZINE 13260 GELCIMISAL 13265 GELCOID 13270 GELFILM 13275 GELFOAM 13280 GELOCAST 13285 GELOX 13290 GELUMINA 13295 GELUSIL 13300 GEMONIL 13305 GENAPAX 13310 GENOPTIC 13315 GENOSCOPOLAMINE PELLET 13320 GENTAMICIN 13325 GENTIAN VIOLET 13330 GENTLAX 13335 GENTLE SPRING DISPOSABLE 13340 GENTRAN 13345 GENTZ WIPE 13350 GEOCILLIN 13355 GEOPEN 13360 GER O FOAM 13365 GER-I-BON 13370 GERATIC FORTE 13375 GERAVITE 13380 GERI-IRON NEO-VADRIN 13385 GERIAMIC 13390 GERIATRIC ELIXIR 13395 GERIATRIC FORMULA 13400 GERIATRIC VITAMINS 13405 GERIBOM 13410 GERIFORT 13415 GERIGINE 13420 GERILETS 13425 GERILIQUID 13430 GERIMIN 13435 GERINATS-T 13440 GERIPLEX 13445 GERITAG 13450 GERITINIC 13455 GERITOL 13460 GERITONIC 13465 GERIX ELIXIR 13470 GERIZYME 13475 GERMICIDAL SOL BARNES-HIND 13480 GEST 13485 GESTEROL 13490 GEvVIZOL 13495 GEVRABON 13500 GEVRAL 13505 GEVRITE 13510 66 MEDICATION CODE LIST, NAMCS 1980 13515 GINSENG 13520 GINSOPAN 13525 GITALIGIN 13530 GLAUCON 13535 GLUCAGON 13540 GLUCOLA 13545 GLUCORON WITH B-12 13550 GLUCOSE 13555 GLUCOVITE 13560 GLUKOR 13565 GLUTAMIC ACID 13570 GLUTAVITE 13575 GLUTETHIMIDE 13580 GLUTOFAC 13585 GLUTOL 13590 GLUTOSE 13595 GL Y-OXIDE 13600 GLYATE SYRUP 13605 GLYCATE 13610 GLYCERIN (FLAVORED) 13615 GLYCEROL 13620 GLYCEROLIZING SOL HAEMO-VAC 13625 GLYCERYL GUAIACOLATE AC SYRUP 13630 GLYCERYL GUAIACOLATE SYRUP 13635 GLYCERYL-T 13640 GLYCINE 13645 GL YCOBARB 13650 GLYCOGEL 13655 GLYCOPYRROLATE 13660 GLYCOPYRROLATE NO. 1 13665 GLYCOPYRROLATE NO. 2 13670 GLYCOPYRROLATE W/ PHENOBARBITAL 13675 GLYCOTUSS 13680 GLYCOTUSS DM 13685 GLYCOTUSS DM SYRUP 13690 GLYCOTUSS SYRUP 13695 GLYCOTUSSIN 13700 GLYCYRRHIZA FLUIDEXTRACT USP 13705 GLYDEINE SYRUP 13710 GLYDM SYRUP 13715 GLYESTRIN 13720 GLYNAZAN 13725 GLYNAZAN ELIXIR 13730 GLYNAZAN EXPECTORANT 13735 GLYROL 13740 GLYSENNID 13745 GLYTUSS 13750 GOLD SODIUM THIOSULFATE 13755 GONIC 13760 GONIO GEL 13765 GONIOSOL 13770 GOURMASE 13775 GRANULEX 13780 GREEN SOAP 13785 GRIFULVIN 13790 GRIS-PEG 13795 GRISACTIN 00 13805 GRISOWEN 13810 GUAIACOL 13815 13820 13825 13830 13835 13840 13845 13850 13855 13860 13865 13870 13875 13880 13885 13890 13895 13900 13905 13910 12915 13920 13925 13930 GUAIADEX SYRUP GUATAMEN GUAIATRATE SYRUP GUAIFENESIN & DM GUAI FENES IN GUAIPHENYL GUANETHIDINE GUANIDINE GUARSOL GUIALYN W/CODEINE COUGH SYRUP GUIATEX GUIATUSCON GUIATUSS GUIATUSS A.C. SYRUP GUIATUSSIN GUIATUSSIN W/ CODEINE GUIATUSS IN W/DEXTROMETHORPHAN GU ISTREY FCRTIS GUSTALAC GUSTASE GUSTASE PLUS GYLANPHEN GYN GYNE-LOTRIMIN 13935 GYNERGEN 13940 13945 13950 13955 13960 13965 13970 13975 13980 13985 13990 13995 14000 14005 14010 14015 14020 14025 14030 14035 14040 14050 14055 14060 14065 14070 14075 GY NET ONE GYNOREST HE&I HeHeRe HeQeC. KIT HeRoHe HeV.B. H-BIG H-H H-R LUBRICATING JELLY HABDRYL HALAZONE HALDOL HALDRONE HALERCOL HALEY'S MO HALIBUT LIVER OIL HALLS COUGH SYRUP HALODRIN HALOG HALCPERIDOL 14045 HALOTESTIN HALOTEX HALOTHANE HAND'S (DR.) TEETHING LOTION HARMONYL HASP HBP 14080 HC 14085 14090 14095 14100 14105 14110 HCV HEAD & SHOULDERS HEB CREAM BASE HEB-CORT HEDULIN HEMA-COMBISTIX 8 14115 14120 14125 14130 14135 14140 14145 14150 14155 14160 14165 14170 14175 14180 14185 14190 14195 14200 14203 14205 14210 14215 14220 14225 14230 14235 14240 14245 14250 14255 14260 14265 14270 14275 14280 14285 14290 14295 14300 14305 14310 14315 14320 14325 14330 14335 HEMAFERRIN HEMALGEN HEMAST IX HEMATEST HEMATINIC HEMATINIC (STUART) HEMATINIC #1 HEMATINICAPS HEMATOVALS HEMATRAN HEMOCAINE HEMOFI L HEMOR-RID HEMORRHOIDAL HEMORRHOIDAL CPI HEMORRHOIDAL HC HEMORRHOIDAL SUPPOSITORY HEMORRHOIDAL-HC HEMOSTATIC AGENT HENOMINT ELIXIR HENOTAL HEP-B GAMMAGEE HEP-IRON ELIXIR HEP-LOCK HEPAFOLIC 12 FORTE HEPAHYDRIN HEPARIN HEPARIN LOCK FLUSH SOLUTION HEPATIC-AID HEPFOMIN 500 HEPICEBRIN HEPP-IRON DROPS HEPRINAR HEPRON HEPTUNA PLUS HERBAL NORFORMS HERPLEX LIQUIFILM HESPER HESPERIDIN C HETRAZAN HEXA-BETALIN HEXABOIL DRAWING SALVE HEXACHLOROPHENE HE XACREST HEXADERM HEXADERM IQ 14340 HEXADROL 14345 14350 14355 14360 14365 14370 14375 14380 14385 14390 14395 14400 14405 HEXADROL ELIXIR HEXADROL PHOSPHATE HEXALET HEXALOL HEXAVIBEX HEXAVITAMIN HEXE STROL HI B—-COMPLEX W/C HI - BEE W/C HI BEECO HI POTENCY VIT B-COMP W/VIT C HI -POTENCY VIT-MIN-IRON HI-TEMP MEDICATION CODE LIST, NAMCS 1980 14410 HI-TEN 14415 HIBICLENS 14420 HIBITANE 14425 HI POTENCY B-VITAMINS W/C 14430 HIGH POTENCY DROPS 14435 HIGH POTENCY VITAMIN NEO-VADRIN 14440 HIPREX 14445 HISCATABS 14450 HISERPIA 14455 HISPRIL SPANSULE 14460 HISTA COMPOUND NO. 5 14465 HISTA-CLOPANE 14470 HISTA-DERFULE 14475 HISTA-VADRIN 14480 HISTABID 14485 HISTADYL E.C. 14490 HISTAGESIC 14495 HISTAJECT 14500 HISTALET 450 ST FOR 14510 HISTALET SYRUP 14515 HISTALET X 14520 HISTALET X SYRUP 14525 HISTALET-DM SYRUP 14530 HISTALOG 14535 HISTAMINE 14540 HISTAMINE PHOSPHATE 14545 HISTASPAN 14550 HISTASPAN-D 14555 HI STASPAN-PLUS 14560 HISTATAB PLUS 14565 HISTATAPP 14570 HISTEFRIN 14575 HISTERONE 14580 HISTEX 14585 HISTIDINE 14590 HISTODRIX 14595 HISTOPLASMIN DILUTED SOLUTION 14600 HISTOREST 14605 HISTREY 14610 HISTREY KERNEL-KAP 14615 HISTREY SYRUP 14620 HIWOLFIA 14625 HMS 14630 HOLD 14635 HOMALYN 14640 HOMATROCEL 14645 HOMATROPINE 14650 HOMATROPINE MUROCOLL 14655 HOMICEBRIN 14660 HOMO-TET 14665 HONEY 14670 HORMOGEN DEPOT 14675 HORMOGEN R-A 14680 HORMOGEN-A 14683 HORMONE 14685 HORMONIN 14690 HOSPITAL LOTION 14695 HOSPITAL RUB 14700 HOUSE DUST CONCENTTD BULK TTMT 14705 HU-TET 14710 HUMAFAC 14715 HUMATIN 14720 HUMIST SPRAY 14725 HUMORSOL 14730 HURRICAINE 14735 HY-E-PLEX 14740 HY-GESTRADOL 14745 HY-GESTRONE 14750 HYALURONIDASE 14755 HYBALAMIN 14760 HYBEPHEN 14765 HYBOLIN 14770 HYCODAN 14775 HYCOFF 14780 HYCOFF SYRUP 14785 HYCOFF-A SYRUP 14790 HYCOFF 14795 HYCOMINE 14800 HYCORACE 14805 HYCOTUSS 14810 HYDELTRA-T.B.A. 14815 HYDELTRASOL 14820 HYDERGINE 14825 HYDEXTRAN 14830 HYDRA MAG 14835 HYDRA SPRAY LA 14840 HYDRALAZINE 14845 HYDRALAZINE COMPLEX 14850 HYDRALAZINE THIAZIDE 14855 HYDRALAZINE-HYDCHLORTHIAZ-RESERP 14860 HYDRAMINE 14865 HYDRAP-ES 14870 HYDRATE 14875 HYDREA 14880 HYDRIODIC ACID 14885 HYDRO ERGOLOID 14890 HYDRO PROPANOCLAMINE SYRUP 14895 HYDRO-CHLOR 14900 HYDRO-CHLORO-THIAZIDE 14905 HYDRO-RESERPINE 14910 HYDRO-Z 14915 HYDROBEXAN 14920 HYDROCHLORIC ACID 14925 HYDROCHLORIC ACID 1/500 14930 HCT 14930 HYDROCHLOROTHIAZIDE 14935 HYDROCHLOROTHIAZIDE W/RESERPINE 14940 HYDROCHLORULAN 4945 HYDROCIL 14950 HYDROCIL FORTIFIED 14955 HYDROCOCONE 14960 HYDROCODONE SYRUP 14965 HYDROCORTISONE 14970 HYDROCORT IODOCHLORHYDROXQUIN 14975 HYDROCORTISONE W/NEOMYCIN 14980 HYDROCORTONE 14985 HY DRODIURIL 14990 HYDROGEN PEROXIDE 14995 HYDROLOSE SYRUP 6b 15000 15005 HYDROMAL HYDROMORPHONE 15010 HYDROMOX 15015 15020 0 15030 15035 15040 15045 15050 15055 15060 15065 15070 15075 15080 15085 15090 15095 15100 15105 15110 15115 15120 15125 15130 15135 15140 15145 15150 15155 HYDROMOX R HYDROPHED HYDROPHI INTM HYDROPLUS HYDROPRES HYDROQUINONE HYDROSERPINE HYDROSTERONE HYDROTENSIN HYDROXOCOBALAMIN HYDROXY~PROGESTERONE HYDROXYCHLOROQUINE HYDROXYPROGEST W/ESTRADIOL HYDROXYQUINONE SKIN BLEACHING HYDROXYSTILBAMIDINE HYDROXYUREA HYDROXYZINE COMPOUND SYRUP HYDROXYZINE HYDROXY ZINE PAMOATE HYGEFEM HYGROTON HYKI NONE HYLATE HYLI VER HYLI VER PLUS HYLUTIN HYLUTIN-EST HYONATOL-B ELIXIR HYOSCYAMUS 15160 HYOSOPHEN 15165 15170 HYPAQUE HYPAQUE-M 15175 HYPAROTIN 15180 15185 HYPER-CHOLATE HYPER-RAUW 15190 HYPER-TET 15195 15200 15205 15210 15215 15220 15225 15230 15235 15238 15240 15243 15245 15250 15255 15260 15265 15270 15275 15280 15285 HYPERAB HYPERHEP HYPERINE HYPERLYTE CONCENTRATE HYPERS AL HYPERSTAT I.V. HYPERT ABS HYPERTUS SIS HYPNAL DYNE HYPNOTIC AGENT HYPOTEARS HYPOTENSIVE AGENT HYPRHO—D = HYPROT IGEN HYPROVAL P.A. HYPTRAN HYREXIN HYSE RP HYSKON HYSONE HYTAKEROL MEDICATION CODE LIST, NAMCS 1980 15290 15295 15300 15305 15310 15315 15320 15325 15330 15335 15340 15345 15350 15355 15360 15365 15370 15375 15380 15385 15390 15395 15400 15405 HYTINIC HYTINIC-PLUS HYTONATOL-B HYTONE HYTUSS HYVA GENT VIOL VAG TAB HYZ INE 1.D. 50 I-L-X W/B-12 I-LITE DROPS I-PAC I-RON I-SEDRIN PLAIN 1-10-S I-10-W 1-5-W IBERET IBERET-FOLIC-500 IBERET-500 IBEROL IBEROL-F IBUPROFEN ICHTHAMMOL ICHTHYMALL 15410 ICHTHYO 15415 15420 15425 15430 15435 15440 15445 15450 15455 15460 15465 15470 15475 15480 15485 15490 15495 15500 15505 15510 15515 15520 15525 15530 15535 15540 15540 15545 15550 15555 15560 15565 15570 15575 15580 ICHTHYOL CONCENTRATE ICN AMIT ICN AZEPOX ICN HYTHIDE ICN ISOX ICN TOLAM ICN 65 COMPOUND ICTOTEST ICY HOT ANALGESIC BALM IDENAL IDOCORT-AVP IDOXUR ID INE ILETIN ILOPAN ILOPAN-CHOL INE ILOSONE ILOTYCIN ILOTYCIN GLUCEPTATE I.V. ILOZYME IMAVATE IMFERON IMIPRAMINE IMMU-6 IMMU-TE TANUS IMMUGLOBIN GLOBULIN IMMUNE SERUM GLOBULIN IMODIUM IMPREGON CONCENTRATE IMURAN INAP SINE INCENDO INCREMIN W/IRON INDERAL INDERIDE 15585 INDIGO CARMINE 15590 INDOCIN 15595 INDOGESIC 15600 INDOMETHACIN 15605 NFADORM DROPS 15610 INFANTOL PINK 15615 INFLAMASE 15620 INFLUENZA VIRUS VAC TYPE A,B 15625 INFRARUB ANALGESIC CREAM 15630 INH 15635 INHISTON 15640 INITIA DROPS 15645 INITIA DROPS W/FLUORIDE 15650 INNOVAR 15655 INOSITOL 15660 INOTON 15665 INPERSOL 15670 INSTA GLUCOSE 15675 INSTRUMENT STERILIZING SOLUTION 15680 INSULIN 15685 INTAL 15690 INTENSIN 15695 INTRALIPID 15700 INTRINSITINIC 15705 INTROPIN 15710 INULIN 15715 INV-10-W 15720 INVERSINE 15725 10DIDES TINCTURE 15730 IODINE 15735 10DIZED LIME 15740 1IODIZED OINT 15745 10DO-NIACIN 15750 I10DOCHLOR 15755 I10DOCHLOR W/HYDROCORTISONE 15760 IODOCHLORHYDROXYQUIN 15765 I10DOCHLORHYDRXQUN HYDROCORTISONE 15770 I10DOCORT 15775 10DOSONE 15780 IONAMIN 15785 IONAX 15790 IONIL 15795 IO0ONOSOL 15800 ICPREP 15805 I0SEL 250 15810 1IPECAC 15815 IPRENOL 15820 IPSATOL P SYRUP 15825 1IPSATOL 15830 IPSATOL/DM SYRUP 15835 IRCON 15840 IRCON-FA 15845 1IRODEX 15850 IROLONG GRANUCAP 15855 IROMIDE 15860 IROMIN-G 15865 IRON &€ B COMPLEX PLUS 15870 IRON PREPARATION 15875 IRON DEXTRAN 15880 IRON QUININE & STRYCHNINE 0S 15885 IRON W/VITAMIN C 15890 IRONATE-B PLUS 15895 [IRONCO B 15900 IRONIZED YEAST 15905 IROPHOS-D 15910 IRRIGATING SOLUT ION 15915 IRRIGOL 5920 ISMELIN 15925 1ISMOTIC 15930 1S0O-BID 15935 I1SOBUTAL 15940 ISOCAL 15945 ISOCLOR 15950 ISOCLOR EXPECTORANT 15955 1SOCLOR TIMESULE 15960 ISODINE 15965 ISOETHARINE 15970 ISOLATE COMPOUND ELIXIR 15975 ISOLEUCINE 15980 ISOLLYL 15985 ISOLYTE 15990 ISONIAZID 15995 ISONICOTINIC ACID 16000 ISOPAQUE 16005 ISOPHRIN 16010 ISOPRO T.De. 16015 ISOPROPAZINE 16020 ISOPROPYL ALCOHOL 16025 [ISOPROTERENOL 16030 ISOPTO ALKALINE 16035 [1SOPTO ATROPINE 16040 1SOPTO CARBACHOL 16045 T1SOPTO CARPINE 16050 ISOPTO CETAMIDE 16055 1SOPTO CETAPRED 16060 ISOPTO ESERINE 16065 ISOPTO FRIN 16070 ISOPTO HOMATROPINE 16075 1SOPTO HYOSCINE 16080 ISOPTO P-ES 16085 1SOPTO PLAIN 16090 ISOPTO TEARS 16095 ISORDIL 16100 ISORDItL W/PHENOCBARBITAL 16105 1SOSORBIDE 16110 ISOTRATE 16115 ISOXSUPRINE 16120 ISTRIAN NUTGALLS yBENZCNE, ZINC 6125 SUPREL COMPOU IXIR 16130 ISUPREL 16135 ISUPREL HCL GLOSSET 16140 ISUPREL HCL MISTOMETER 16145 I1VADANTIN 16150 IVY-CHEX 16155 1VY-RID LOTION 16160 JACOBSON'S SOLUTION 16165 JANIMINE 16170 JECTASONE 16175 JECTO-SAL 16180 JERI-BATH MEDICATION CODE LIST, NAMCS 1980 16185 16190 16195 16200 16205 16210 16215 16220 16225 16230 16235 16240 16245 16250 16255 16260 16265 16270 16275 16280 16285 16290 16295 16300 16305 16310 16315 16320 16325 16330 16335 16340 16345 16350 16355 6360 16370 16375 16380 16385 16390 16395 16400 16405 16410 16415 16420 16425 16430 16435 16440 16445 16450 16455 16460 16465 16470 16475 16480 JEVERON K-B=P W/0PIUM K-C K-CILLIN 500 K-G ELIXIR K-LOR K-LYTE DS K-L YTE/CL K-P K-PAVA K-PEN K-PHEN K-PHOS K-PHOS (MODIFIED) K-PHOS NEUTRAL K-PHOS NO. 2 K-PHOS W/S.A.P. K-Y STERILE LUBRICATING JELLY KA-PEK KADAL EX KAFOC IN KALCINATE KAMADROX KANALKA KANAMYC IN KANTREX KANULASE KAO NOR KAOCHLOR KAOCHLOR-EFF KAOLIN KAOLIN & PECTIN KAOMEAD KAOMEAD PG KAOMEAD W/BELLADONNA ALKALOIDS KAON KAOPECTATE KAOPHEN KAOPHIL KAPECTOLIN KAPECTOLIN PG KAPECTOLIN W/BE LLADONNA KAPECTOLIN W/PAREGORIC KAPPADIONE KARAYA GUM KARI-RINSE KARIDIUM KARIGEL KA SOF KATO KAVRIN KAY CIEL KAY-POTE KAYBOVITE KAYEXALATE KAYLIXIR KAYTRATE-30 KAYKAP KEFF KEFLEX KEFLIN 16365 16485 KEFZOL 16490 KELEX 16495 KEMADRIN 16500 KENACORT 16505 KENALOG 16510 KENPECTIN 16515 KENPECTIN-P 16520 KERALYT 16525 KERI 16530 KERID DROPS 16535 KERODEX 16540 KESSADROX 16545 KESSO-BAMATE 16550 KESSO-MYCIN 16555 KESSO-PEN 16560 KESSO-TETRA 16565 KESSODRATE 16570 KESTRIN 16575 KESTRONE 16580 KETAJECT 16585 KETALAR 16590 KETAMINE 16595 KETO-DIASTIX 16600 KETOCHOL 16605 KETOSTIX 16610 KEY-PLEX 16615 KEY-PRED 16620 KI-N 16625 KIDDI KOFF SYRUP 16630 KIDDISAN 16635 KIE 16640 KINESED 16645 KINEVAC 16650 KLARON 16655 KLEBCIL 16660 KLEER 16665 KLEER COMPOUND 16670 KLEER EXPECTORANT 16675 KLEER EXPECTORANT DH 16680 KLEER IMPROVED 16685 KLEER JUNIOR 16690 KLEER SYRUP 16695 KLEER-MILD 16700 KLEER-TUSS 16705 KLIX 16710 KLOR-CON 16715 KLORIDE ELIXIR 16720 KLORVESS 16725 KLORVESS EFFERVESCENT GRANULE 16735 KOATE 6740 KOLANTYL 16745 KOLDEZE 16750 KOLYUM 16755 KOMAZINE 16760 KOMED HC LOTION 16765 KOMED 16770 KOMED MILD LOTION 16775 KOMEX 16780 KONAKION 16785 KONDREMUL 1G 16790 KONDREMUL W/ CASCARA 16795 KONDREMUL W/PHENOLPHTHALEIN 16800 KONSYL 16805 KONYNE 16810 KORIZOL 16815 KORO-SULF 16820 KOROMEX 16825 KOROSTATIN 16830 KREM 16835 KRONOHIST KRONOCAP 16840 KRUSCHEN SALTS 16845 KU-ZYME 16850 KU-ZYME HP 16855 KUDROX 16860 KUTAPRESSIN 16865 KUTRASE 16870 KWELL 16875 L.A. FORMULA 16880 L.F.B.-12-100 16885 L-CAINE 16890 L-T-S 16895 L-THYROXINE 16900 LA-12 16905 LABID 16910 LABSTIX 16915 LACRI-LUBE 16920 LACRIL 16925 LACT-AID 16930 LACTATED PEPSIN ELIXIR 16935 LACTATED RINGER'S (HARTMANN®S) 16940 LACTIC ACID 16945 LACTICARE LOTION 16950 LACTINEX 16955 LACTOBACILLUS ACIDOPHILUS 16960 LACTOCAL 16965 LACTOCAL-F 16970 LACTOSE 16975 LACTULOSE 16980 LANABAC 16985 LANABARB 16990 LANABEE-C 16995 LANABROM ELIXIR 17000 LANABURN 17005 LANACILLIN VK 17010 LANAMINS 17015 | ANASED 17020 LANATOSIDE 17025 LANATRATE 17030 LANATUSS EXP ECTORANT 17035 LANAURINE 17040 LANAVITE 17045 LANAZETS 17050 LANESTRIN 17055 LANIAZID 17060 LANNATES ELIXIR 17065 LANOKALIN 17070 LANOLIN 17075 LANOLINE 17080 LANOLOR 17085 LANOPHYLLIN MEDICATION CODE LIST, NAMCS 1980 17090 LANOPLEX 17095 LANOPLEX ELIXIR 17100 LANOPLEX FORTE 17105 LANOCRINAL 17110 LANOTHAL PILL 17115 LANOXIN 17120 LANTEEN JELLY 17125 LANTRISUL 17130 LANVISONE 17135 LARGON 17140 LAROBEC 17145 LARODOPA 17150 LAROTID 17155 LARYLGAN 17160 LASAN 17165 LASIX 17170 LASSAR'S PASTE 17175 LAUD-IRON 17180 LAUD-IRON FOLIC 17185 LAUD-IRON FORTE 17190 LAUD-IRON PLUS 17195 LAUDACIN SYRUP 17200 LAVACOL 17205 LAVATAR 17210 LAVENDER OIL 17215 LAVERIN 17220 LAX-HERB 17225 LAXADANE 17230 LAXADANE SUPULE 17235 LAXAMEAD P.H.M. 17240 LAXINATE 100 17243 LAXATIVE 17245 LAXOGEN 17250 LC-65 CLEANING 17255 LEAD & OPIUM WASH LOTION 17260 LEAD ACETATE 17265 LECITHIN 17270 LEDERCILLIN VK 17275 EDERPLEX 17280 LEMIVITE 17285 LEMON OIL 17290 LENS-MATE 17295 LENSINE 17300 LENSRINS 17305 LEPTINOL 17310 LERITINE 17315 LEUCINE 17320 LEUCOVORIN 17325 LEUKERAN 17330 LEVAMINE 17335 LEVO UROQID ACID 17340 LEVO-DROMORAN 17345 LEVODOPA 17350 LEVOID 17355 LEVOPHED 17360 LEVOPROME 17365 LEVOTHROID 17370 LEVOTHYROXINE 17375 LEVSIN 17380 LEVSIN-PB DROPS 17385 LEVSIN/PHENOBARBITAL 17390 LEVSINEX 17395 LEVSINEX/PHENOBARBITAL 17400 LEVUCAL 17405 LEXAVITE 17410 LEXOR 17415 LEXTRON 17420 LI-BAN SPRAY 17425 LIB-E 17430 LIBCO-12 17435 LIBIGEN 17440 LIBRAX 17445 LIBRITABS 17450 LIBRIUM 17455 LICOPLEX DS 17460 LICORICE COMPOUND 17465 LIDA-MANTLE 17470 LIDA-MANTLE-HC 17475 LIDEX 17480 LIDINIUM 7485 L1 AIN 17490 LIDOCAINE HCL 7495 0 N 17500 LIDOJECT 17505 LIDONE 17510 LIDOSPORIN 17515 L1F12 17520 LIFOJECT 17525 LIFOLBEX 17530 LIMBITROL 17535 LIME WATER 17540 LINCOCIN 17545 LINCOMYCIN 17550 LINOLESTROL 17555 LINSEED OIL 17560 LIORESAL 17565 LIPAN 17570 LIPO GANTRISIN 17575 LIPO THREE 17580 LIPO-ADRENEX IN OIL 17585 LIPO-HEPIN 17590 LIPO-NICIN 17595 LIPODERM 17600 LIPOFLAVONOID 17605 LIPOGEN 17610 LIPOMUL-ORAL 17615 LIPONOID 17620 LIPOSPERSE 17625 LIPOSYN 17630 LIPOTRIAD 17635 LIPOVITE 17640 LIQUAEMIN 17645 LIQUALGINE 17650 LIQUAMAR 17655 LIQUI-CEE 17660 LIQUI-DOSS 17665 LIQUID LUBRICATING JELLY 17670 LIQUIFILM 17675 LIQUIMAT 17680 LIQUIPRI W INEPHR IN cs 17685 LIQUIX C 17690 LIQUOPHYLLINE 17695 LIQUOR CARBONIS DETERGENS 17700 LISTEREX 17705 LISTERINE 17710 LITHANE 17715 LITHIUM 17720 LITHIUM CITRATE SYRUP 17725 LITHOBID 17730 LITHONATE 17735 LITHOTABS 17740 LIVA-ZYME 17745 LIVER DESICCATED W/B-12 17750 LIVER EXTRACT 17755 LIVER 17760 LIVERy IRON & VITAMINS 1 7765 LIVITAMIN 17770 LIVITAMIN LIQUID 17775 LIVITAMIN PRENATAL 17780 LIVITAMIN W/ INTRINSIC FACTOR 17785 LIVITOL W/ VITAMIN C B-12 & IRON 17790 LIVONAMINE 17795 LIVOREX W/B-12 17800 LIXAGESIC ELIXIR 17805 LIXAMINOL AT ELIXIR 17810 LIXAMINOL 17815 LMD 17820 LO-TUSSIN SYRUP 17825 LO/OVRAL 17830 LOBANA HOSPITAL LOTION 17833 LOCAL ANESTHETIC 17835 LOCORTEN 17840 LOESTRIN 17845 LOFENALAC 17850 LOFENE 17855 LOFLO 17860 LOMANATE 17865 LOMOTIL 17868 LOMUSTINE 17870 LONALAC 17875 LONITEN 17880 LOPERAMIDE 17885 LOPRESSOR 17888 LORAZEPAM 17890 LORELCO 17895 LORFAN 17900 LORIDINE 17905 LOROXIDE 17910 LOROXIDE-HC 17915 LORYL 17920 LOTIO ALSULFA 17925 LOTRIMIN 17930 LOTUSATE 17935 LOWILA 17940 LOXAPINE 17945 LOXITAN 17950 LUBAFAX 17955 LUBATH 17960 LUBRASEPTIC 17965 LUBRASOL BATH OIL MEDICATION CODE LIST, NAMCS 1980 17970 LUBRICATING JELLY 17975 LUBRIDERM 17980 LUFA 17985 LUFTODIL 17990 LUFYLLIN 17995 LUFYLLIN-EPG 18000 LUFYLLIN-GG 18005 LUFYLLIN-400 18010 LUGOL"*S SOLUTION 18015 LUMINAL 18020 LR IDE DROPS 18025 LV PENICILLIN 18030 LYCOLAN ELIXIR 18035 LYNAPAP 18040 LYNETUSS 18045 LYO B-C 18050 LYSINE 18055 LYSODREN 18060 LYSOL 18065 LYTEERS 18070 LYTREN 18075 M.A.H. #2 18080 M.V.l. 18085 M-C ANTACID 18090 M-CAL-M 18095 M-CILLIN 18100 M-M-R 18105 M-R-VAX II 18110 M-TETRA 250 18115 M-VAC (EDMONSTON STRAIN) 18120 M-Z 18125 MAALOX 18130 MACRODANTIN 18135 MACRODEX 18140 MAG-CAL 18145 MAGAN 18150 MAGCYL 18155 MAGMALIN 18160 MAGNAGEL 18165 MAGNALOX 18170 MAGNAL UM 18175 MAGNATRIL 18180 MAGNESIA 18185 MAGNESIA & ALUMINA 18190 MAGNESIUM CHLORIDE 18195 MAGNESIUM CITRATE 18200 MAGNESIUM GLUCONATE 18205 MAGNESIUM OXIDE 18210 MAGNESIUM PHOSPHATE TRIBASIC 18215 MAGNESIUM SULFATE 18220 MAGNESIUM TRISILICATE 18225 MAJOR B W/C 18230 MAKROGEN 18235 MALATAL 18240 MALLAMINT 18245 MALLENZYME 18250 MALLERGAN 18255 MALLERGAN SYRUP PLAIN 18260 MALLERGAN W/CODEINE SYRUP 18265 MALLISOL SURGICAL SCRUB 18270 18275 18280 18285 18290 18295 18300 18305 18310 18315 18320 18325 18330 18335 18340 18345 18350 18355 18360 18365 18370 18375 18380 18385 18390 18395 18400 18405 18410 18415 18420 18425 18430 18435 18440 18445 18450 18455 18460 18465 18470 18475 18480 18485 18490 18495 18500 18505 18510 18515 18520 18525 18530 18535 MALLO-PECT IN MALLOPRESS MALOGEL MALOGEN MALOTUSS SYRUP MALT SUPEX MAMMOL MANACID MANDEL AHAB MANDELAMINE MANDEX MANDOL MANESE MANGANESE GLUCONATE MANI RON MANNITOL MANNITOL HEXANITRATE MANNITOL HEXANITRATE W/PB MANNITOL 18 GA MANTADIL MAOLATE MARAX MARAZIDE MARBAXIN MARBLEN MARCAINE MARCAINE HCL W/EPINEPHRN MAREZ INE MARFLEX MARFLEX PLUS MARFLEX 60 MARGESIC COMPOUND NO. 65 MARGES IC MARMINE MARNAL MARPL AN MARUATE SPANTAB MASSE MASSENGILL DISPOSABLE DOUCHE MASSENGILL DISP DOUCHE-VINEGAR MASSENGILL LIQUID CONCENTRATE MASSENGILL POWDER MATERNA MATROPINAL MATULANE MAXAFIL MAXAMAG MAXI-E MAXIBOLIN MAXIDEX MAXITROL MAYT REX MCT OIL MEADLAX-DSS ES v1 VA £9 18565 MED-DEPO 18570 MEDACHE 18575 MEDAPED 18580 MEDI - MARKER 18585 MEDI -QUIK SPRAY 18590 MEDI-TRATING 18595 MEDIATRIC 18600 MEDICATED COUGH DROPS 18605 MEDICATED FOOT POWDER 18610 MEDICONE DRESSING 18615 MEDICONET 18620 MEDIHALER-EPI 18625 MEDIHALER-ERGOTAMINE 18630 MEDIHALER-ISO 18635 MEDRALONE 18640 MEDROL 18645 MEFOXIN 18650 MEGA-B 18655 MEGACE 18660 MEGADOSE 18663 MEGESTROL 18665 MELFIAT 18670 MELLARIL 18675 MELLOBATH BATH OIL 18680 MELOZETS WAFER 18685 MELPHALAN 18690 MENADIONE 18695 MENEST 18700 MENIC 18705 MENINGOVAX-C 18710 MENOJECT 18715 MENOMUNE A/C 18720 MENOMUNE 18725 MENOMUNE-C 18730 MENOTAB 18735 MENRIUM 18740 MENTHALGESIC 18745 MENTHOL 18750 MENTHOLATUM 18755 MEPERGAN 18760 MEPERIDINE 18765 MEPHOBARBITAL 18770 MEPHOHAB 18775 MEPHYTON 18780 MEPRED 18785 MEPRO COMPOUND 18790 MEPRO-PETN 18795 MEPROBAMATE 18800 MEPROSPAN 18805 MEQUIN 18810 MER ESTRONE 18815 MERBROMIN 18818 MERCAPTOMERIN 18820 MERCAPTOPURINE 18825 MERCODOL W/DECAPRYN COUGH SYRUP 18830 MERCRESIN 18835 MERCURIC CHLORIDE 18840 MERCURIE IODIDE 18845 MERCURIC PREPARATION 18850 MERCURIC SULFIDE MEDICATION CODE LIST, NAMCS 1980 18855 MERCUROCHROME 18860 MERCUROPHYLLINE 18865 MERCURY OXIDE 18870 MERCUTHEOLIN 18875 MERPHOL 18880 MERSA 18885 MERSALYL-THEOPHYLLINE 18890 MERSALYN 18895 MER SOL 18900 MERTHIOLATE 18905 MERUVAX II 18910 MERVALDIN 18915 MESANTOIN 18920 MESTINON 18925 METAHYDRIN 18930 METAMUC IL 18935 METANDREN 18940 METAPREL 18945 METAPROTERENOL 18950 METARAMINOL 18955 METASEP 18960 METATENSIN 18965 METH 18970 METH CHOL INE 18975 METHABOLIC 18980 METHACHOLINE 18985 METHADONE 18990 METHAKOTE 8995 METHALATE 19000 METHAMPEX 19005 METHAMPHETAMINE 19010 METHANOL 19015 METHAPYRILENE HCL POWDER 19019 METHAQUALONE 19020 METHAVIN 19025 METHAZOLAMIDE 19030 METHAZOLATE 19035 METHENAMINE 19040 METHENAMINE MANDELATE 19045 METHERGINE 19050 METHICILLIN 19055 METHIONINE 19060 METHISCHOL 19065 METHO-500 19070 METHOCARBAMOL 19075 METHOCARBAMOL W/ASA 19080 METHOCARBAMOL W/ASPIRIN 19085 METHOSARB 19090 METHOTREXATE 19095 METHOXAL 19100 METHOXANOL 19105 METHOXSALEN 19110 METHSCOPOLAMINE 19115 METHULOSE 19120 METHYL ALCOHOL 19130 METHYL SALICYLATE 19135 METHYL-CYST 19140 METHYLCELLULOSE 19145 METHYLCELLULOSE 1500 CPS 19150 METHYLCELLULOSE 4000 CPS 19155 19160 19165 19170 19175 19180 METHYLDOPA METHYLENE BLUE METHYL ONE METHYL PARABEN METHYLPHENIDATE ETHYL PREDNISOLONE 19185 ETHYL TESTOSTERONE 19190 EZ ETI-DERM 19195 19200 19205 19210 19215 19218 19220 19225 19230 19233 19235 19240 19245 19250 19255 19260 19265 19270 19275 19280 19285 19290 19295 19300 19305 19310 19315 19320 19325 19330 19335 19340 19345 19350 19355 19360 19365 19370 19375 19380 19385 19390 19395 19400 19405 19410 MET ICORT EL ONE METICORTEN MET IMYD METOLA ZONE METOPIRONE METOPROLOL METRA METRAZOL METRETON METRONIDAZOLE METUBINE METYCA INE MEVANIN-C MEVATINIC-C MEXATE MEXS ANA MG-BLUE MI-CEBRIN MI-THERIC MICATIN MICO-ONE MICONAZOLE MICRAININ MICRHOGAM MICRIN PLUS MICROCORT LOTION MICRONEFRIN MICRONOR MICROSTIX MICROSUL FON MICROSYN ACNE LOTION MICTROL MIDICEL MIDOL MIDRAN DECONGESTANT MIDRIN MIGRAL MILK OF BISMUTH MILK OF MAGNESIA MILK OF MAGNESTA-CASCARA SAGRADA MILK OF MAGNESIA-MINERAL OIL MILK SKIM BAKER®S READY-4 MILKINOL MILONT IN MILPATH MILPREM 19415 MILTOWN 19420 MILTRATE 19425 19430 19435 19440 MIN-GERA MIN-HEMA MIN-HEMA-PLUS MIN-PLEX 300 19445 19450 19455 19460 19465 19470 19475 19480 19485 19490 19495 19500 19505 19508 19510 19515 19520 19525 19530 19535 19540 19545 19550 19555 19560 19565 19570 19575 19580 19585 19590 19595 19600 MINERAL OIL MINI-LIX MINI PRESS MINOCIN MINOCYCLINE MINOPLEX MINOTAL MINRO-PLEX MINT-0-MAG MINTEZOL MIBCEL MIOCHOL MIOSTAT MIGTIC AGENT MIRADON MISSION PRENATAL MISSION PRENATAL F.A. MISSION PRENATAL H.P. MISTOL MISTOL W/EPHEDRINE MITEY VITES MITHRACIN MITHRAMYCIN MITY-MYCIN MITY-QUIN MIVERT MI ZYME MOBAN MOBIDIN MOBIGESIC MOBI SYL MODANE MODERIL 19605 HMODICON 19610 19615 19620 19625 19630 MODI CUM MODIFIED PROTEIN HYDROLYSATE MOEB IQUIN MOL—-IRON MOL-IRON W/ VITAMIN C 19635 19640 19645 19650 19655 19660 19665 19670 MONI STAT MONI STAT 7 MONOSI NE MORPHINE MORPHINE & ATROPINE MORRHUATE MORUGUENT MORUSAN 19675 MOTRIN 19680 19685 19690 19695 19700 19705 19710 19715 19720 19725 19730 MOUTHWASH MOUTHWASH & GARGLE MOUTHWASH ASTRINGENT MOVI COL GRANULE MSC TRIAMINIC TIMED-RELEASE MUCI LAX MUCILOSE MUCOMYST MUCOMYST W/ISOPROTERENOL MUCOMYST-10 MUCOPLEX 19735 MUDRANE MEDICATION CODE LIST, NAMCS 1980 19740 19745 19750 19755 19760 19765 19770 19775 19780 19785 19790 19795 19800 19805 19810 19815 MULL-SOY MUL TA-GEN 12 MULTALAN MUL TI-B-PLEX MUL TI-CHEWZ MUL TI-HEMA MULTI-SYMPTOM MULTI-V MULTI-VI-FLOR DROPS MUL TI-VITAMIN MULTI-VITES MULTICEBRIN MUL TIFUGE SYRUP MULTIGEST MULTIPLE B-C MULTIPLE VITAMIN 19820 MULTIPLE VITAMIN W/IRON 19825 MULTIPLE VITAMINS 19830 9835 19840 [9845 19850 19855 19860 19865 19870 19875 19880 19885 19890 19895 19900 19905 19910 MULTIPLE VITAMINS W/IRON MULTIPLE VITAMINS W/MINERALS MULTIVITAMIN MUL TIVITAMIN CONCENTRATE MULTIVITAMIN FORMULA MUL TIVI TAMIN THERAPEUTIC MULTIVIT THERAPEUTIC W/MINS MUL TIVITAMIN/MULTIMI NERAL MUL TIVITAMINS & MULTIMINERALS MUL TIVITAMINS ROWELL MUL TOHAB MULVIDREN MUL VIDREN-F SOFTAB MUL VITAB MUMPS SKIN TEST ANTIGEN MUMPS VIRUS VACCINE LIVE MUMPSVAX JERYL LYNN 19915 19920 19925 19930 19935 19940 19945 19950 19955 19960 19965 19970 19975 19980 19985 19990 19995 20000 20005 MURCIL MURI-LUBE MUR INE MUR IPSIN MUROCARB MUROCEL MUROCOLL MUSTARD OIL MUSTARGEN MUTAMYCIN MUVICA MY-B-DEN MY-CORT MY-CORT #1 MY-CORT #2 MY-CORT LOTION MYADEC MYAMBUTOL MYAMEAD 20010 MYCELEX 20015 20020 20025 20030 20035 MYCHEL MYCI SPRAY MYCIFRADIN SULFATE MYC IGUENT MYCITRACIN 20040 MYCO TAC 20045 MYCO TRIACET 20050 MYCOFAN 20055 MYCOLOG 20060 MYCOSTATIN 20065 MYDFRIN 20070 MYDRAPRE 20075 MYDRIACY 20078 MYDRIATI 20080 MYLANTA 20085 MYLAXEN 20090 MYLERAN 20095 MYLICON 20100 MYOBID DIALSPAN 20105 MYOCHRYS INE 20110 _ MYOFLEX 20115 MYOLIN 20120 MYOTONACHOL 20125 MYRINGACAINE 20130 MYRRH 20135 MYSOLINE 20140 MYSTECLIN F 20145 MYTELASE 20150 MYTRATE 20155 N-MILTISTIX 20160 N-URISTIX 20165 NABILENE 20170 NAFCIL 20175 NAFCILLIN 20180 NAFEEN 20185 NAILIFE 20190 NALDECON 20195 NALDEGESIC 20200 NALDELATE SYRUP 20205 NALDETUSS SYRUP 20210 NALFON D AGENT NINN ~N 20215 NALIDIXIC ACID 20220 NAMIDE 20225 NAMIDE-C 20230 NANDROLATE 20235 NANDROL IN 20240 NANDROLONE 20245 NANDROLONE PHENPROPIONATE 20250 NAPAL 20255 NAPHAZOL INE 20260 NAPHCON 20265 NAPHCON FORTE 20270 NAPHCON-A 20275 NAPLOPAN 20280 NAPRIL PLATEAU CAP 20285 NAPROSYN 20290 NAPROXEN 20295 NAPTRATE 20300 NAQA 20305 NAQUIVAL 20310 NARCAN 20313 NARCOTIC ANTAGONIST 20315 NARDIL 20320 NASAHIST 20325 NASAL DECONGESTANT Gg 20330 20335 20340 20345 20350 20355 20360 20365 NASAL DECONGESTANT ELIXIR NASAL DECONGESTANT NEO-VADRIN NASCOBARB NASO-MIST NASOBID GRANUCAP NASOCON NATABEC NATABEC RX 20370 NATABEC W/ FLUORIDE 20375 NATABEC-FA 20380 20385 NATACYN NATAFORT 20390 20395 20400 20405 20410 20415 20420 20425 20430 20435 20440 20445 20450 20455 20460 20465 20470 20475 20480 20485 20490 20495 20500 20505 20510 20515 20520 20525 ! NATALINS NATALINS RX NATRONA COMPOUND NATURAL VEGETABLE LAXATIVE NATURE'S REMEDY NATURETIN NATURETIN W/K NAUS ATROL NAUS ETROL NAVANE NAZAC DECONGESTANT ND-GESIC ND-STAT NEBCIN NECHLORIN NECORT-AVP NECTA SWEET NECT ADON NEFROSUL NEGATAN NEGG RAM NELEX-100 NEMA SE NEMBUTAL NEOC—-BETALIN 12 CRYSTALLINE NEO-CALGLUCON SYRUP NEC-CHOLEX NEO-CORT-DOME OTIC 20530 20535 20540" NEO-CORTEF NEC-CULTOL NEC—-DELTA-CORTEF 20545 20550 20555 20560 20565 20570 20575 20580 20585 £ NEC-FLO NEOC-HYDELTRASOL NEC—-MEDROL NEC-MIST NEO-MULL-SOY NEC—-NI LOREX NEO-OXYLONE NEC-POLYCIN NEOC—-SYNALAR 20590 NEO-SYNEPHRINE COMPOUND COLD 20595 NEO-SYNEPHRINE 20600 NEO-SYNEPHRINE II 20605 NEO-SYNEPH MENTH SPRAY 20610 NEO-TEARS 20615 NEOBID GRANUCAP 20620 NEOBIOTIC 20625 NEOCASTADERM MEDICATION CODE LISTs NAMCS 1980 20630 20635 20640 20645 20650 20655 20660 20665 20670 20675 20680 20685 20690 20695 20700 20705 20710 20715 20720 20725 20730 20735 20740 20745 20750 20755 20760 20765 20770 20775 20780 20785 20790 20795 20800 20805 20810 20815 20820 20825 20830 20835 20840 20845 20850 20855 20860 20865 20870 20875 20880 20885 20890 20895 20900 20905 20910 20915 20920 20925 NEOCET NEOCHOLAN NEOCURTASAL NEOCYLATE NEOCYTEN NEODECADRON NEODECA SPRAY NEOFED NEOGESIC NEOLAX NEOLOID NEOMIXIN NEOMYC IN NEON IC NEOP AP NEOPAVRIN FORTE NEOQUESS NEOSONE NEOSORB PLUS NEO SPECT NEOSPORIN NEOSTIGMINE NEO STIGMINE METHYLSULFATE NEOTABS NEO TAL NEOTEP NEOTHYLL INE NEO THYLL INE-GG NEOTRIZINE NEOVICAPS NEOXYN LOTION NEOZIN NEOZYL NEPHRAMINE NEPHROX NEP TAZANE NER VOCA INE NESACA INE NEUROSINE NEUROVAL ELIXIR NEUT NEUTRA-PHOS NEUTRA-PHOS-K NEUTRACOMP NEUTRALOX NEUTROGENA NEVROTOSE NI-FWRIN NI-SPAN NIAC NIACAL NIACIN NIACINAMICE NIACINAMIDE NIALEX NIALEXO-C NIAPENT NIASCORB NICALEX NICL 20930 20935 20940 20945 20950 20955 20960 20965 20970 20975 20980 20985 20990 20995 21000 21005 21010 21015 21020 21025 21030 21035 21040 21045 21050 21055 21060 21065 21070 21075 21080 21085 21090 21095 21100 21105 21110 NICO-METRAZOL NICO-SPAN NICO-VERT NICO-400 NICOBID NICOCAP NICOLAR NICOTAL NICOT INAMIDE NICOTINEX ELIXIR NICOTINIC ACID NICOTINYL ALCOHOL NICOTINYL ALCOHOL TARTRATE NICOZIN-C NICOZOL NIDAR NIFEREX NIFEREX W/VITAMIN C NIFEREX-PN NIFEREX-150 NIFEREX-150 FORTE NIGROIDS NIKETHAMIDE NIL VAGINAL NILAIN NILCOL NILPRIN NILSTAT NINE VITAMIN NINE VITAMIN W/ IRON NIORIC NIPIRIN NIPRIDE NISENT IL NISINE ELIXIR NITRAZINE PAPER NITREX 21115 NITRIN 21120 21125 21130 21135 21140 21145 21150 21155 21160 NITRINE NITRO T.D. NITRO-BID NITROCAP NITRODAN NITROFURANTOIN NITROFURAZONE NITROGEN NITROGLYCERIN 21165 NITROGLYN 21170 N NITROL 21175 N NITROL IN 1180 NITROS PAN 1185 ~N f NITROSTAT N 1190 195 ~N NT NITROSULE NITROUS OXIDE 200 NIVEA 21205 21210 21215 NOCTEC NODOZ NOLAMINE 21220 NOLUDAR 21225 NOLV ADEX 9g 21230 NOR LIEF 21235 NOR MIL 21240 NOR PRES 21245 NOR-Q. DD. 21250 NOR-TET 21255 NORAFED 21260 NORAPHEN SPANCAP 21265 NORAVITE PLUS 21270 NORAZINE 21275 NORDRYL 21280 NORF LEX 21285 NORFRANIL 21290 NORGESIC 21295 NORIMEX PLUS 21300 NORINYL 21305 NORI SODRINE 21310 NORISODRINE SULFATE AEROHALER 21315 NORISODRINE W/ CALC IODIDE 21320 NORLAC 21325 NORLAC RX 21330 NORLESTRIN 21335 NORLUTATE 21340 NORLUTIN 21345 NORMACID 21350 NORMADERM 21355 NORMAT ANE DC EXPECTORANT 21360 NORMATANE ELIXIR 21365 NORMAT ANE 21370 NORMATANE T.D. 21375 NORMOL 21380 NORMOSOL 21385 NOROXI NE 21390 NORPACE 21395 NORPANTH 21400 _NORPRAMIN 21403 _NORTRIPTYL INE 21405 NOSE DROPS 21410 NOSPAZ 21415 NOVACEBRIN 21420 NOVACEBRIN W/FLUORIDE 21425 NOVAFED 21430 NOVAFED A 21435 NOVAFED 120 21440 NOVAHI STINE 21445 NOVAHI STINE COUGH FORMULA 21450 NOVAHISTINE DH 21455 NOVAHISTINE DMX 21460 NOVAHISTINE ELIXIR 21465 NOVAHISTINE EXPECT ORANT 21470 NOVAHISTINE FORTIS 21475 NOVAHISTINE LP 21480 NOVAHISTINE MELET 21485 NOVAHISTINE SINUS 21490 NOVALAX 21495 NOVIPLEX 21500 NOVOCAIN 21505 NOVRAD 21510 NP-27 21515 NTZ 21520 NU THERA MEDICATION CODE LISTy NAMCS 1980 21525 21530 21535 21540 21545 21550 21555 21560 21565 21570 21575 21580 21585 21590 21595 21600 21605 21610 21615 21620 21625 21630 21635 21640 21645 21650 21655 21660 21665 21670 21675 21680 21685 21690 21695 21700 21705 21710 21715 21720 21725 21730 21735 21740 21745 21750 21755 21760 21765 NU-FLCW SHAMPOO NU-IRON NU-IRON ELIXIR NU-IRON-V NU'LEVEN NUBAIN NUCOFED NUJOL NULICAINE NUM ZIT NUMORPHAN NUMZ IDENT NUPERCAINAL NUPERCAINE NUROSAL NWR SOY NUTRACORT NUTRADERM NUTRAL NUTRAMIGEN NUTRAPLUS NUTRI-1000 NUXAPHEN NYCRALAN NYDRAZID NYLIDRIN NYLMERATE NYLOXIN NYOMIN NYQUIL NYRAL NYSOLONE NYST-OLONE NYSTAFORM NYSTATIN NYSTATIN VAGINAL TABLET NYTILAX O0.B. VITAMIN 0-V STATIN OBALAN OBEGYN OBEPHEN OBEVAL OBOTAN OBRON-6 OCEAN MIST ocusoL ODOR-SCRIP OGEN 21770 21775 21780 21785 1790 1795 1800 21805 21810 21815 21820 NINN) OIL RETENTION ENEMA OILATUM SOAP OLEIC ACID OLIVE OIL OMNI-TUSS OMNIPEN ONCOVIN OP-THAL-ZIN OPACEDRIN OPASAL OPECTO ELIXIR 21825 21830 21835 21840 21845 21850 21855 21860 21865 21870 21875 21880 21885 21890 21895 21900 21905 21910 21915 21920 21925 21930 21935 21940 21945 21950 21955 21960 21965 21970 21975 21980 21985 21990 21995 22000 22005 22010 22015 22020 22025 22030 22035 22040 22045 22050 OPHTHA LIPO OPHTHA P/S OPHTHAINE OPHTHALGAN OPHTHETIC OPHTHOCHLOR 1855 OP HT HOCORT _ OPIUM OPIUM AND BELL ADONNA OPIUM CAMPHORATED OPIUM EXTRACT OPIUM POWDERED OPT-EASE OPTEF OPTILETS OPTIMINE OPTIMYD OPTISED ORA-TESTRYL ORABASE HCA PASTE ORABASE ORABASE W/BENZOCAINE PASTE ORACIN ORAGRAFIN ORALPHYLLIN ORAMINIC ORANGE OIL ORAPAV TIMECELLE ORAPHEN-PD ORAS ONE ORATRAST ORATROL ORDWAY'S SOLUTION ORENZYME ORET IC ORETICYL ORETON OREX IN ORGANIDIN ORGAPHEN ORIMUNE POLIC VAC LIVE ORAL ORINASE ORLEX H.C. OTIC ORLEX OTIC ORMAZI NE ORNACOL 22055 ORNADE 22060 22065 22070 22075 22080 22085 22090 22095 22100 22105 22110 ORNEX ORPHENADRINE ORPHENADRINE W/A.P.C. ORTHO PERSONAL LUBRICANT ORTHO-CREME ORTHO-GYNOL ORTHO-NOVUM ORTHOXICOL ORTHOX INE ORTHOXINE AND AMINOPHYLL INE 0S-CAL 22115 OS-CAL PLUS 22120 OS-CAL 500 LS 22125 OS-CAL-FORTE 22130 OS—-CAL-GESIC 22135 0S-CAL-MONE 22140 OSMI TROL 22145 OSMOGLYN 22150 OSTEOLATE 22155 OTALL DROPS 22160 OTIC DROPS 22165 OTIC PLAIN DROPS 22170 OTIC-HC 22175 OTOBIONE 22180 OTOBIOTIC 22185 OTOCOR 22190 OTOLGESIC 2219 OTOREI D-H 0 RI 22205 OUABAIN 22210 OVCON 22215 OVRAL 22220 OVRETTE 22225 OVULEN 22230 OX BILE EXTRACT 22233 OXACILLIN 22235 OXAINE M 22240 OXALIC ACID 22245 OXALID H 22255 OXLOPAR 22260 OXO-THALEIN 22265 0OXOIDS 22270 OXSORALEN 22275 OXY-KESSO-TETRA 22280 OXY-LOTICN 22285 OXY-SCRUB 22290 OXCEL COTTON-TYPE PLEDGET 22295 OXYCEL GAUZE-TYPE PAD 22300 OXYCHINOL 22305 OXYCODONE HCL & ACETAMINOPHEN 22310 OXYDESS 5S XYG EN 22320 OXYLONE 22325 OXYMETHOLONE 22330 OXYNITRAL W/VERATRUM VIRIDE 22335 OXYQUINOLINE 22340 OXYTETRACYCLINE 22345 OXYTOCIN 22350 OYSTER SHELL & VITAMIN D 22355 P NS 22360 P.A.S. SODIUM 22365 P.E.T.N. 22370 P.E.T.N. W/ PHENOBARBITAL 22375 P.H. 22380 P.H. MIXTURE 22385 P.H. PLUS 22390 P.H. 1500 22395 P.Re. SYRUP 22400 P.V. CARPINE LIQUIFILM 22405 P&S 22410 P-A-C COMPOUND 22415 P-A-C COMP W/COD SULF MEDICATION CODE LIST, NAMCS 1980 22420 P-A-V 22425 P-B-SAL 22430 P-I-N FORTE 22435 P-V TUSSIN 22440 P-W VERMIFUGE 22445 P-200 22450 PABA 22455 PABAFILM 22460 PABAGEL 22465 PABALAN 22470 PABALATE 22475 PABANOL LOTION 22480 PABIRIN BUFFERED 22485 PAGITANE 22490 PAIN REL IEVER-E 22495 PAL-A-DEX 22500 PALADAC 22505 PALAFLOR 22510 PALODYNE 22515 PAMA 22520 PAMELOR 22525 PAMINE 22530 PAMPRIM 22535 PAN ULTRA 22540 PAN-APC 22545 PAN-B-1 22550 PAN-B-6 22555 PAN-KLORIDE ELIXIR 22560 PAN-SONE 22565 PANACARB 22570 PANACID 22575 PANACORT R-P 22580 PANADYL 22585 PANAFIL 22590 PANAHIST 22595 PANALGESIC 22600 PANAMIN 22605 PANAQUIN 22610 PANASCORB 22615 PANAZID 22620 PANAZID W/B-6 22625 PANCEBRIN 22630 PANCREASE 22635 PANCREATIC HORMONE 22640 PANCREATIN 22645 PANEX 22650 PANGYN 22655 PANHEPRIN 22660 PANHYDROSONE 22665 PANI SOLONE 22670 PANMYCIN 22675 PANOXYL 22680 PANSCOL 22685 PANTENYL 22690 PANTERIC 22695 PANTHODERM 22700 PANTHODERM LOTION 22705 PANTHOL IN 22710 PANTISONE 22715 PANTOPAQUE 22720 PANTOPON 22725 PANTOTHENIC ACID 22730 PANTOTHENOL (D) ALCOHOL 22735 PAN ITEX PRENATAL W/F 22740 PANWARFIN 22745 PANZYME 22750 PAPACON CONCAP 22755 PAPASE 22760 PAPAVATRAL 22765 PAPAVATRAL W/PHENOBARBITAL 22770 PAPAVER INE 22775 PAPAYA 22780 PAPZANS 22785 PARA AMINOBENZOIC ACID 22790 PARA-AMINOSALICYLIC ACID 22795 PARACORT 22800 PARADIONE 22805 PARAFLEX 22810 __PARAFON FORTE 22815 PARAL 22820 PARALDEHYDE 22825 PARASAL SODIUM 22827 PARASYMPATHOLYTIC AGENTS 22828 PARASYMPATHOMIMETICGAGENTS 22830 PARATHYROID 22835 PARBOCYL 22840 PAREDRINE W/BORIC ACID 22845 PAREGORIC 22850 PAREGORIC & SODA COMPOUND 22855 PARELIXIR 22860 PAREPECTOLIN 22865 PAREST 22870 PARGEL 22875 PARLODEL 22880 PARLODION STRIPS 22885 PARNATE 22890 PARSIDOL 22895 PARTIAL ECLIPSE 22900 PAS 22905 PATCHOULY OIL 22910 PATHIBAMATE 22915 PATHILON 22920 PATHILON SEQUEL 22925 PATHILON W/PHENOBARBITAL 22930 PATHILON W/PHENOBARBITAL SEQUEL 22935 PATHOCIL 22940 PAVA-MEAD 22945 PAVABID 22950 PAVACAP 22955 PAVACEN 22960 PAVADUR 22965 PAVAGEN 22970 PAVAKEY S.A. 22975 PAVAR 22980 PAVASED 22985 PAVASULE PELLSULE 22990 PAVATRAN 22995 PAVERIL 23000 PAVEROLAN LANACAP 23005 PAVULON 89 23010 PAZO HEMORRHOID 23015 PBZ 23020 23025 23030 23035 23040 23045 23050 23055 23060 23065 23070 23075 23080 23085 23090 23095 23100 23105 23110 23115 23120 23125 23130 23135 23140 23145 23150 23155 23158 23160 23165 23170 23175 23178 23180 23185 23190 23195 23200 23205 23210 23215 23220 23223 23225 23230 23235 23240 23245 23250 23255 23260 23265 23270 23275 23280 23285 23290 PBZ PBZ ELIXIR PBZ LONTAB PBZ W/COD & EPHED EXP PBZ W/EPHEDRINE PBZ W/ EPHEDRINE EXPECTORANT PBZ-SR PDM PE-DE-EM PEANUT OIL USP PEATON PECTOCEL PECT CCOMP PECTOKAY PEDAMETH PEDI-DENT CV PEDI-DENT PEDIACOF PEDIACON PEDIACON W/IRON PEDIAFLOR DROPS PEDIALYTE PEDI AMYC IN PEDIAQUIL PEDIATRIC COUGH SYRUP PEDI ATROL PEDIATROL W/ IRON PEDI AZOLE PEDI ERAN W/ IRON PEDICULICIDE PEDRIC PEECE PEGANONE PEKT AMALT PEMOLINE PEN A/N PEN-VEE K PENALATE ELIXIR PENAMP PENAPAR VK PENBRITIN ‘PENICILL AMINE PENICILLIN PENICILLIN G PENICILLIN PROCAINE PENICILLIN V PENICILLIN VK PENS YN PENTACRESOL ORAL PENT AERYTHRITOL TETRANIT-PB PENTAERYTHRITOL PENTAFIN PENTAFORT PENT ATHYN PENT AZINE PENT AZINE EXPECTORANT PENTAZINE EXP W/CODEINE PENT AZOC INE PENT AZYME MEDICATION CODE LIST, NAMCS 1980 23295 23300 23305 23310 23315 PENTE TRA PENTHRANE PENTIDS PENTOBARBITAL PENTOTHAL 23320 PENTRASPAN-80 23325 23330 23335 23340 23345 23350 23355 23360 23365 23370 23375 23380 23385 23390 23395 23400 23405 23410 23415 23420 23425 23430 23435 23440 23445 23450 23455 23460 23465 23470 23475 23480 23485 23490 23495 23500 23505 23510 23515 23520 23523 23525 23530 23535 23540 23545 23550 23555 23560 23565 PENTRAX PENTRITOL PENTYLAN PEPPERMINT OIL PEPPERMINT SPIRIT PEPPERMINT WATER PEPSIN PEPTALAC ELIXIR PEP TAVLON PEPTO-BISMOL PERCHLORACAP PERCOBARB PERCOCET-5 PERCODAN PERCODAN-DEMI PERCOGE SIC PERCORTEN ACETATE PERCORTEN PIVALATE PERDIEM GRANULE PERESTAN PERGONAL PERI-COLACE PER I-DO SS PER IACTIN PERIDIAL PERIDIN-C PER IES PERIFOAM PER IHAB PER IHEMIN PERIOCAL-D PERITINIC PERITONEAL DIALYSIS W/DEXTROSE PERITRATE PERITRATE W/PHENOBARBITAL PERMAPEN AQUEOUS SUSPENSION PERMITIL PERNAEMON PERNAVIT PERNOX PERPHENAZINE PER SA-GEL P ER SADOX PERSANT INE PERSISTIN PERTINEX PERTOFRANE PER TROP IN PERTUSS IN PETN 23570 23575 23580 23585 PETRO-PHYLIC SOAP PETRO-SYLLIUM PETROGA LAR PETROLATUM 23590 23595 23600 23605 23610 23615 23620 23625 23630 23635 23640 23645 23650 23655 23660 23665 23670 23675 23680 PETROLEUM ETHER PETROZIN PFIZER-E PFIZERPEN PHAZYME PHAZYME PB PHAZYME-95 PHE-MER-NITE PREOP TINC PHEDRAL PHEL ANT IN PHEMIT HYN PHEN-DIMETRAZ INE PHENACETIN PHENACETOPHEN PHENAPHEN PHENAPHEN NO. 2 PHENAPHEN NO. 3 PHENAPHEN NO. 4 PHENAPHEN W/CODE INE 23685 PHENASEPTIC MOUTHWASH 23690 23695 23700 23705 23710 23715 23720 23725 23730 23735 23740 23745 PHENATANE EXPECTORANT PHENATAPP ELIXIR PHENATE PHENAZ INE PHENAZODINE PHENAZOPYRIDINE PHENAZOPYRIDINE-SULFISOXAZOLE PHENCASET (IMPROVED) PHENCEN-50 PHENCOID PHENDIMEAD PHENDIMETRAZINE 23750 23753 22755 23760 23765 23770 e31758 23780 23785 23790 23795 23800 23805 23810 23815 23820 23825 23830 23833 23835 23840 23845 23850 23855 23860 23865 23870 23875 PHENEEN PHENEL ZINE PHENERGAN PHENERGAN COMPOUND PHENERGAN EXPECTORANT PLAIN PHENERGAN EXPECTORANT W/CODE INE PHENERGAN EXP W/DM PHENERGAN SYRUP PHENERGAN VC EXPECTORANT PLAIN PHENERGAN VC EXP W/COD PHENERGAN-D PHENETRON PHENETRON COMPOUND PHENETRON LANACAP PHENETRON LANATAB PHENETRON SYRUP PHENHIST DH W/CODEINE PHENISTIX PHENMETRAZ INE PHENO NUX PHENO-SQU AR PHENOBARBITAL PHENOBARBITAL & ATROPINE PHENOBARBITAL & BELL ADONNA PHENOBARBITAL & HEXE STROL PHENOBARBITAL & HYOSCYAMUS PHENOBARBITAL ELIXIR PHENOBARBITAL GREEN 6S 23880 PHENOBARBITAL LAVENDER 23885 PHENOBARBITAL ORANGE 23890 PHENOBARBITAL PINK 23895 PHENOBARBITAL S.C. 23900 PHENOBARBITAL SCORED 23905 PHENOBARBITAL SODIUM 23910 PHENOBARBITAL WHITE 23915 PHENOBEL 2392) PHENOBELLA 23925 PHENODYNE 23930 PHENOJECT-50 23935 PHENOL 23940 PHENOL AX WAFER 23945 PHENOLPHTHALEIN 23950 PHENOLSULFONPHTHALEIN 23955 PHENOPTIC 23960 PHENOXENE 23965 PHENSPRAY 23970 PHENTERMINE 23975 PHENTROL 23980 PHENURONE 23985 PHENYL SALICYLATE 23990 PHENYL-APAP 23995 PHENYLALANINE 24000 PHENYLAZO 24005 PHENYLBUATZONE 24010 PHENYLBUTAZONE ALKA 24015 PHENYLEPHRINE 24020 PHENYLIN 24025 PHENYL PROPANOL AMINE 24030 PHENYLPROPANOLAMINE W/ CAFFEINE 24035 PHENYLZIN 24040 PHENYLZONE-A 24045 PHENYTQIN 24050 PHENZINE 24055 PHILJECT 24060 PHILLIPS OINTMENT 24065 PHISOAC 24070 PHISODAN 24075 PHISODERM 24080 PHISGCHEX 24085 PHISOSCRUB 24090 PHOS—-CAL W/VITAMIN D 24095 PHOS-CAL W/VITAMIN D & IRON 24100 PHOS-FLUR 24105 PHOS -PHAID 24110 PHOSPHACAL-D 24115 PHOSPHALJEL 24120 PHOSPHATE ENEMA 24125 PHOS PHO-SODA 24130 PHOS PHOL INE 24135 PHOSPHORIC ACID 24140 PHRENILIN 24145 PHYLDROX 24150 PHYLLOCONTIN 24155 PHYSIOSOL IRRIGATION 24160 PHYSOSTIGMINE 24165 PHYSOSTIGMINE SUL FATE 24170 PHYSPAN 24175 PHYTONADIONE MEDICATION CODE LIST, NAMCS 1980 24180 PIL-DIGIS 24185 PILO 24190 PILOCAR 24195 PILOCARPINE 24200 PILOCARPINE NITRATE 24205 PILOCEL 24210 PILOMIOTIN 24215 PILOPTIC 24220 PIMA SYRUP 24225 PIPERAZINE 24230 PIRACAPS 24235 PIRIN PLUS 24240 PIRIN 24245 PIRSEAL 24250 PISEC GEL 24255 PITOCIN 24260 PITRESSIN 24265 PITUITARY POSTERIOR 24270 PITUITRIN 24275 PIZ BUIN 24280 PLACEBO 24285 PLACIDYL 24290 PLAGUE VACCINE 24295 PLANTAMUCIN GRANULE 24300 PLAQUENIL 24305 PLASBUMIN 24310 PLASMA-LYTE 24315 PLASMA-PLEX 24320 PLASMANATE 24325 PLASMATEIN 24328 PLASMODICIDE 24330 PLATINOL 24335 PLEBEX 24340 PLEGINE 24345 PLEXONAL 24350 PMB 24355 PNEUMOVAX 24360 PNU IMUNE 24365 PODOPHYLLIN 24370 PODOPHYLLUM 24375 POINT-TWO 24380 POISON ANTIDOTE KIT 24385 POISON IVY 24390 POISON IVY EXTRACT 24395 POISON IVY SPRAY 24400 POLARAMINE 24405 OPV 24405 POLICMYELITIS VACCINE 24410 POLLEN ANTIGEN 24415 POLY-VI-FLOR 24420 POLY-VI-SOL 24425 POLYBACIN 24430 POLYBRADE GEL 24435 POLYCILLIN 24440 POLYCILLIN-PRB 24445 POLYCITRA 24450 POLYCOSE 24455 POL YDINE 24460 POL YMAGMA 24465 POL YMOX 24470 24475 24480 24485 24490 24495 24500 24505 24510 24515 24520 24525 24530 24535 24540 24545 24550 24555 24560 24565 24570 24575 24580 24585 24590 24595 24600 24605 24610 24615 24620 24625 24630 24635 24640 24645 24650 24655 24660 24665 24670 24675 24680 24685 24690 24695 24700 24705 POLYMYX IN POLYONIC POLYSAL POLYSORB POLYSORBIN POLYSORBIN DROPS POLYSORB IN-F DROPS POLYSPECTRIN POLYSPORIN POLYTAR POLYTRACIN POLYVITAMIN POLYVITAMIN DROPS POLYVITAMIN FLUORIDE POLYVITAMIN FLUORIDE DROPS PONDIMIN PONSTEL PONTOCA INE PONTOCAINE HCL PONTOCAINE HCL NIPHANOID PONTOCAINE W/DEXTROSE PORTAGEN POSTACNE LOTION POSTER ISAN POT CHLOR POTABA POTABA ENVULE POTABA POWDER POTABA+6 POTASALAN ELIXIR POTASSIUM & IODINE POTASSIUM ACETATE POTASSIUM BICARBONATE POTASSIUM BITARTRATE POTASSIUM BROMIDE POTASSIUM CARBONATE POTASSIUM POTASSIUM CHLORIDE & SOD CHLOR POTASSIUM CHLORIDE MEQ D5-W POTASSIUM CHLORIDE W/DEXTROSE POTASSIUM CHLORIDE 20 MEQ D5-W POTASSIUM CHLORI 30M ~ POTASSIUM CHLORIDE 40 MEQ D5-W POTASSIUM CITRATE POTASSIUM ESTRONE SULFATE POTASSIUM GLUCONATE POTASSIUM GUAIACOLSULFONATE POTASSIUM HYDRCXIDE 24710 POTASSIUM 10DIDE 24715 POTASSIUM NITRATE 24720 POTASSIUM PERCHLORATE 24725 POTASSIUM PERMANGANATE 24730 POTASSIUM PHOSPHATE 24735 POTASSIUM SODIUM TARTRATE AR 24740 POTASSIUM THIOCYANATE 24745 POTASSIUM TRIPLEX 24750 POTASSIUM 99 24755 POVAN 24760 POVIDINE 24765 POVIDONE 09 24770 24775 24780 24785 24790 24795 24800 24805 24810 24815 24820 24825 24830 24835 24840 24845 24850 24855 24860 24865 24870 24875 24880 24885 24890 24895 24900 24905 24910 24915 24920 24925 24930 24935 24940 24945 24950 24955 24960 24965 24970 24975 24980 24985 24990 24995 25000 25005 25010 25015 25020 25025 25030 25035 25040 25045 25050 25055 25060 25065 ININ POXY COMPOUND-65 POYAMIN PRAG MATAR PRAMET FA PRAMILET FA PRAMOS ONE PRANTAL PRAZOSIN PRE-DEP PRE-ENTHUS F PRE- INE PRE-MENS FORTE PRE-NATAL VITAMINS PRE-PEN PRE-SATE PRE~-SERT PRED FORTE PRED MILD PRED-5 PREDALONE R.P. PRED ANEX PREDCOR PREDNI CEN-M PREDNISOLONE PREDNISONE PREDOX INE PREDULOSE PREFLEX PREF RIN PREG ENT PREGESTIMIL PREGNYL PREL AN PRELAN F.A. PRELESTRIN PRELUDIN PREMARIN PREMARIN INTRAVENOUS PREMARIN VAGINAL PREMARIN W/METHYLTESTOSTERONE PRENABEX PRENATAL FORMULA (VITAMINS) PRENATAL STUART PRENATAL W/ FOLIC ACID PRENATAMIN PREPARATION H PREPODYNE PREPRO PREPTIC PRESALIN PRES AMINE PRES UN PRETTS PRIMAQUINE PRIMATENE MIST PRI MAT ENE-M PRIMATENE-P PRIMIDONE PRIMOLINE PRINCI PEN MEDICATION CODE LIST, NAMCS 1980 25070 25075 25080 25085 25090 25095 25100 25105 25110 25115 25 120 25125 25130 25135 25140 25145 25150 25155 25160 25165 25170 25175 25180 25185 25190 25195 25200 25205 25210 25215 25220 25225 25230 25235 25240 25245 25250 25255 25260 25265 25270 25275 25280 25285 25290 25295 25300 25305 25310 25315 25320 25325 25330 25235 25340 25345 25350 25355 25360 25365 PR INC IPEN W/PROBENECID PRINCIPEN/N PRINN V.S. VAGINAL PRISCOL INE PRIVINE PRO- AMID PRO-BANTHINE PRO-BANTHINE W/DARTAL PRO-BANTHINE W/ PHENOBARBITAL PRO-P HEN PROAQUA PROBALAN PROBAMPACIN PROB ANA PROBEN-C PROBENCILIN PROBENECID PROBENECID W/COLCHICINE PROBENI MEAD PROBENIMEAD W/COLCHICINE PROBILAGOL PROBOLIK PROCAINAMIDE PROCAINE PROCAINE HCL W/EPINEPHRINE PROCAMIDE PROCAN SR PROCAPAN PROCARBAZ INE PROCHLOR ISO PROCHLORPERAZINE PROCO PROCOLIN PROCTALME PROCTOCORT PROCTODON PROCTOFOAM PROCTOFOAM-HC PROCUTE LOTION PRODERM PROFASI HP PROFILATE PROFLAVINE PROGELAN PROGESIC COMPOUND-65 PROGESIC PROGESTASERT PROGESTERONE PROGLYCEM PROGYNON PROKETAZINE PROKLAR PROLIXIN PROLOID PROMAJECT PROMAPAR PROMAZ PROMAZ INE PROME THAMEAD PROMETHAZ INE 25370 25375 25380 25385 25390 25395 25400 25405 25410 25415 25420 25425 25430 25435 25440 25445 25450 25455 25460 25465 25470 25475 25480 25485 25490 25495 25500 25505 25510 25515 25520 25525 25530 25540 25545 25550 25555 25560 25565 25570 255715 25580 25585 25590 25595 25600 25605 25610 25615 25625 25630 25635 25640 PROMETHAZINE COMPOUND PLAIN PROMETHAZINE COMPOUND W/CODE INE PROMETHAZINE EXPECTORANT PROMETHAZINE EXP PROMETHAZ INE EXP PROMETHAZINE HCL PROMETHAZINE HCL PROMETHAZINE HCL PROMETHAZINE HCL PROMETHAZINE HCL PROMETHAZINE HCL DM PED W/COD EXPECTORANT EXP W/COD SYRUP VC EXPECTORANT W/COD EXP W/DM PROMETHAZ INE SYRUP FORTIS PROMETHAZINE VC W/CODE INE PROMETHAZINE W/PHENYLEPHR INE PROMEX PROMEX W/CODEINE PRONEMIA PRONESTYL PROPADRINE PROPAHIST PROPANOLOL PROPANTHEL INE EXPECTORANT PROPANTHEL INE BROMIDE W/PB PROPARACAINE PROPHYLLIN PROPION GEL PROPLEX PROPOXYPHENE COMPOUND 65 PROPOXYPHENE PROPOXYPHENE HCL PROPOXYPHENE HCL PROPOXYPHENE HCL COMPOUND COMPOUND 65 W/AP.Ce PROPOXYPHENE W/ACETAMIN PROPOXYPHENE HCL W/APAP PROPOXYPHENE W/APAP PROPYLENE GLYCOL PROPYLPARABEN PROPYLTHIOURACIL PROREX PROSOBEE PROSTAPHLIN PROSTIGMIN PROSTIN PROT ABOLIN PROT AMINE PROT ENATE PROT ERNOL PROTHAZINE PROT OP AM PROV-U-SEP PROV-U-SEP FORTE PROV AL PROVERA 5645 PROVIGAN 25650 25660 25665 25670 25675 25680 PROXAGESIC PROXENE PROXENE COMPOUND PROX IGEL PROZEX PROZ INE 50 19 25685 PRULET 25690 PRUNICODEINE 25695 PSEUDOEPHEDRINE 25700 PSORELIEF 25705 PSORIASIS CREAM 25710 PSORIGEL 25715 PSP-1V 25720 PSYCHOZINE 25725 PSYLLIUM SEED BLONDE 25730 PURETANE 25735 PURETANE EXPECTORANT 25740 PURETANE EXPECTORANT D.C. 25745 PURETAPP 25748 PURGATIVE 25750 PURINETHOL 25755 PURODIGIN 25760 PURPOSE 25765 PVPI- 25770 PYLORA #1 25775 PYOCIDIN 25780 PYOPEN 25785 PYRACORT 25790 PYRADYNE 25795 PYRADYNE COMPOUND 25800 PYRAZINAMIDE 25805 PYRIDIATE 25840 PYRI NAL 25845 PYRINYL 25850 PYRIZINE 25855 PYRODI NE 25860 PYROGALLOL REAGENT CRYSTAL 25865 PYROXINE 25870 PYRRALAN EXPECTORANT DM 25875 PYRROXATE 25880 PYRROXATE W/CODE INE PHOSPHATE 25885 PIE 25890 P2EL 25895 P3E 25900 P4EL 25905 P6EL 25910 QUAALUDE 25915 QUADETTS 25920 QUADRA HIST 25925 QUADRAMOID 25930 QUADRINAL 25935 QUADSUL 25940 QUAN-III 25945 QUARZAN 25950 QUELICIN 25955 QUEL IDRINE 25960 QUELTUSS 25965 QUES TRAN 25970 _QUIBRON 25975 __QUIBRON BIDCAP MEDICATION CODE LIST, NAMCS 1980 25980 25985 25990 25995 26000 26005 26010 26015 26020 26025 26030 26035 26040 26045 26050 26055 26060 26065 26070 26075 26080 26085 26090 26095 26100 QUIBRON ELIXIR QUIBRON PLUS QUIBRON PLUS ELIXIR QUIBRON-300 QUICK TANNING QUIDE QUIESS QUINAGL UTE QUINAMM QUINE QUINIDEX EXTENTAB QUINIDINE QUINIDINE SULFATE QUININE SULFATE QUINITE QUINOLOR COMPOUND QUINORA QUINSONE QUINTESS QUOTANE REC SPRAY R-GENE 10 RABIES VACCINE RAC EPHEDRINE RACET 26105 RAGWEED & RELATED POLLENS ALLERG 26110 26115 26120 26125 26130 26135 26140 26145 26150 26155 26160 26165 26170 26175 26180 26185 26190 26195 26200 26205 26210 26215 26220 26225 26230 26235 26240 26245 26250 26255 26260 26265 26270 26275 RAMSES "10 HOUR™ VAGINAL RANTEX CLOTH WIPE RASPBERRY SYRUP RATIO RAU-SED RAUDIXIN RAULOYDIN RAUPOID RAURINE RAUSERPA RAUSERP IN RAUTENSIN RAUTRAX RAUVAL RAUWILOID RAUWOLFIA RAUZIDE RBC PLUS RECOUP RECTACORT RECTAL MEDICONE RECTAL MEDICONE-HC RECTAL OINTMENT RECTALAD ENEMA RECTALAD-AMINOPHYLLI NE RECTOID REDISOL REDUCETS REGITINE REGLAN REGONOL REGROTON REGUL-AID SYRUP REGUL-AIDS 26280 26285 26290 26295 REGULOID REGUTOL REIDACOL REIDAMINE 26300 26305 26310 26315 26320 26325 26330 26335 RELA RELAXADON RELEFACT TRH REMSED RENACIDIN RENALGIN RENELATE RENESE 26340 RENESE-R 26345 26350 26355 26360 26365 26370 26375 26380 26385 26390 26395 26400 26405 26410 26415 26420 26425 26430 26435 26440 26445 26450 26455 26460 26465 26470 26475 26480 26485 26490 26495 26500 26505 26510 26515 26520 26525 26530 26535 26540 26545 26550 26555 26560 26565 26570 26575 RENO-M RENOGRAF IN RENCQUID RENOVIST RENOVUE REP ESTRA REP-PRED REPBIMONE REPEN-VK REPESTROGEN REPOISE REPOST ERONE #2 REPRO COMPOUND-65 REPSTRONE RESECT ISOL RESERPATABS RESERPINE RESERPOID RESORC INOL RESPINOL-G RESP IROL RESPITAL RESULIN RETET RETICULEX RETICULOGEN RET IN-A REVAC SUPPRETTE REXOLATE REZAMID RHEABAN RHEI-MINT ELIXIR RHEOMACRODEX IN DEXTROSE RHINAFED RHINALL RHINALL LONG ACTING SPRAY RHINALL VAPORISE RHINALL-10 RHINATE RHINDECON RHINEX RHINEX D-LAY RHINEX DM RHINICOMP JR. RHINIDRIN RHINIHAB JR. RHINITIS 29 26580 RHINOCAPS 26585 RHINOGESIC 26590 RHINOGESIC GG 26595 RHINOGESIC JUNIOR 26600 RHINOSYN 26605 RHINOSYN PEDIATRIC SYRUP 26610 RHINOSYN SYRUP 26615 RHINOSYN-DM PEDIATRIC SYRUP 26620 RHINOSYN-X SYRUP 26625 RHINSPEC 26630 RHODAVITE 26635 RHUBARB & SODA MIXTURE 26640 RHULICAINE 26645 RHULICREAM 26650 RHULIGEL GEL 26655 RHULIHIST LOTION 26660 RHULISPRAY 26665 RHUS TOX ANTIGEN 26670 RHUSTICON LOTION 26675 RHUTOX LOTION 26680 RIASOL LOTION 26685 RIBOFLAVIN 26690 RICOR 26695 RID 26700 RIDP-A-COL 26705 RIFADIN 26710 RIFAMATE 26715 RIFAMPIN 26720 RIMACTANE 26725 RIMACTANE/INH DUAL PACK 26730 RIMSO0-50 26735 RINGER'S 26740 RIOPAN 26745 RIOPAN CHEW 26750 RIOPAN PLUS 26755 RIOPAN SWALLOW 26760 RITALIN 26765 RO-BILE 26770 ROBALATE 26775 ROBAMOX 26780 ROBATHOL BATH OIL 26785 ROBAXIN 26790 ROBAXISAL 26795 ROBIEILLIN VK 26800 ROBIMYCIN 26805 ROBINUL 26810 ROBINUL FORTE 8 B = 26820 ROBINUL-PH FORTE 26825 ROBITET 26830 ROBITUSSIN 26835 ROBITUSSI —~C SYRU 26840 ROBI JUSSIN-CF 26845 ROBITUSSIN-CF SYRUP 26850 ROBITUSS IN-DAC SYRUP 26855 ROBITUSSIN-DM COUGH CALMERS 26860 ROBITUSSIN-DM SYRUP 26865 ROBITUSSIN-PE SYRUP 26870 ROCALTROL 26875 ROCHELLE SALT MEDICATION CODE LIST, NAMCS 1980 26880 ROERIBEC 26885 ROGENIC 26890 ROLAIDS 26895 ROLOX 26900 ROMEX DECONGEST COUGH & COLD 26905 ROMILAR 26910 ROMILAR CHILDREN'S COUGH SYRUP 26915 ROMILAR COUGH & COLD 26920 ROMILAR IIt 26925 ROMYCIN 250 26930 RONDEC 26935 RONDEC-DM 26940 RONDGCMYCIN 26945 RONIACOL 26950 ROPANTH 26955 ROSE WATER 26960 RP-MYCIN 26965 RU-A-DRON 26970 RU-ANDROSPAN 26975 RU-CORT 26980 RU-CORT SPAN 26985 RU-EST 26990 RU-EST-SPAN 26995 RU-EST-TEST-SPAN 27000 RU-HEM 27005 RU-HIST 27010 RU-HY-T 27015 RU-K-N 27020 RU-LOR 27025 RU-LOR-N 27030 RU-PHEN 27035 RU-SPAS NO. 2 27040 RU-THIDSAL 27045 RU-TUSS 27050 RU-TUSS EXPECTORANT 27055 RU-TUSS PLAIN 27060 RU-TUSS W/HYDROCODONE 27065 RU-VERT 27070 RUBBING ALCOHOL 27075 RUBELLA VIRUS VACCINE LIVE 27080 RUBESOL 27085 RUBRAMIN PC 27090 RUBRAPLEX 27095 RUBRAVITE 27100 RUC-DANE 27105 RUFOLEX 27110 RUGAR 27115 RUHEXATAL W/RESERPINE 27120 RWOX 27125 RUM-K SYRUP 15% 27130 RUTIN 27135 RUTIN & ASCORBIC ACID 27140 RUTIN W/C 27145 RUVITE 27150 RV PABA STICK 27155 RVP 27160 RVPAQUE 27165 RVWPLUS 27170 RYNA 27175 RYNA-C 27180 RYNA-CX 27185 RYNA-TUSSADINE EXPECTORANT 27190 RYNATAN 27195 RYNATUSS 27200 RYNEX 27205 RYTHROSITE 27210 S.A.S.-500 27215 S.B.P. PLUS 27220 S.0.S. FIRST AID SPRAY 27225 S.T. 37 27230 S-A-C 27235 S-8-C 27240 S-B-T 27245 S-M-A IMPROVED 27250 S-P-1 27255 S—PAINACET 27260 S-250 GRANUCAP 27265 SACCHARIN 27270 SAFFLO-BEE 27275 SAFFLOWER OIL 27280 SAFFLOWER OIL W/B-6 27285 SAL DEX 27290 SAL HEPATICA 27295 SAL-FAYNE 27300 SALAGEN 27305 SALATAR 27310 SALATIN 27315 SALATIN W/CODEINE 27320 SALETO 27325 SALETO-D 27330 SALICRESIN FLUID 27335 SALICYLAMIDE 27340 SALICYLATE 27345 SALICYLIC ACID 27350 SALICYLIC ACID & SULFUR SOAP 27355 SALICYLIC ACID SOAP 27360 SALIGEL 27365 SALIMEPH FORTE 27368 SALINE 27370 SALINE INDUCTION KIT 27375 SALDCOL 27380 SALCL & BISMUTH COMPOUND W/OPIUM 27385 SALOMENTH 27390 SALPABA W/VITAMIN C 27395 SALPHENINE 27400 SALPHENYL 27405 SALPIX 27410 SALTPETER 27415 SALURON 27420 SALUTENSIN 27425 SANDRIL 27430 SANMETTO 27435 SANOREX 27445 SANTYL 27450 SARAKA GRANULE 27455 SARATOGA 27460 SARO-C 27465 SAROCYCLINE 27470 SARODANT £9 27475 SAROFLEX 27480 SAROLAX 27485 SARONIL 27490 SAROPEN VK 27495 SASTID 27500 SAVACORT-D 27505 SAVAPLEX 27510 SAXIN 27513 SCABICIDE 27515 SCADAN SCALP LOTION 27520 SCARLET RED 27525 SCHAMBERG LOTION 27530 SCLAVOTEST PPD 27535 SCODONNAR 27540 SCOPODEX PELLET 27545 SCOPOLAMINE 27550 SCOPOLAMINE HYDROBROM IDE 27555 SCOPOLAMINE MUROCOLL NO. 19 27560 SCOTT'S EMULSION 27565 SCRIP SPRAY 27570 SCRIP STICK 27575 SCRIP ZINC COMPOUND 27580 SCRIP-GESIC 27585 SCRIP-LAX 27590 SEBA-NIL 27595 SEBACI DE 27600 SEBASORB LOTION 27605 SEBASUM CLEANSER 27610 SEBAVEEN SHAMPOO 27615 SEBEX SHAMPOO 27620 SEBICAL CREAM SHAMPOO 27625 SEBIZON LOTION 27630 SEBUCARE LOTION 27635 SEBULEX 27640 SEBUTONE 27645 SECOBARBITAL 27650 SECONAL 27655 SECOPHEN 27660 SECRAN 27665 SECRAN PRENATAL 27670 SECRETIN-BOOTS 27675 SED-TENS SE 27680 SEDABAMATE 27685 SEDADROPS 27690 SEDADYNE 27695 SEDAMYL 27700 SEDAPAP ELIXIR 27705 SEDATANS 27708 SEDATIVE 27710 SEDATUSSIN 27715 SEDRALEX 27720 SELACRYN 27725 SELENIUM 27730 SELSUN 27735 SELSUN BLUE 27740 SEMETS TROCHE 27745 SEMICID 27750 SENILEX 27755 SENNA COMPOUND 27760 SENNA EXTRACT MEDICATION CODE LIST, NAMCS 1980 27765 SENNA LEAF FLUIDEXTRACT 27770 SENNA 27775 SENNA PODS, WHOLE 27780 SENOKAP DSS 27785 SENOKOT 27790 SENOKOT GRANULE 27795 SENOKOT S 27800 SENOKOT SYRUP 27805 SENOKOT W/PSYLLIUM 27810 SENDOLAX 27815 SENORAL-M ELIXIR 27820 SEPP ANTISEPTIC APPLICATOR 27825 SEPP SCRUB APPLICATOR 27830 SALOL & BISMUTH W/OPIUM 27835 SEPTRA 27840 SEPTRA DS 27845 SER—-AP-ES 27850 SERATHIDE 27855 SERAX 27860 SERENIUM 27865 SERENTIL 27870 SERFOLIA 27875 SEROMYCIN 27880 SERPALAN 27885 SERPANRAY 27890 SERPASIL 27895 SERPASIL ELIXIR 27900 SERPASIL-APRESOLINE 27905 SERPASIL-ESIDRIX 27910 SERPATE 27915 SERUTAN GRANULE 27920 SESAME OIL 27925 SETAMINE 27930 SHERAFED 27935 SHERAFED C EXPECTORANT 27940 SHERAFED SYRUP 27945 SHERITAL ELIXIR 27950 SIBLIN 27955 SIDONNA 27960 SIGTAB 27965 SILAIN-GEL 27970 SILENCE IS GOLDEN SYRUP 27975 SILICIC ACID REAGENT 27980 SILMAGEL 27985 SIL VADENE 27990 SILVER IODIDE 27995 SILVER NITRATE 28000 SILVER NITRATE TOUGHENED STICKS 28005 SILVER NITRATE WAX AMPUL 28010 SILVER PROTEIN MILD NF 28015 SILVER SULFADIAZINE 28020 SIMAAL GEL 28025 SIMECO 28030 SIMETHICONE 28035 SIMILAC 28040 SIMPLE SYRUP 28045 SIMRON 28050 SINACON 28055 SINAREST 28060 SINE-AID 28065 28070 28075 28080 28085 28090 S INE-OFF SINE-OFF SPRAY SINE-OFF W/APAP S INEMET S INEQUAN SINEX LONG ACTING SPRAY 28095 28100 28105 28110 28115 28120 28125 28130 28135 28140 28145 28150 28155 28160 28165 28170 28175 28180 28185 28190 28195 28200 28205 28210 28215 28220 28225 28230 28235 28240 28245 28250 28255 28260 28265 28270 28275 28280 28285 28290 28295 28300 28305 SINEXIN SINGLET SIND TUSS SINO-COMP SINOGRAFIN SINOPHEN SINDZE SINU-LETS SINUBID SINUGESIC SINULIN SINUNIL PELLSULE SINUS SINUSULE SINUTAB SINUTAB EXTRA STRENGTH SINUTAB LONG ACTING SINUS SPRAY SINUTAB W/CODE INE SINUTAB-II SINUTREX SIROIL EMULSION SK-AMITRIPTYLINE SK-AMPICILLIN SK-APAP SK—-APAP W/ CODEINE SK-BAMATE SK-BISACODYL SK-CHLORAL HYDRATE SK—-CHLOROTHIAZIDE SK-DEXAMETHASONE SK-DIGOXIN SK-DIPHENHYDRAMINE SK-DIPHENOXYLATE SK-ERYTHROMYCIN SK-HYDROCHLOROTHIAZIDE SK-LYGEN SK-NIACIN SK-PENICILLIN G SK-PENICILLIN VK SK—-PHENOBARBITAL SK-PRAMINE SK-PREDNISONE SK-QUINIDINE SULFATE 28310 28315 28320 28325 28330 28335 28340 28345 28350 28355 28358 SK-RESERPINE SK-SOXAZOLE SK-TETRACYCLINE SK-TOLBUTAMIDE SK-TRIAMCINOLONE SK-65 SK-65 APAP SK-65 COMPOUND SKELAXIN SKIM INFANT FORMULA SKIN PREPARATION 28360 SKIODAN SODIUM 28365 SLO-PHYLLIN 28370 SLO-PHYLLIN GG 8375 Y 28380 SLO-PHYLLIN SYRUP 28385 SLOAN'S LINIMENT 28390 SLOW-K 28395 SMALLPOX VACCINE 28400 SOACLENS 28405 SOAKARE SOAKING 8410 SOAP LINIMENT 28415 OAP_LIQUID 28420 SODA MINT 28425 SODANUX 28430 SODESTRIN 28435 SODIUM ACETATE 28440 SODIUM AMINOSALICYCLATE 28445 SODIUM ASCORBATE 28450 SODIUM BENZOATE 28455 SODIUM BICARBONATE 28460 SODIUM BIPHOSPHATE 28465 SODIUM BISULFATE 28470 SODIUM BISULFITE REAGENT 28475 SODIUM BORATE 28480 SODIUM EROMIDE 28485 SODIUM CACODYLATE 28490 SODIUM CARBONATE 28495 SODIUM CHLORIDE 28500 SODIUM CHLORIDE (BENZ ALC PRSVD) 28505 SODIUM CHLORIDE (PARABEN PRSVD) 28510 SODIUM CHLORIDE (PHYSIOL SALINE) 28515 SODIUM CHLORIDE & DEXTROSE 28520 SODIUM CHLORIDE BACTERIOSTATIC 28525 SODIUM CHLORIDE CONCENTRATE 28530 SODIUM CHLORIDE DISPOS—-A-VIAL 28535 SODIUM CHLORIDE ENSEAL 28540 SODIUM CHLORIDE FLIP TOP 28545 SODIUM CHLORIDE FOR IRRIGATION 28550 SODIUM CHLORIDE HAEMO-VAC 28555 SODIUM CHLORIDE INFLEX UNIT 28560 SODIUM CHLORIDE IRRIGATION 28565 SODIUM CHLORIDE MURO 128 28570 SODIUM CHLORIDE TEAR TOP 28575 SODIUM CITRATE 28580 SODIUM CITRATE W/BELLADONNA 28585 SODIUM DEHYDROCHOLATE 28590 SODIUM DICHROMATE 28595 SODIUM FLUORIDE 28600 SODIUM GLUTAMATE 28605 SODIUM HYDROXIDE 28610 SODIUM HYPOCHLORITE 28615 SODIUM INDIGOTINDISULFONATE 28620 SODIUM IODIDE 28625 SODIUM LACTATE 28630 SODIUM LAURYL SULFATE 28640 SODIUM MORRHUATE 28645 SODIUM NICOTINATE 28650 SODIUM NITRITE 28655 SODIUM NITROFERRICYANIDE AR MEDICATION CODE LIST, NAMCS 1980 28660 28665 28670 28675 28680 28685 28690 28695 28700 28705 28710 28715 28720 28725 28730 28735 28740 28745 28750 28755 28760 28765 28770 SODIUM PERBORATE SODIUM PHOSPHATE SODIUM PHOSPHATE & BIPHOSPHATE SODIUM SALICYLATE SODIUM SUCCINATE SODIUM SULFATE SODIUM SULFITE SODIUM THIOSALICYLATE SODIUM THIOSULFATE SODIUM VERSENATE SOFLENS CLEANER SOFNER SOFT THERM SOFT'N SOOTHE SOF TNER SOLAR SOLARCAINE SOLATENE SOLBAR SOLFOTON SOLGANAL SUSPENSION SOLU PREDALONE SOLU PREDCOR 28775 SOLU-B MIX-0-VIAL 28780 28785 28790 28795 28800 28805 28810 28815 28820 SOLU-B W/ASCORBIC ACID SOLU-B-FORTE W/TRANSFER NEEDLE SOLU-CORTEF SOLU-EST SOL U-MEDROL SOLUJECT SOL UREX SOMA COMPOUND SOMA COMPOUND W/CODEINE 28825 SOMA 28830 28835 28840_S 28845 28850 28855 28860 28865 28870 28875 28880 28885 28890 28895 28900 28905 28910 28915 28920 28925 SOMBULE X SOMLYN W/PHENOBARBITAL SOMOPHYLLIN SONAC IDE SONAZ INE SONIL YN SONIPHEN SOOTHE EYE SOO THOCA INE SOOTHOGEL SOPOR SOPRODOL SOPRODOL COMPOUND SOPRONOL SOQUETTE SORBIDE T.D. SORBITOL SORBITOL-MANNITOL IRRIGATION SORBITRATE SORETTS 28930 SORQUAD 28935 28940 28945 28950 28955 sosoL SOTRADECOL SOY BAKER'S READY-4 SOY-DOME CLEANSER SOY-SITZ COLLOID BATH 28960 28965 28970 28975 28980 28985 28990 28995 29000 29005 29010 29015 29020 29025 29030 29035 29038 29040 29045 29050 29055 29060 29065 29070 29075 29080 29085 29090 29095 29100 29105 29110 29115 29120 29125 29130 29135 29140 29145 29150 29155 29160 29165 29170 29175 29180 29185 29190 29195 29200 29205 29210 29215 29220 29225 29230 29235 29240 29245 SOYALOID COLLOID BATH SOYBEAN OIL SPAL IX SPAN NIACIN SPAN-EST-TEST SPANCAP C SPANCAP NO. 1 SPANESTRIN P SPANTAC SPANTROL SPANTUSS SPARINE SPARINE SYRUP SPASMATE SPASMOJECT SPASMOLIN SPASMOLYTIC AGENT SPASMOPHEN SPASQUID ELIXIR SPASTOLATE SPASTOSED SPD ANALGESIC SPEARMINT OIL SPEC-T ANESTHETIC SPECTRO-BIOTIC SPECTROCIN SPEEDRIN SPEN-COLD ADULT COUGH SYRUP SPEN-COLD IMPROVED SPEN-COLD PEDIATRIC COUGH SYRUP SPEN-COLD SPRAY SPEN-HISTINE DH SPEN-HISTINE ELIXIR SPEN-HISTINE EXPECTORANT SPEN-IRON DROPS SPEN-O-LETS M SPENAURAL DROPS SPENAX IN SPENBOL IC SPENCORT SPENCORT FORTIFIED SPENCORT H.P. SPENDEC DM SPENDEPIOL SPENDRIS IN SPENDUO SPENGINE SPENIACOL SPENOX SPENPATH SPENSOMIDE SPENTACID SPENT ANE SPENTANE DC EXPECTORANT SPENTANE EXPECTORANT SPENTAPP SPENZIDE SPIRITEX SPIRONAZ IDE 29250 SPIRONOLACTONE G9 29255 SPIRONOLACTONE W/HCLTHIAZIDE 29260 SPRX-1 29265 SPRX-2 29270 SPRX-3 29275 SSKI 29280 ST JOSEPH DROPS FOR CHILDREN 29285 STADOL 29290 STANBACK ANALGESIC 29295 STANDARD CHEW 29300 STANDARD DROPS 29305 STANDARD F CHEW 29310 STAPHAGE LYSATE ANTIGEN 29315 STAPHCILLIN 29320 STAPH TOXOID DIGEST MODIFIED 29325 STARCH 29330 STAT OBEX 29335 STATOMIN MALEATE 29340 STATROL 29345 STATUSS 29350 STAY-FLO 29355 STEARIC ACID 29360 STEARYL ALCOHOL usP 29365 STELAZINE 29370 STERA-FORM 29375 STERABEL 29380 STERACHOL 29385 STERADRIN 29390 STERAFED 29395 STERAFED SYRUP 29400 STERAFED-C EXPECTORANT 29405 STERAHEX 29410 STERAJECT-50 29415 STERAMAG #1 29420 STERAMINE 29425 STERANE 29430 STERAPHENATE 29435 STERAPRED 29440 STERAPRES 29445 STERASOL T.De. 29450 STERASOL INE 29455 STERASPASMOL 29460 STERATANE 29465 STERATAP PA 29470 STERAZIDE 29475 STERAZYME 29480 STERIBOLIC 29485 STERIDIUM 29490 STERINAL-200 29495 STEROFORM 29498 STEROID(S) 29500 STEROTATE 29505 STILBESTROL 29510 STILPHOSTROL 29515 STIM-TABS 29518 STIMULANT 29520 STIMULAX 29525 STOKES EXPECTORANT MIXTURE 29530 STOMASEPTINE 29533 STOOL SOFTENER 29535 STOP MEDICATION CODE LIST, NAMCS 1980 29540 29545 29550 29555 29560 29565 29570 29575 29580 29585 STR! 29590 29595 29600 29605 29610 29615 29620 29625 29630 29635 29640 29645 29650 29655 29660 29665 29670 29675 29680 29685 29690 29695 29700 29705 29710 29715 29720 29725 29730 29735 29740 29745 29750 29755 29760 29765 29770 29775 29780 29785 29790 29795 29800 29805 29810 29815 29820 29825 29830 29835 STOPAIN STOPIT STOPIT-25 STOXIL STREPTASE STREPTOMYCIN STRESS FORMULA STRESS- VITES STRESSCAPS STRESSTABS 600 STRESSVICON STRI-DEX PADS STRYCHNINE POWDER STUART FORMULA STUARTINIC STUARTNATAL 1+1 STUDAFL UCR STULEX SU-TINIC SU-TON SU-Z0L SUBL IMAZE SUBY'S SOLUTION G SUCAR YL SUCC INYLCHOLINE SUCCINYLSULFATHIAZOLE SUCOSTRIN SUCRETS SUCRETS CHILDREN'S SUCRETS COLD DECONGE STANT SUCRETS COUGH CONTROL SUCROSE SUDA-PROL SUDAFED SUDAFED PLUS SUDAFED PLUS SYRUP SUDAFED S.A. SUDAFED SYRUP SUDAHIST SUDDEN TAN BRONZING FOAM SUDO- 60 SUDOL IN SUDR IN SUFAMAL SUFEDR IN SUFEDRIN SYRUP SUL TRIG-MM NO. 2 SUL-BLUE SHAMPOO SULADYNE SULAMYD SODIUM SULDIAZO SULF-10 SULFA VAGINAL SULFACEL-15 SULFACET-R LOTION SULFACETAMID SULFACYTINE SULFADIAZINE SUL FALAR SULFALOID 29840 29845 29850 29855 29860 29865 29870 29875 29880 29885. 29890 29895 29898 29900 29905 29910 29915 29920 29925 29930 29935 29940 29945 29950 29955 29960 29965 29970 29915 SULFAMETHOXAZOLE SULFAMYLON SULFANILAMIDE SULFAPRED SULFAPYRIDINE SULFASALAZINE SULFASOX SULFASOXAZOLE SULFATHALIDINE SULFATHIAZOLE SULFAVITIN SULFEM SULFINPYRAZOLE SULFISOXAZOLE SULFISOXAZOLE W/PHENAZOPYR SULFIZ IN SULFO-LO SULFOBROMOPHTHAL EIN SULFONAMIDES DUPLEX SULFORCIN BASE SULFOSALICYLIC ACID SULFOSE SULFOXYL LOTION REGULAR SULFSTAT FORTE SULFUR SULFUR & RESORCIN COMPOUND SULFUR COLLOIDAL POWDER SULFUR FLOWERS SULFUR PRECIPITATED 29980 29985 29990 29995 29998 30000 30005 30010 30015 30020 30025 30030 30035 30040 30045 30050 30055 30060 30065 30070 30075 30080 30085 30090 30095 30100 30105 30110 30115 30120 30125 SULFUR SOAP SULFUR SUBLIMED SULFUR W/CREAM OF TARTAR SULFUR-8 CONDITIONER SULINDAC SULFURATED POTASH USP SULLA SULPHUR SULTRIN SUMMER'S EVE DISPOSABLE SUMOX SUMSCREEN SUMYCIN SUNBRELLA LOTION SUNDARE CLEAR LOTION SUNDOWN SUNGARD SUNRIL SUNSTICK LIP PROTECTANT SUNSWEPT SUPAC SUPEN SUPER AFKO-HIST SUPER ANAHIST SUPER ANAHIST SPRAY SUPER B COMP W/C LIV IRON & B-12 SUPER D PERLE SUPER DOSS SUPER MANIVIM SUPER POTENCY B-COMP W/C SUPER ULEX 99 30130 30135 30140 30145 30150 30155 30160 30165 30170 SUPERBEE SUPE RT AH SUPP AP SUPPORT SUPPRESSATE SURBEX SURBEX W/C SURBEX 750 W/ IRON SURBEX 750 W/ ZINC 30175 SURBEX-T 30180 SURBEX-T W/DEXTROSE 5% 30185 SURFACAINE 30190 SURFADIL 30195 SURF AK 30200 SURFOL BATH OIL 30205 SURG-C 30210 SURGI-SEP 30215 SURGICEL 30220 SURGILUBE 30225 _SURIN 30230 SURITAL 30235 SURMONTIL 30240 SUS=PHRINE AQUEOUS SUSPENSION 30245 SUSTACAL 30250 SUSTAGEN 30255 SUSTAIRE 30260 SUSTAVERINE 30265 SUX-CERT 30270 SWEDISH TANNING SECRET LOTION 30275 SWEET OIL 30280 SWEETA 30285 SWEETER 30290 SWIM-EAR 30295 SYLLACT 30300 SYLLAMALT 30305 SYMMETREL 30308 SYMPATHOLYTIC AGENT 30309 SYMPATHOMIMETC AGENT 30310 SYMPTOM 1 30315 SYMPTOM 2 30320 SYMPTOM 3 30325 SYMPTOMAX 30330 SYNALAR 30335 SYNALGOS 30340 SYNALGOS-DC 30345 SYNASAL 30350 SYNCELAX 30355 SYNCELAX-RS 30360 SYNCURINE 30365 SYNEMOL 30370 SYNKAYVITE 30375 SYNOPHYLATE 30380 SYNOPHYLATE-GG 30385 SYNTHALOIDS 30390 SYNTHETAR 30395 SYNTHROID 30400 SYNTOCINON 30405 SYNTROGEL 30410 SYRACOL 30415 SYTOBEX MEDICATION CODE LIST, NAMCS 1980 30420 30425 30430 30435 30440 30445 30450 30455 30460 30465 30470 30475 30480 30485 30490 30495 30500 30505 30510 30515 30520 30525 30530 30535 30540 30545 30550 30553 30555 30560 30565 30570 30575 30580 30585 30590 T CAINE TeD. THERALS GRANUCAP T.E. IONATE P.A. TeEePe TeHePe. TeTePe T=E=P T-I1-GAMMAGEE T-IONATE-P.A. T-PLEX 7-250 TABRON TACARYL TACE TAGAFED TAGAMET TAGATAP TAKA-COMBEX TAKA-DIASTA SE TALC TALOIN TALTAPP TAL WIN COMPOUND TAL WIN TALWIN LACTATE TAMINE TAMINE DC TAMOXIFEN TANDEARIL TANICA INE TANNIC ACID TANUROL TAO TAP AR TAPAZOLE TAR 30595 30600 30605 30610 30615 30620 30625 30630 30635 30640 30645 30650 30655 30660 30665 30670 30675 30680 30685 TARACTAN TARBONTIS TARCORTIN TARPASTE TARTARIC ACID TA SHAN TAUROPHYLLIN TAVIST TA XOL TAYSTRON TEAR-EFRIN TEARISOL TEARS NATURALE TEARS PLUS TEDFERN TEDRAL TEDRAL ANTI-H TEDRAL ELIXIR TEDRAL EXPECTORANT 30690 30695 30700 30705 30710 TEDRAL SA TEDRAL-25 TEEBACIN TEEBACONIN TEEBACONIN W/VITAMIN B-6 30715 TEENAC 30720 30725 30730 30735 30740 30745 30750 30755 30760 30765 30770 30775 30780 30785 30790 30795 30800 30805 30810 30815 30820 30825 30830 30835 30840 30845 30850 30855 30860 30865 30870 30875 30880 30885 30890 30895 30900 30905 30910 30915 30920 30925 30930 30935 30940 30945 30950 30955 30960 30965 30970 30975 30980 30985 30990 30995 31000 TEEV TEGOPEN TEGRETOL TEGRIN TELDRIN TELEPAQUE TEM-PAIN TEMARIL TEMPRA TEMPTEE-CEE TEN-SHUN TENAX TENOL TENSILON TENUATE TEPANIL TERAMINE TERBUT ALINE TERFONYL TERPHAN EL IXIR TERPIN HYDRATE TERPIN HYDRATE & CODEINE ELIXIR TERPIN HYDRATE & DM TERPIN HYDRATE ELIXIR TERPIN HYDRATE W/DM TERRA-CORTRIL TERRAMYCIN TERRAMYCIN W/POLYMYXIN B SULFATE TERSA-TAR TERSASEPTIC TES-TAPE TESLAC TESSALON PERLE TESTA ESTRA C TESTA-C TESTADIATE DEPO TESTAQUA TESTATE TESTOJECT TESTOSTERONE TESTOSTERONE CYPIONATE W/ESTRAD TESTOSTERONE DEPOT TESTOSTERONE ENANTHATE TESTOSTERONE ENANTHATE W/ESTRAD TESTOSTERONE IN OIL TESTOSTERONE METHYL TESTOSTERONE PROPIONATE TESTOSTERONE SUSPENSION TESTOSTERONE-ESTRONE TESTOSTERONE-100 TESTOSTERONE-25 TESTOSTROVAL P.A. TESTRED TESTRINE TET 250 TET-CONN-G TETANUS ANTITOXIN 31005 31010 TETANUS DIPHTHERIA TOXOID TETANUS IMMUNE GLOBULIN L9 MEDICATION CODE LIST, NAMCS 1980 015 TETAN TOX0 31315 THERABID 31600 THYMOL 31020 TETRA 31320 THERACEBRIN 31605 THYMOL IODIDE 31025 TETRACAIN 31325 THERAGRAN 31610 THYPINONE 31030 TETRACAINE HCL 31330 THERAGRAN HEMATINIC 31615 THYRAR 31035 TETRACHEL 31335 THERAGRAN-M 31620 THYRO-TERIC 31040 TETRACON 31340 THERAGRAN-Z 31625 THYROGLOBULIN 31045 TETRACYCLINE 31345 THERALAX 31630 THYROID 31050 TETRACYCLINE HCL 31350 THERALETS 31635 THYROLAR 31055 TETRACYN 31355 THERALS T.De. 31640 THYTROPAR 31060 TETRALAN 31360 THERAMEAD 31645 TICAR 31065 TETRAMINE 31365 THERAPADS 31650 TICARCILLIN 31070 TETRASTATIN 31370 THERAPAV 31655 TIGAN 31075 TETREX 31375 THERAPEUTIC B COMP W/VIT C 31658 TIMOLOL 31080 TEX SIX Te.Re 31380 THERAPEUTIC FORMULA VITAMIN 31660 TIMOPTIC 31085 TEXACORT 31385 THERAPEUTIC LIQUID 31665 TIMOTHY & REL POLLENS ALLERG 31090 THALFED 31390 THERAPEUTIC MULTIVITAMIN 31670 TINACTIN 31095 THAM 31395 THERAPEUTIC VITAMIN 31675 TINDAL 31100 THAM-E 31400 THERAPEUTIC VITAMIN & MINERAL - 31680 TING 31105 THANTIS 31405 THERAVILAN 31685 TING SOAP 31110 THEELIN AQUEOUS SUSPENSION 31410 THERAVIT HEMATINIC 31690 TING SPRAY 31115 THEO-COL ELIXIR 31415 THEREVAC 31695 TINWER LOTION 31120 THEO-DUR 31420 THERMOTABS 31700 TIREND 31125 THEO-GUAIA 31425 THERON 31705 TIS-U-SOL 31130 THEO-LIX 31430 THI-CIN 31710 71SIT 31135 THEO-NAR 31435 THI-CO-LIX ELIXIR 31715 TITAN 31140 THEO-ORGANIDIN 31440 THIACIDE 31720 TITRALAC 31145 THEOBID 31445 THIAHEP 31725 TOBRAMYCIN 31150 THEOBROMINE 31450 THIAMINE ELIXIR 31730 TOCOPHER 31155 THEOBROMINE SODIUM SALICYLATE 31455 THIAMINE 31735 TOCOPHERYL 31160 THEDCAP 31460 THIAMINE HCL ELIXIR 31740 TOFRANIL 31165 THEOCLEAR 31465 THIAMINE-PYRIDOXINE 31745 TOKOLS 31170 THEOCLIMAN 31470 THIAPHYLL 31750 TOLAZOLINE 31175 THEODRINE 31475 THIASERP 31755 TOLBUTAMIDE 31180 THEOFED 31480 THIMEROSAL 31760 TOLECTIN 31185 THEOFEDRAL 31485 THIODYNE 31765 TOLERON 31190 THEOFENAL 31490 THIOGUANINE 31770 TOLFRINIC 31195 THEOKIN 31495 THIOLATE 31775 TOLINASE 31200 THEOLAIR 31500 THIOMER IN 31780 TOLMETIN 31205 THEOLATE 31505 THIOPENTAL 31785 TOLU-SED 31210 THEOLIXIR 31508 THIORIDAZINE 31790 TONACON 31215 THEOPHED 31510 THICSAL 31795 TONAVITE M 31220 THEOPHOZINE 31515 THIOSULFIL 31800 TONEBEC 31225 THEOPHYL 31520 THIOSULFIL DUD-PAK 31805 TONECOL 31230 THEOPHYLLIN ELIXIR 31525 THIOSULFIL FORTE 31810 TONELAX 31235 THEOPHYLLINE 31530 THIOSULFIL-A 31815 TONESTAT 31240 THEOPHYLLINE COMPOUND 31535 THIOSULFIL-A FORTE 31820 TONDO B WAFER 31245 THEOPHYLLINE ELIXIR 31540 THIOTEPA 31825 TOPIC 31250 THEOPHYLLINE KI ELIXIR 31545 THIURETIC 31830 TOPICORT EMOLL IENT 31255 THEOPHYLLINE T.D. 31550 THORAZINE 31835 TOPICYCLINE 31260 THEOPHYLLINE-EPHEDRINE-PB 31555 THREACON EXPEC TORANT 31840 TOPSYN GEL 31265 THEOSPAN 31560 THREAMINE 31845 TORA 31270 THEOTABS 31565 THREE BROMIDES ELIXIR 31850 TORECAN 3 5 THEOQV 31570 THREONINE 31855 TOROCOL 31280 THEOZINE 31575 THRIOCAINE 31860 TOSSECOL 31285 THERA SPANCAP 31580 THROAT DISC 31865 TOTA-VI-CAPS 31290 THERA-AMCAPS 31585 THROAT LOZENGE NEO-VADRIN 31870 TCTACILLIN 31295 THERA-COMBEX 31588 THROAT PREPARATION 31875 TOTAL 31300 THERA-FLUR GEL-DROPS 31590 THROMBIN TOPICAL 31880 TOTAL B W/C 31305 THERA-9 31595 THROMBOLYSIN 31310 THERABEX 31598 THROMBOLYTIC AGENT 31890 TOTAMIN 89 31895 31900 31905 31910 31915 31920 31925 31930 31935 31940 31945 31950 31955 31960 31965 31970 31975 31980 31985 31990 31995 32000 32005 32010 32015 32020 32025 32030 32035 32040 32045 32050 32055 32060 32065 32070 32075 32080 32085 32090 32095 32100 32105 32110 32115 32120 32125 32130 32135 32140 32145 32150 32155 32160 32165 32170 32175 32180 32185 32190 TPN ELECTROLYTES TRAC TRACILON TRAGACANTH TRAL TRAL MAG TRANCOPAL TRANME P TRANSACT GEL TRANTOIN TRANXENE TRATES GRANUCAP TRAV-AREX TRAVAD TRAVAD PREFILLED B.E. TRAVASE TRAVASOL TRAVASOL NUTRITION KIT TRAVASOL W/ELECTROLYTE TRAVERT TRAVERT IN SODIUM CHLORIDE TRECAT OR-SC TREMIN TRENDAR TREST TREX IN TRI HEXABAMATE TRI HIST TRI HIST DM SYRUP TRI HIST SYRUP TRI KORT TRI TINIC TRI-CONE TRI-HEMIC 600 TRI-HYDROSERPINE TRI-IMMUNOL TRI-K TRI-MEDEX EXPECT ORANT TRI-MEDICOL COUGH SYRUP TRI-MINE TRI-MINE EXPECTORANT TRI-MINE SYRUP TRI-0PH TRI-PHEN-CHLOR TRI-PHEN-GES IC TRI-QUAD TRI-SULFA #2 TRI-THALMIC TRI-VERT TRI-VI-FLOR TRI-VI-SOL TRIACET TRIACIN TRIACIN-C TRIACT TRIACTIN TRI AFED TRIAFED C EXPECTORANT TRIAFED SYRUP TRIAM MEDICATION CODE LIST, NAMCS 1980 32195 32200 32205 32210 32215 32220 32225 32230 32235 32240 32245 32250 32255 32260 32265 32270 32273 32275 32280 32285 32290 32295 32300 32305 32310 32315 32320 32323 32325 32330 32335 32340 32345 32350 32355 32360 32363 32365 32370 32375 32380 32385 32390 32395 32400 32405 32410 32415 32420 32423 32423 32424 32425 32430 32433 32435 32438 32440 32445 32450 TR IAMC I NOLONE TR IAMCINOLONE ACETONIDE TRIAMCINOLONE DIACETATE TR IAMC INOLONE NYSTATIN TRIAMINIC DM TR IAMINIC EXPECTORANT TRIAMINIC EXPECTORANT DH TRIAMINIC EXPEC TORANT W/CODEINE TRIAMINIC INFANT DROPS TRIAMINIC JUVELET TIMED-RELEASE TRIAMINIC TRIAMINIC TIMED-RELEASE TR IAMINICIN TRIAMINICIN SPRAY TRIAMINICOL SYRUP TRIAMOLONE 40 TRIAMTERENE TRIAPON TRIASYN B TRIA TROPHENE TRIAVIL TRIBARB TR ICHLORMETHI AZ 1DE TRICHLORMETHIAZIDE W/RESERPINE TRICHLOROACETIC ACID TR ICHLOROE THYLE NE TR ICHOLAN TR ICHOMONAC IDE TR ICHOTINE TRICLOS TR ICODENE PEDIATRIC TR ICONOL TR IDESILON TRIDIHEXETHYL CL MEPROBAMATE TRIDIONE TRIETHANOLAMINE TR IFLUOPERAZI NE TR IGESIC TRIGOT TR IHEXANE TRIHEXIDYL TRIHEXY-PHENIDYL TR IHEXYPHENID YL TRILAFON TRILION TRILISATE TRILOG TRILONE TRILOX TRIMETHOPRIM W/SULFAMETHOXAZOLE TRIMETHOPRIM W/SULFA SOXAZOLE TR IMIPRAMINE TRIMINOL COUGH SYRUP TR IMOX TRIMPEX TRIMSTAT TR IME THOPRIM TRIMTABS TRIND SYRUP TRINIAD 32455 32460 32465 32470 32475 32480 32485 32490 32495 32500 32505 32510 32515 32520 32525 32530 32535 32540 32545 32550 32555 32560 32565 32570 32575 32580 32585 32590 32595 32600 32605 32610 32615 32620 32625 32630 32635 32640 32645 32650 32655 32658 32660 32665 32670 32675 32680 32685 32690 32690 32695 32700 32705 32710 32715 32720 32725 32730 32735 TRINIAD PLUS 30 TRINSICON TRINSITRATE TRIOGEN TRIOSULFON DMM TRIPALAX TRIPALGEN SYRUP TRIPELENNAMINE TRIPHED TRIPHEDRINE TRIPHENYL TRIPIRIN TRIPLE ANTIBIOTIC TRIPLE ANTIGEN ULTRAFINED TRIPLE DYE TRIPLE SULFA TRIPLE SULFOID TRIPLE VITAMIN DROPS TRIPLEN TRIPROLIDINE TRIPROLIDINE W/PSEUDOEPHEDR INE TRIPTAVEL TRIPTAZINE TRISEM TRISOGEL TRISORALEN TRISORBIN DROPS TRISORBIN-F DRCPS TRISTOJECT TRISULFAPYRIMIDINES TRISUREID TRIVITAMIN DROPS TROBICIN TROCA INE TROCAL TROCINATE TROFAN TROISUL TRONOT HANE TROPH-IRON TROPHITE TROPICAMIDE TRYPTOPHAN TRYSUL VAGINAL TU-CILLIN TUAMINE TUBERCUL IN TINE TEST TUBERCUL IN PPD (HEAF) T.B. TINE TEST TUBERCULIN PPD TINE-TEST TUBERSOL TUBOCURARINE TUCAZYME H.P. TUCKS TUDECON TUINAL TULOIDIN TUMOL TUMS 32740 TURGASEPT 69 32745 32750 32755 32760 32765 32770 32775 32780 32785 32790 32795 32800 32805 32810 32815 32820 32825 32830 32835 32840 32845 32850 32855 32860 32865 32870 32875 32880 32885 32890 32895 32900 32905 32910 32915 32920 32925 32930 32935 32940 32945 TURG EX TURPENTINE TURPOIN TUS ORAMINIC SPANCAP TUSAL TUSS -ORN ADE TUSSADON T.D. TUSSAGES IC TUSSAMINIC T IMED-RELEASE TUSSANIL EXPECTORANT TUSSANIL-DH TUSSAR DM TUSS AR TUSSAR-2 SYRUP TUSSCAPINE TUSSCAPS TUSSEND TUSSEND EXPECT ORANT TUSSEX COUGH SYRUP TUSS I-ORGANIDIN DM TUSST-ORGANIDIN TUSSTHAB SYRUP TUSS IONEX TUSSTAT EXPECTORANT TUSSTROL TUZON TWELVE L-A TWINK TWINBARBITAL NO. 2 TWO-DYNE TYBATRAN TYCOPAN TYLENOL TYLENOL NO. TYLENOL NO. TYLENOL NO. TYLENOL NO. TYLENOL W/ CODEINE TYLENOL W/CODEINE ELIXIR TYLOST ERONE TYLOX WN = TYMATRO 32950 32955 32960 32965 32970 32975 32980 32985 32990 32995 33000 33005 33010 33015 33020 33025 33030 33035 33040 TYMPAGESIC TYPHOID VACCINE TYPHUS VACCINE TYROBENZ TYROHI ST TYROHIST SPRAY TYROSUM CLEANSER TYZINE TYZOMINT UR D UeRe Ie U-TRACT U-TRAN us ULCORT ULGEST IN ULO SYRUP ULTAR MEDICATION CODE LIST, NAMCS 1980 33045 ULTRA DERM BATH 33050 ULTRA MIDE LOTION 33055 W TRA TEARS 33060 ULTRACAINE 33065 ULTRACAINE W/EPINEPHRINE 33070 WL TRADINE 33075 ULTRAPAQUE 33080 ULTROGEN 33085 ULTROGEN-D.A. 33090 ULVICAL 33095 WNAVIT 33100 UNDECYLENIC ACID 33105 UNDECYLENIC ACID COMPOUND 33110 UNGUENTINE 33115 UNGUENTUM BOSSI 33120 UNIAD 33125 UNIAD PLUS 33130 UNIBASE 33135 UNIBON 33140 UNICAP 33145 UNIDOTE 33150 UNIGESIC-A 33155 UNIPEN 33160 UNIPRES 33165 UNISOL 33170 UNISOM 33175 UNITENSEN 33180 UNNA'S GELATIN PASTE 33185 UNPROCO 33190 URACEL 33195 URACIL MUSTARD 33200 URADAX 33205 URANAP 33210 WRAZIDE 33215 WEA 33220 UREAPHIL 33225 URECHOL INE 33230 UREMIDE 33235 URETHAN 33240 WREX 33245 URI-PAK 33250 UWRIDON 33255 URIFON 33260 URIFON-FORTE 33265 WRIHAB 33270 UR IMED 33275 URINARY ANTISEPTIC 33280 WRISED 33285 URISEDAMINE 33290 URISEP 33295 URISEPTIC 33300 UWRISPAS 33305 WISTAT 33310 WISTIX 33315 URITABS 33320 WITEN 33325 URITHOL 33330 WITIN 33335 WITRAL 33340 URO-PHOSPHATE 33345 33350 33355 33360 33365 33370 33375 33380 33385 33390 33395 33400 33405 33410 33415 33420 33425 33430 33435 33440 33445 33448 33450 33455 33460 33465 33470 33475 33480 33485 33490 33495 33500 33505 33510 33515 33520 33525 33530 33535 33540 33545 33550 33555 33560 33565 33570 33515 33580 23585 33588 33590 33595 33600 33605 33610 336 15 33620 33625 33630 URO-VES UROBILISTIX UROBIOTIC-250 URODINE UROGES IC UROLENE BLUE UROLOGIC G UROQID-ACID UROTOIN URSINUS UT IBID UTICILLIN VK UTICORT UT IMOX V GAN V.A. DOUCHE V-CILLIN V=-CILLIN K V-CORT V-LAX VA-TRO-NOL DROPS VACCINATION VACON VACUETTS VAGESIC PLUS VAGIDINE VAGILIA VAGIMINE VAGISEC VAGITROL VAL TEP VALACET VALADOL VALAX VALCAINE VALDRENE VALENOL VALERGEN VALERIAN VALERTEST VALIMENT L INIMENT VALINE VALISONE VALIUM VALMID VALOBAR VALPIN 50 VANCERIL VANCOCIN VANCOCIN HCL VANCOMYCIN VANILLIN VANOBID VANOX IDE VANOXIDE-HC LOTION VANQUISH VANSEB LOTION VAPO-1SO VAPONE FRIN VAPORUB oL 33635 33640 33645 33650 33655 33660 33665 33670 33675 33680 33685 33690 33695 33700 33705 VAPOSTEAM VARIDASE VARIDASE W/CARBOXYMETHYLCELL VARMEL VAS-0-SPAN VASAL GRANUCAP VASCORAY VASELINE VASELINE GAUZE VASO-80 UNICELLE VASOCIDIN VASOCL EAR VASOCOL SYRUP VASOCON VASOCON-A 33710 VASODI LAN 33713 33715 33720 33725 33730 33735 33740 33745 33750 33755 33760 33765 33770 33775 33780 33785 33790 33795 33800 33805 33810 33815 33820 33825 33830 33835 33840 33845 33850 33855 33860 33865 33870 33875 33880 33885 33890 33895 33900 33905 33910 33915 33920 33925 VASODILATOR VASOGESIC VASOGLYN UNICELLE VASOMI DE VASOMINIC T.D. VASOPRESSIN VASOPRINE VASOSPAN VASOSULF VASOTHERM VASOTRATE SUBLINGUAL VASOXYL VASTRAN VECTRIN VEETIDS VEGOIL W/ISOCAINE VEHICLE/N VEINAMINE VELBAN VELOSEF VELTANE VELTANE EXPECTORANT VELTAP VELVACHOL VENTACOL VENTACOL EXPECTORANT VENTAIRE VERA-67 VERACILLIN VERACOLATE VERATRUM VIRIDE VERAZINC VERCYTE VERDEFAM VEREQUAD VERILOID VERMIZINE SYRUP VERMOX VERNATE VERS AL VERS APEN VERS AP EN-K VERS TAT VERSTRAN MEDICATION CODE LIST, NAMCS 1980 33930 33935 33940 33945 33950 33955 33960 33965 33970 33975 33980 33985 33990 33995 34000 34005 34010 34015 34020 34025 34030 34035 34040 34045 34050 34055 34060 34065 34070 34075 34080 34085 34090 34095 34100 34105 34110 34115 34120 34125 34130 34135 34138 34140 34145 34150 34155 34160 34165 34170 34175 34180 34185 34190 34195 34200 34205 34210 34215 34220 VERV VES ICHOLINE VESPRIN VI-AQUA VI-AQUAMIN VI-BETA-C VI-CERT VI-DAYLIN VI-DAYLIN DROPS VI-DAYLIN F VI-DAYLIN F ADC DROPS VI-DAYLIN F ADC PLUS IRON DROPS VI-DAYL IN F DROPS VI-DAYLIN F PLUS IRON DROPS VI-DAYLIN OVER 4 VI-DAYLIN OVER 4 PLUS IRON VI-DAYLIN PLUS IRON VI-DAYLIN PLUS IRON ADC DROPS VI-DAYLIN PLUS IRON DROPS VI-DAYLIN PLUS IRON SYRUP VI-MAGNA VI-PENTA F VI-PENTA F INFANT DROPS VI-PENTA F MULTIVITAMIN DROPS VI-PENTA INFANT DROPS VI-PENTA VI-SYNERAL VI-SYNERAL ONE-CAPS VI-THEL VI-ZAC VIBEDOZ VIBRA VIBRAMYC IN VICALTEIN VICAM VICKS INHALER VICODIN VICON VICON FORTE VICON-C VICON-PLUS VICON-T VIDARAB INE VIFEX VIGRAN VIGRAN PLUS IRON VIMAH VINCRISTINE VINGESIC NO. 3 VINI-RUB VIO-BEC VIG-BEC FORTE VIO-GERIC VIO-HYDROCORT VIO-HYDROSONE VID-PRAMOSONE VIO-SERPINE VIOFORM VIOFORM-HYDROCORTI SONE VIOKASE 34225 VIOTAG 34230 VIPEP 34235 VIRA-A 34240 VIRILON 34245 VIROMED 34250 VISALENS WETTING 34255 VISCULOSE 34260 VISINE 34265 VISTACON 34270 VISTARIL 34275 VISTAZINE 34280 VISTRAX 34285 VITA IRON 34290 VITA TOT 34295 VITA-GLEN 34300 VITA-KAPS 34305 VITA-METRAZOL 34310 VITABEE 34315 VITACEE 34320 VITACEE SYRUP 34325 VITADEX-B 34330 VITADYE 34335 VITAGETT 34340 VITAKAPS-M 34345 VITAL 34350 VITALITY-E 34355 VITALYNE 34360 VITAMIN A 34365 VITAMIN A + VITAMIN D 34370 VITAMIN AED 34375 VITAMIN A & D HI-POTENCY 34380 VITAMIN A NATURAL 34385 VITAMIN A NATURAL NEO-VADRIN 34390 VITAMIN A PALMITATE 34395 VITAMIN A PLUS D 34400 VITAMIN A SOLUBILIZED 34405 VITAMIN A SOLUBLE 34410 VITAMIN A SOLUBLE NEO-VADRIN 34415 VITAMIN A SYNTHETIC 34420 VITAMIN A WATER SOLUBLE 34425 VITAMIN B COMPLEX 34430 VITAMIN B COMPLEX B-12 W/C 34435 VITAMIN B COMPLEX EL IXIR 34440 VITAMIN B COMPLEX HI POTENCY 34445 VITAMIN B COMPLEX W/B-12 34450 VITAMIN B COMPLEX W/VITAMIN C 34455 VITAMIN B-1 34460 VITAMIN B-1 & B-12 34465 VITAMIN B-1 & B-12 TONIC 34470 VITAMIN B-1 & B-12 W/IRON 34475 VITAMIN B-1 ELIXIR 34480 VITAMIN B-1 NEO-VADRIN 34485 VITAMIN B-1 W/B-6 34490 VITAMIN B-1 W/B-6 & B-12 34495 VITAMIN B-12 34500 VITAMIN B-12 CRYSTALLINE 34505 VITAMIN B-2 34510 VITAMIN B-6 34515 VITAMIN B-6 NEO-VADRIN 34520 VITAMIN C 74 34525 VITAMIN C & E 34810 34530 VITAMIN C CRYSTAL 34535 VITAMIN C FLAVORED 34540 VITAMIN C NEO-VADRIN 34545 VITAMIN C PLUS E 34550 VITAMIN C SYRUP 34555 VITAMIN C T.D. 34560 VITAMIN CHEWABLE CHILDREN'S 34565 VITAMIN D 34570 VITAMIN D-2 IN OIL 34575 VITAMIN 34580 VITAMIN E € C 34585 VITAMIN E NATURAL 34590 VITAMIN E NATURAL NEC-VADRIN 34595 VITAMIN E NATURAL SOLUBLE 34600 VITAMIN E NEO-VADRIN 34605 VITAMIN E SKIN OIL 34610 VITAMIN E SYNTHETIC 34615 VITAMIN E WAFER 34620 VITAMIN E ZESTAB 34623 VITAMIN K 34625 VITAMIN 34628 VITAMINS AND MINERALS 34630 VITANATE 34635 VITANATE FA 34640 VITAPHEN 34645 VITERRA 34650 VITERRA C 34655 VITERRA HIGH POTENCY 34660 VITERRA RDA 34665 VITERRA RDA PLUS IRON 34670 VITERRA-E 34680 VITRON-C PLUS 5 I 34690 VIVARIN I 34700 VLEM-DOME LIQUID CONCENTRATE 34705 VLEMINCKX® SOLUTION 34710 VOCALZONES 34715 VOLAXIN MODIFIED 34720 VOLEX IN SODIUM CHLORIDE 34725 VONCE SHAMPOO 34730 VONTROL 34735 VORANIL 34740 VOSOL 34745 VOSOL HC 34750 VULVAELINE 34755 VYTONE 34760 W.D.D. 34765 W-T LOTION 34770 WAMPOCAP 34775 WARFARIN 34780 WART OFF 34785 WASH °N DRI TOWELETTE 34790 WATER DISTILLED 34795 WATER FOR INJECT ION 34800 WATER FOR IRRIGATION 34805 WATER PURIFIED WEHA MI NE MEDICATION CODE LIST, NAMCS 1980 34815 WEHDRYL 34820 WEHGEN 34825 WEHLESS 34830 WEHVERT 34835 -WEIGHTROL 34840 WESCOHEX SURGICAL SCRUB 34845 WESLAX 34850 WESTADONE 34855 WESTCORT 34860 WET N SOAK 34865 WETTING & SOAKING 34870 WETTING SOLUTION BARNES HIND 34875 WHEAT GERM OIL 34880 WHITE PINE COMPOUND 34885 WHITFIELD S 34890 WIGRAINE 34895 WILD CHERRY SYRUP 34900 WINE 34905 WINGEL 34910 WINSTROL 34915 WINTERGREEN OIL 34920 WITCH HAZEL 34925 WOLFINA 34930 WOOD ALCOHOL 34935 WUN-TABS 34940 WUN-TABS W/ IRON 34945 WYAMINE SULFATE 34950 WYAMYCIN 34955 WYANOID 34960 WYANOIDS 34965 WYANDIDS HC 970 34975 MWYCILLIN INJ & PROBENECID TAB 34980 WYDASE 34985 WYGESIC 34990 WYMOX 34995 X SEB H.C. SCALP LOTION 35000 X SEB SHAMPOO 35005 X SEB T SHAMPOO 35010 X-0TAG 35015 X-PREP 35020 X-PREP BOWEL EVACUANT KIT 35025 XERAC AC 35030 XERAC BP10O 35035 XERAC BPS 04% XER 35045 XERO-LUBE 35050 XEROFOAM DRESSING 35055 XYLO-PFAN 35060 XYLOCAINE 35070 XYLOCAINE HCL 35075 XYLOCAINE HCL W/DEXTROSE 35085 XYLOCAINE VISCOUS 35090 XYLOMET SPRAY 35095 XYLOMETAZOLINE 35100 XYLOSE CP 35105 YEAST 35110 YELLOW FEVER VACCINE 35115 YODOXIN 35120 YOHIMBINE 35125 Z.B.T. BABY POWDER 35130 Z-BEC 35135 Z-PRO-C 35140 ZACTANE 35145 ZACTIRIN 35150 ZARONTIN 35155 ZAROXOLYN 35160 ZARUMIN 35165 ZEASORB MEDICATED 35170 ZEM HISTINE 35175 ZEMACOL MEDICATED LOTION 35180 ZEMACON 35185 ZEMALO SKIN CLEANSER 35190 ZENI-B W/C 35195 ZENIVITES M 35200 ZENTINIC 35205 ZENTRON 35210 ZEPHIRAN CHLORIDE 35215 ZESTE 35220 ZESTE M.T. 35225 ZETAR 35230 ZETRAN 35235 IDE 35240 ZINC 35245 ZINC GLUCONATE 35250 ZINC OXIDE 35255 ZINC PASTE (LASSARS) 35260 ZINC SULFATE 35265 ZINC SULFATE COMPOUND 35270 ZINC SULFATE MUROCOLL NO. 35275 ZINC SULFATE NEO-VADRIN 35280 ZINC SULFIDE COMPOUND LOTION 35285 ZINC-20 35290 ZINC-220 35295 ZINCATE ZINCFRI 35305 ZINCOFAX 35310 ZINCON 35315 ZIPAN 35320 ZIRADRYL LOTION 35325 ZOLINE-M 35330 ZOLINE-S 35335 ZOLYSE 35338 ZOMAX 35340 ZONIUM CHLORIDE 35345 ZYLAN 35350 ZYLOPRIM 35355 ZYMACAP 35360 ZYMALIXIR 35365 ZYMASYRUP 35370 ZYMAVITES 35375 ZYMENOL 35380 ZYMME 35385 2-PROPANOL 35390 2/6 35395 2G/DM 35400 4-WAY LONG ACTING 35405 4&-WAY NASAL SPRAY cL 35410 ANTICONVULSANT AGENT 35415 ANTIDEPRESSANT AGENT 35420 ANTIDIABETIC AGENT 35425 ANTI DI ARRHEAL AGENT 35430 ANTIEMETIC AGENT 35435 ANTIEPILEPSY AGENT 35440 ANTIFLATULENT AGENT 35445 ANTI FUNGAL AGENT 35450 ANTIHISTAMINE 35455 ANTIINFECTIVE AGENT 35460 ANTI INFLAMMATORY AGENT 35465 ANTILIPEMIC AGENT 35470 ANTIMALARIAL AGENT 35475 ANTINAUSEANT AGENT 35480 ANTINEOPLASTIC AGENT 35483 ANTIOBESITY AGENT 35485 ANTI PROTOZOAL AGENT 35490 ANTIPRURITIC AGENT 35495 ANTIPYRETIC AGENT 35500 ANTITHYROID AGENT 35505 ANTI TUBERCULAR AGENT 35510 ANTIVIRAL AGENT 99980 OTHER 99999 ILLEGIBLE MEDICATION CODE LIST, NAMCS 1980 Appendix IV. Coding proce- dures for medication entries, NAMCS 1980 CODING PROCEDURES FOR MEDICATION ENTRIES NAMCS 1980 SOURCE MATERIALS — Medication Code List 1980 (MCL) — Attachment 1: “Common Abbreviations Used in Medical Orders” — Attachment 2: “Caution! 1,000 Drugs Whose Names Look-Alike or Sound-Alike” — Attachment 3: Drug Form I (Coder A) — Attachment 4: Drug Form I (Coder B) — Attachment 5: Drug Form II (Adjudicator) GENERAL These instructions apply to the coding of the information contained on the 1980 Patient Record in Item 11: MEDICATION THERAPY THIS VISIT. On the lines provided in Item 11, the physician records all medications ordered or provided at the visit. These include new or continued medications which may be prescription or non- prescription (over-the-counter) drugs. Immunizing and desensitizing agents are also included. For making the entry of a medication, the physician uses one of two names: EITHER 1. the generic name. This is the “official” or “nonproprietary’’ name. Usually, it is the contraction of a complex, chemical name. OR 2. the brand or trade name. This is the commercial or “proprietary” name, the name that is used to advertise a drug to the medical profession. For example, tetracycline is the generic name of a widely used antibiotic. A physician may write a prescription for this drug by its generic name, or may prescribe it under the brands names Achromycin (made by Lederle), Sumycin (made by Squibb), or Tetracyn (made by Pfizer). Item 11 has two parts. Item 11a provides space for listing medications ordered or provided for the principal diagnosis shown in Item 9a. Item 11b provides space for listing medications ordered or provided for all other reasons. The basic coding document is the Medication Code List 1980 (MCL). It contains 7,081 brand and generic names of medications, alphabetized and coded serially from 00005 to 35405. In addition, there is a code (99999) for 73 illegible entries and an ‘other’ code (99980) for entries that are legible but cannot be found on the Medication Code List 1980. One of three coding choices will be applied to every medication entry: a code from the name list, the illegible code, or the ‘other’ code. The name that appears on the Medication Code List 1980 is the core identifier of a medication. That is, it usually contains just enough information to set it clearly apart from all other medications. In most cases the core identifier will not contain specific information about — the ‘labeler’ (manufacturer of a drug) — the dosage form (tablet, capsule, inhalant, ointment, injection, suppository, etc.) — the route of administration, that is, the path by which a drug product enters or is applied to the body (intramuscular or I.M., implant, inhalation, intravenous or I.V., nasal, ophthalmic, oral, otic, rectal, topical, etc.) — the strength of a drug product (milligram or mg, milliliter or ml, percent or %, etc.) —instructions for use. For example, “Take as needed (p.r.n.).”” or “Take twice a day (b.i.d.).” — other qualities of the drug such as color or flavor Though the above descriptors are not usually included as part of the core identifier of a medication in the MCL, they may appear in many of the medication entries that the physician makes on the Patient Records. Mostly, they appear because many physicians will write the same entries on the Patient Records that they wrote on the patient’s prescription. Therefore, the coders will have to take them into consideration in their search for a core identifier to be matched against the Medication Code List 1980. For the specimen entries below, the core identifiers have been circled, matched against the Medication Code List 1980, and the proper code identified. Specimen entry Code Phillip’s chocolate 19375 blue 14620 (Prednisone ) 5 mg tid x 5d! 24890 im.2 04235 IDosage strength 5 milligrams. Take 3 times per day for 5 days. Injected intramuscularly. As an aid in isolating the core identifiers, the coder will find it useful to refer to Attachment 1: “Common Abbreviations Used in Medical Orders.” Once the core identifier has been stripped of all (or as much as possible) of the extraneous material that sur- rounds it, it will be matched alphabetically against the Medication Code List 1980. A perfect fit will not be possible in every case but the coder will always attempt to MAKE THE CLOSEST MATCH POSSIBLE The specimen entries above appear in typed form and have been edited for correct spelling, facilitating the isola- tion of their core identifiers. This will not be the case in the actual coding operation, where the entries are more likely to look like the following real facsimiles of 1980 Patient Records. 74 Example A 11. MEDICATION THERAPY THIS VISIT (J NONE | Using brand or generic names, record all new and continued medications ordered, injected, administered, or otherwise provided at this vist. Include immunizing and desensitizing agents] a. FOR PRINCIPAL DIAGNOSES IN ITEM Qa. b. FOR ALL OTHER REASONS. b A cele X 2 3, 3 4 4 Example B 11. MEDICATION THERAPY THIS VISIT TINONE [ Using brand or generic names, recond all new and continued medications ordcred, injected. administered, or otherwise provided at this vist. Include innnunizing and desensitizing agents | A. FOR PRINCIPAL DIAGNOSES IN ITEM 9a, D. FOR ALL OTHER REASONS t; Np rn 2 Ngee ! 3 3; Example C 11. MEDICATION THERAPY THIS VISIT (J NONE [ Using brand or generic names, record ail new and continued meaications ordered, injected, administered, or otherwise provided ct this vist. Include immunizing and desensitizing agents] |, AEC PPINGIRAL DIAGNOSES IN ITEM Oo. Ls evan Lhd L 2 A Aaron. ¢ “h i J 3, 4, 4. b. ENR ALL OTHER REASONS. Example D 11. MEDICATION THERAPY THIS VISIT CJ NONE | Using brand or gencric names, record all new and continued medications ordered, injected, administered, or otherwise provided at this vist. Include immunizing and desensitizing agents] a. FOR PRINCIPAL DIAGNOS! J EM Ya b. FOR ALL OTHER BEASONS. CN A= ~~ | 1S 2498 Tears >A) & oe, : 2 t 2, 3 3 4. 4 75 76 Exawmple E ld = uh $A, IP | 11. MEDICATION THERAPY THIS VISIT [J NONE [ Using brand or generic names, record all new and continued medications ordered, injected, administered, or otherwise provided at this vist. Include immunizing and desensitizing agents a. FOR PRINCIPAL DIAGNOSES IN ITEM 9a. b. FOR ALL OTHER REASONS. _lravnmeneos - 2 q edn LID : APY ru 3 4. 4, 13 WASPATIENT | 44 DISPOSITION THIS VISIT | 48 DURATION Example = 11. MEDICATION THERAPY THIS VISIT (J NONE Using brand or generic names, record all new and continued medications ordered, injected, administeped, or otherwise provided ut this vist, Include immunizing and desensitizing agents | a. FOR PRINCIPAL DIAGNOSES IN ITEM Sa, b. FOR ALL OTHER REASONS. " Pl yess " : < ~~ Cc 5? Example es 11. MEDICATION THERAPY THIS VISIT (OJ NONE [Using brand or generic names, record all new and continued meaications ordered, injected, administered, or otherwise provided at this vist. Include immunizing and desensitizing agen rs] a F PRINCIPAL DIAGNOSES IN ITEM 9a, b. FOR ALL OTHER REASONS. ' : We NW, h pry fP0 NG rp 4 2 A (iruets cor 3 £2 Hexallal . Here are the correct core identifiers for these examples: patient Nil Item 11a Item 11b Example A 1 Phenobarbital Acelax 2 Dilantin Example B 1 Ampicillin 2 Aspirin Example C 1 Aldoril 2 Persantine 3 Atromid Example D 1 Diabinese Lasix 2 Tenuate Example E 1 Garamycin 2 Cyclogyl 3 Neosynephrin A Sodium Chloride 5 Alcaine Example F 1 Fluress 2 Cyclogyl 3 Neosynephrin Example G 1 Aristocort 2 Hexadrol It is readily apparent that legibility is going to be a major problem. At least two things will help the coder over- come this problem. 1. Experience: Names and spellings of medications that appear impossible or difficult at the beginning of the coding operation will become more and more familiar as the coder proceeds through an increasing number of Patient Records. 2. Familiarity with the writing peculiarities of a given physician. Patient Records completed by the same physician will be batched together and, in this fashion, presented by the coding supervisor to the coder. Thus, an early entry which cannot be deciphered may later in the batch become legible after the coder has gained a grasp of the physician’s writing style. 77 Another problem (complicated by legibility) lies in the fact that some drugs look or sound like others. To help solve this problem, the coder should consult Attachment 2: “Caution: 1,000 Drugs Whose Names Look-Alike or Sound-Alike.” Entries will be numbered and coded sequentially from top to bottom. If there is more than one entry on a line, each will be included in the sequence, numbering from left to right (See Example B). No more than four entries will be coded in Item 11a and no more than four in Item 11b. Any additional entries, such as “Alcaine” in Example E, will not be coded. It is the core identifier of a medication that is numbered, not any extraneous, descriptive material. Thus Aristo- cort is entry number 1 in Example G, even though it appears in the third space. THE CODING PROCESS The specific instructions on coding and reporting that follow are provisional instructions, that is, they are subject to alteration as our experience in the coding and reporting process develops over time. NORC staff are invited to suggest any changes that will improve the efficiency and/or economy of the operation. Rule: There will be 100 percent, independent coding of every medication entry. Functions Personnel Control and D1I0W DUNCHON. « vs vv » wuss 550m wm mime ove ams * sms es Coding supervisor Clerical aide Coding function ........... iii iii iiinenennn Coder A Coder B Adjudicating function. ............ iii ie ee Adjudicator* *Specially qualified by education and experience in drug nomenclature and use. Should preferably have on- the-job familiarity with the prescribing practices of physicians. The responsibility for all three of these functions will rest with NORC. Steps 1. The coding supervisor will assign Patient Records to the coders, batched so that all the Records from the same physician are kept together. 2. Coders A and B will independently code the medication entries, each listing the selected codes on the appropriate DRUG FORM 1 (see Attachments 3 and 4). The coders will not request any assistance from the Adjudicator at the time of coding and will not consult each other regarding specific coding. 3. The coders will deliver completed DRUG FORM I to the clerical aide. For every entry where the code assigned by Coder A agrees with the code assigned by Coder B, the clerical aide will write that code directly on the Patient Record. When the coders disagree on a code and/or for every entry coded illegible or ‘other’, the clerical aide will put a check mark (+/ ) opposite these codes on the copies of DRUG FORM I, and deliver the forms to the Adjudicator along with the appropriate Patient Records. Reference to Attachments 3 and 4 will demonstrate the processes described in steps 2 and 3. 4. The Adjudicator rules on any disagreement between Coder A and Coder B, either: (1) choosing A’s code or B’s code or (2) choosing some other selection from the Medication Code List 1980. Also every illegible or ‘other’ entry will be carefully checked in the adjudication process, even if both coders agree that it is illegible or ‘other’. For entries which require adjudication, including all illegibles or ‘others’, the Adjudicator will use DRUG FORM II to record the codes assigned by Coders A and B and to enter his (her) own coding choice. The process is illustrated in Attachment 5. Finally, the adjudicated coding choice will then be entered directly on the Patient Record (by either the Adjudicator of the clerical aide). 78 5. A weekly written report will be made to: Hugo Koch Ambulatory Care Statistics Branch, NCHS Rm. 2-28, Federal Center Building 3700 East West Highway Hyattsville, Maryland 20782 The report will contain the following: a. Number of Patient Records processed. b. Number of Patient Records with at least one medication entry. c¢. Total number of medication entries. d. Number of medication entries in which a difference existed between Coder A and Coder B. Enclose copies of appropriate DRUG FORM II’s and Patient Records. e. List of all medication entries coded ‘other’. Additional duties of the Adjudicator: [n addition to his (her) contribution to the direct, coding process already discussed, the Adjudicator will further: 1. Conduct a sample check of Patient Records where no difference existed between Coder A and Coder B, with the aim of discovering possible areas of consistent bias. 2. Develop list of medications that are most commonly misspelled and supply them to the coders to aid them in the on-going coding of the 1980 medication entries. When the coding of the 1980 medication entries has been completed, the Adjudicator will supply a list of the 100 most commonly misspelled medications to Hugo Koch (address above). 3. Develop lists of the medications most frequently associated with common principal diagnoses and supply them to the coders to aid them in coding entries for Item 11a. 4. Conduct periodic training sessions for the coders to correct specific patterns of error that the adjudicator observes developing, and to share his (her) specialized drug knowledge. 5. Continually evaluate the adequacy of the medication coding instruments and processes, and communi- cate suggestions for improvement in frequent, informal contact with: Hugo Koch Phone: (301) 436-7132 *® ok kk ok ok 79 Attachment 1. 80 COMMON ABBREVIATIONS USED IN MEDICAL ORDERS Word Abbreviation Meaning ana... BB, 88 oo vvvniinivmsn of each antecbum ................. 1 A before meals or food B80 cs vramvams ste we HR BE ad ................ to, up to auriodextra . ................ BG ov vin sees maa right ear adlibitum .................. adiih, reece es at pleasure auriolaeva ................. 8h cocsenerranamrna left ear ante meridiem ............... AM. ones vinnmamuas morning AQUB «ovr mn ve nn EE BY sn vw mens nas mas water aqua destillata . .............. 80.968 ..... 00. meas distilled water aurio sinister . . . ............. BE sve ms rma wens left ear auresutrae . ................ Bil un we vane oe sea each ear bisindie .................. bid: ....vncenninnin twice daily bowel movement . . ............ BM cor nvionsaninmss bowel movement blood pressure . .............. BP. cv vrs minnrnnsnnsn blood pressure CUM vs mme mus mpsnts assis Hp with capsula ................... CAPS. «vv vm ovnwn sas n capsule compositus . ................ BOMP: vv vbw an 58005 compound des ..................... 0 comer ses mE rs day dive ..................... Qh: covrvrvnememany dilute dispensa . . ................. disp. ............... dispense divide .................... + AEE divide dentur talesdoses . . ........... did. ... ern Emim rs give of such a dose elixir ........ 0... 8 iu imramiERiTrs elixir Blo int ews mim ma nena rennin nm and fac, fiat, fiant ................ fof ihe mins make, let be made gramma . ..........eeee. GM. 0. ov ersns wns gram QUAN... nna sree mE an QF cn hE Ee aE nn grain GUID, conn ms we me gi. . rm meEa mee ey a drop DOTA ovine ns me BE wa 3m I im srwnng ms naam hour hOrasomni . ........z:00smea hs, hor.som. -........ at bedtime BHQUDY =. -vv 559 3 2 ad nin kok wn I smevesmvnmros ems a liquor, solution MISOE ... oso am mswmm nm msinve ss M. cosa dmans nos mix MOBAICIOF «svn ssmosmsnsiams MmAick ..iov.uvinanas as directed MRE coins smonmsmesas MIKE ovo vonemas mans a mixture DUMBIUS .. . : sc vsomsmusmemusims NO. sx ssiwaimmemasio nus number AOCIUMEl =; cvoimuiasnsmemusas POC: so v:nseaswuisuss in the night nonrepetatur ................ NON TBD. wv weiwusmsss do not repeat, no refills OClANUS. . 7 somsnms ping messisns 0,0Oct. .............. a pint oculusdexter ................ OM: 50 0 ss ms mne rs nes right eye oculuslaevus . ............... Oh sowevs noms sswsa left eye oculus sinister . .............. 08: vovwsmesunsasmus left eye OCU UBIGUB . - v5 vv cv navman OM, ow vps mw vw mes bows each eye POSLOIDOS cs mnninmsivawni ns PL, post. Oi. «usm after meals POStMBACIBN +. so cv vrs vrmesvs PM. «vv: vnrrmismnns afternoon or evening PEIOS .v:asumvsmsnasavems PO. tev nn mim utmnns by mouth PrOrenala ;...xsvuscn mur ns PIT ovr oss ss mus mens as needed PUVIB isin ssmermsamennsma PIV. “ov rss r im mr ary a powder quiaque hora . ............... Qh cosa ni mses mnn every hour QUAI INGIB: «vx wiw vn nmran Qi. +s canis agra four times a day quantum sufficiat . . . ........... 8: sim n si ® Bk wr a sufficient quantity *The listing of commonly used abbreviations is included as an aid in interpreting medical orders. Word Abbreviation quam volueris . . . ............. AV. conn recipe . . . vii RY ..%.covnsmms repetatur . ................. TOD: + «vv bv mk os SINE ,. vip inmsmY ih nd BEE Be in ein nee wn ne secundumartem .............. 8h. . inn nn sataratus .................. SBE ose BIONA +s ss sv msn ss mse n rn wo Sig. ............ solutio . ................... sol. ............ SEMIS . . oo §5.,SS .......... siopus sit .................. 30 X- statim . ................... stat. ........... suppositorium . . ..... LLL SUPP. + vee SYTUPUS . . «vv vee BY ve me a 20ONA. .. ws rns mE re a al. ie ie wa Bring ...... ccs c0c vemos 2d. css cnn va ae HACIIS: oc cin ms Beh & nT 4000 nee 1/11 A Larisa ema. ms unguentum . .......... Lo... UNG: =: inuiwonsws MOCHIN =. cvs vss stm: msi 1 A. while awake ................ Wl: «savin me as much as you wish take, a recipe let it be repeated without according to art saturated label, or let it be printed solution one-half if there is need at once, immediately suppository syrup tablet three times a day tincture triturate ointment as directed while awake 81 82 COMMON SYSTEMS OF WEIGHT AND MEASURE* Metric Weight 1 microgram pg (mcg) 1 milligram mg 1 centigram cg 1 decigram dg 1 gram g 1 dekagram Dg 1 hectogram hg 1 kilogram kg METRIC SYSTEM 0.000,001 0.001 0.01 0.1 1.0 10.0 100.0 1000.0 QUO Metric Liquid Measure 1 microliter ul = 0.000,001 = 1 milliliter m = 0.001 = 1 centiliter cl = 0.01 = 1 deciliter dl = 0.1 = 1 liter L = 1.0 = 1 dekaliter DI = 10.0 = 1 hectoliter hl = 100.0 = 1 kiloliter ki = 1000.0 = Note—The abbreviation pug or mcg is used for microgram in pharmacy rather than gamma (y) as in biology. Apothecary Weight APOTHECARY SYSTEM Apothecary Liquid Measure 1 grain gr = 1gr 1 minim m = 1m 1 scruple = 20¢r 1 fluidram 3 = 60m 1 dram 3 = 60gr. =32 1 fludounce f3 = 480m = 81{3 1 ounce 3 = 480¢gr = 83 1 pint pt 7680m = 16f3 1 pound Bb =5760gr = 123 1 gallon gal =61440m = 8 pt AVOIRDUPOIS SYSTEM Avoirdupois Weight 1 ounce =1 oz. 1 pound =1 Ib. 437.5 grains (gr.) 16 ounces (0z.) = 7000 grains (gr.) NOTE: The grain in each of the above systems has the same value, and thus serves as a basis for the interconversion of the other units. *The listing of common systems of weight and measure is included to aid the practitioner in calculating dosages. rrr r APPROXIMATE PRACTICAL EQUIVALENTS* Weight Equivalents 1 gram 1 kilogram 1 ounce avoirdupois 1 ounce apothecary 1 pound avoirdupois 1 grain Measure Equivalents 1 milliliter 1 fluidounce 1 liter 1 tablespoonful 1 wineglassful 1 teacupful 1 tumblerful 1 pint 1 gallon =1Gmorg = 15.432 = 1Kg = 220 = 102 = 28.35 =13% = 31.1 =1lb = 454, =1 gr = 648 =1ml = 18 =113 = 29 = 1 tbs or tbsp w 15 = 413 = 120. =815 = 240. = 1 pt or O or Oct = 473. = 1galor Cor Cong = 3785. grains pounds avoirdupois(lb) grams grams grams milligrams .23 minims (m) 57 ml .8 fluidounces (f3) mi ml ml ml ml mi *The listing of approximate practical equivalents is included to aid the practitioner in calculating and converting dosages among the various systems. Reproduced with permission of the “A merican Drug Index 1979” 83 Attachment 2. Updated CAUTION! 1,000 Drugs Whose Names Look-Alike or Sound-Alike Ben Teplitsky, Pharmacist Pharmacy Editor, Pharmacy Times » The accurate reading and interpretation of prescriptions is one of your key responsibili- ties as a pharmacist. That is why you must be so very careful (a) when you take a prescrip- tion over the phone or (b) when you attempt to decipher a doctor’s scribbling. You take spe- cial care to avoid dispensing a drug product not intended by the prescriber. Physicians’ Scribbling Dispensing errors can result from sound- alike or look-alike drugs. Similarities might be due to pronunciation—e.g., Donnatal and Donnagel. These drug names sound alike to the listener. Or, there may be similarities in the way the drug names look when quickly scribbled by a physician or quickly read by a pharmacist; for example, phenobarbital and pentobarbital, Decadron and Decaderm, and Brethine and Banthine. In some cases, drugs might be both look- alikes and sound-alikes. Amoxil & Amcill and digitoxin & digoxin are examples of these similarities. Most drugs on the list were sent to us by pharmacists. Fortunately, pharmacists contact the pre- scriber—immediately—when doubts exist. On the following two pages, PHARMACY TIMES publishes an updated, alphabetical list of 1,000 look-alike and/or sound-alike drugs. We suggest that you familiarize yourself with the entire list, because it is published in pairs of drugs. Thus, you may have a prescription for a drug which is the second half of a listing, and you might not see it right away. For example, if a doctor phones in a pre- scription for ampicillin and you check to see if it is on the list, you will find nothing listed under the first column of the A’s for ampicillin. However, if you are familiar with the list, you will recall that there is a listing under the C’s for “Compocillin . . . Ampicillin.” Mail Other ‘Alikes’ to Pharmacy Times After you have studied the list carefully, we suggest that you post it in the prescription aepartment as a handy reference. Pharmacists are urged to mail us the names of drugs they have encountered which are sound-alikes and/or look-alikes. An additional list of such problem drugs will be published in a future issue. Just mail your “alikes” to Ben Teplitsky, Pharmacy Editor of PHARMACY TIMES, 80 Shore Road, Port Washington, New York 11050. More » Reproduced with permission of the PHARMACY TIMES. Look-Alikes and Sound-Alikes A Aarane ........ Anturane Aarane ........ Artane Aberel ........ Iberol Acetohexamide . . Acetazolamide Achromycin ....Aureomycin Adapin ........ Atabrine Adroyd ........ Android Aerolone ...... Aralen Aerolone ...... Arlidin Afrin LLL... Afrinol Afrin ......... Aspirin Agoral ........ Argyrol Aldactone ..... Aldactazide Aldoril ........ Aldomet Alu-Cap ....... Aluscop Ambenyl ...... Ambodryl Ambenyl ...... Aventyl Amodrine ...... Amonidrin Amoril ......s. Amcill Ananase ....... Orinase Ananase ....... Tolinase Anavar ........ Antepar Ancobon ...... Oncovin Anturane ...... Antuitrin Anturane ...... Artane Anusol ........ Aplisol Anusol ........ Aquasol Aplisol ........ Apresoline Aplisol ........ Atropisol Appedrine ..... Ephedrine Apresoline ..... Priscoline Aralen ........ Arlidin Arfonad ....... Afrin Arthralgen ..... Auralgan Asminyl ....... Asmolin Asminyl ....... Esimil Asminyl ....... Ismelin Atarax ........ Enarax Atarax ........ Marax Ativan ........ Avitene Auralgan ...... Ophthalgan Azathioprine ... Azulfidine Azene ........ Azatadine Azolate ....... Azolid Azotrex ....... Azo-Stat ‘Azotrex ....... Tetrex B Bactocill ...... Pathocil Bactrim ....... Bacitracin Banesin ....... Benisone Banthine ...... Bentyl Belladonna ....Belladenal Benadryl ...... Belladenal Benadryl ...... Benty! Benadryl ...... Benylin Benadryl ...... Caladryl Bendopa ...... Bendectin Benemid ...... Beminal Benoxyl ....... PanOxyl Bentyl ........ Aventyl Bentyl ........ Bontril Benzedrex ..... Benzedrine Betalin ........ Benylin Betapar ....... Betapen Dexameth ..... Dexamyl Dextran ....... Dexedrine Diabinese ...... Dianabol Bicillin ........ V-Cillin Diafen ........ Delfen Bicillin ........ Wycillin Dialog ........ Halog Bontril sninenss Vontrol Dialose ....... Dialog Brethine ....... Banthine Dialume ....... Dalmane Brondecon ..... Bronkotabs Dialume ....... Dialose Butabarbital .... Butalbital Diasal ........ Diasone Butibel ........ Butabell Diazepam ...... Diazoxide Butisol ........ Butabell Dicarbosil ..... Dacarbazine Butisol ........ Butazolidin Digitoxin ...... Digoxin Digoxin ....... Desoxyn C Dilantin ....... Delalutin Calcidin ....... Calcidrine Dilantin ....... Deltalin Calcitriol ...... Calcitonin Dilantin ....... Phelantin Capastat ...... Cépastat Dilaudid ....... Dilantin Catapres ...... Catarase ‘Dimacol ....... Dimercaprol Catapres ...... Diupres Dimetane wien et Dimentabs Cephalexin . Cephalothin Dipaxin ....... Digoxin Cephapirin ..... Cephradine Disipal ........ Disophrol Chlorambucil . Chloromycetin Disomer ....... Disophrol Chloromycetin . . Chlor-Trimeton | Disophrol ....... Isuprel Chlorpromazine . . Chlorpropamide Disophrol ...... Stilphostrol Clinitest ....... Citanest Disopyramide - Dipyridamole Clofibrate ...... Clorazepate Diuril ......... Doriden Clonidine ...... Quinidine Diutensen ..... Salutensin Clonopin ...... Clonidine Diutensen ..... Unitensen Clonopin ...... Clopane Dobutamine - Dopamine Codeine ....... Coldene Dolene .......s Dilone Codeine ....... Cordran Dolene ........ Dolonil Coenzyme-B . Cotazym-B Dolohll css Dilone Colestid ....... Colistin Donnatal ...... Dianabol Colistin ....... Colestipol Donnatal ...... Donnagel Combid ....... Combex Donnazyme - Entozyme Combipres ..... Catapres Dopar ........ Dopram Compazine . Compocillin Dopram ....... Dopamine Compocillin . . .. Ampicillin Doriden ....... Doxidan Consotuss ..... Cotussis Doriden ....... Loridine Coramine ...... Calamine Doxan ........ Doxidan Cortone ....... Cort-Dome Doxinate ...... Doxan Cuprimine ..... Cuprex Dyazide PRT El Diasone Cyclopar ...... Cytosar Dyazide ....... Thiacide Cytarabine ..... Vidarabine Dyclone ....... Dilone Cytellin ....... Cytoferin Dyclonine ..... Dicyclomine Cytoxan ....... Cytosar Dyrenium ...... Pyridium D E Dalmane ...... Demulen Ecotrin ........ Edecrin Danocrine ..... Dacriose Elase ...:-vsu-s Alidase Dantrium ...... Danthron Elavil ........» Aldoril Daranide ...... Daraprim Elavil .....:0:0 Mellaril Dancon .....:- Darvon Emetine ....... Emetrol Darvon-N ...... Darvocet-N Enarak ......o: Marax Decadron ...... Decaderm Endecon ...... Edecrin Decadron ...... Percodan Endecon ...... Enduron Decagesic ..... Duragesic Enderin ....... Empirin Decholin ...... Daxolin Enduron ...... Enderin Delalutin ...... Deladumone Endure ...... Eutron Delta-Dome .. . Deltasone Enduron ...... Imuran Demerol ....... Demulen Ephedrol ...... Tedral Demerol ....... Dicumarol Equagesic ..... Decagesic Demerol ....... Dymelor Esimil .... 0a Estinyl Demerol ....:«. Pamelor Esimil ........ Ismelin Deprol ...:s4. Demerol Eskatrol ....... Hexadrol Desferal ....... Disophroi Estiny! ........ Estomul Desipramine ...Deserpidine Estomul ....... Esimil Desoximetasone . . Dexamethasone | Estomul ....... Ismelin Desoxyn ...:::.: Digitoxin Ethabid ....... Ethamide Ethamide ...... Ethionamide Ethinamate ....Ethamide Ethionamide ...Ethinamate Burax .......c. Urex Euthroid ...... Thyroid Evtonyl on Eutron F Felsules ....... Feosol Feosol ..:vv.v. Feostat Feostat ....... Feostim Fer-in-Sol ..... Feosol Festal ......... Feosol Fiogesic ....... Wygesic Flagyl ...convm. Flexical Flexeril “....... Flexical Fluocinonide ...Fluocinolone Folbesyn ...... Fulvicin Fostex ......vn pHisoHex Furgoin ....... Fulvicin G Gamastan ..... Garamycin Ganatrex ...... Kantrex Gantrisin ...... Gantanol Garamycin ..... Terramycin Gelfoam ....... Ger-O-Foam Geritol ........ Cheracol Gevral ........ Gevrine Glucagon ...... Glucoron Glutethimide ...Guanethidine H Haldol ........ Winstrol Haldrone ...... Haldol Halodrin ....... Haldrone Halog .uiocorein Haldol Halog ......... Mycolog Halotestin ..... Halothane Halotex ....... Halotestin Harmonyl ...... Hormonin Hexadrol ...... Hexaderm Hexadrol ...... Hexalol Hexalol- ....... Hexestrol FHISprl .cusmcnn Hiprex Homapin ...... Hormonin Hycodan ...... Hycomine Hycodan ...... Vicodin Hycomine ..... Hycodan Hydropres ..... Catapres Hydropres ..... Diupres Hydroxyzine ....Hydroxyurea Hygroton ...... Hykinone Hypersal .....:« Hyperstat Hyperstat ...... Hyper-Tet ‘Hyperstat ...... Nitrostat Hytone ......-.» Hytrona Hytone .... svn Vytone 1 llosone ........ lonosol Imipramine ....Imferon WOPARY ate svi dis Imferon Inderal ........ Enderin Inderal ........ Imuran Inderal ........ Isordil Indocin ....... Lincocin Indocin ....... Minocin Intropin ....... Ditropan 85 Isopto Carpine ..lIscpto Eserine Isordil ........ Isuprel Isuprel ........ Ismelin K Kafooin ....... Keflin +. (SR Kaolin Keflex .....ovn Keflin Keflex ........ Kelex Kemadrin ...... Coumadin Ketalar ........ Kenalog Kel®F ovine Kaochlor KehOr ocuvsman Klor L Lactinex ....... Lactocal LBrOSIA -uivoive imo Lomotil LasiX ..ouimsmen Esidrix LASIX ..vvwsnan Laxsil Levophed ...... Levoprome Levorphano! ... Levallorphan Lidaform cess Vioform LIGBX : ovine Lasix LIGONG ooiniv nn Dilone Lincocin uuu. Cleocin Loriding ....... Leritine Luminal ....... Tuinal Lurie ices Loryl M Maalox .....0e Camalox Maalox ........ Maolate MBRIOX .u.uiuve Marax Mebaral ....... Mellaril Mebaral ....... Tegretol Medrol ........ Mebaral Mellaril ....... Meltrol Mellaril ....... Moderil Mephenytoin . Mesantoin Meprobamate . ..Mepergan Meprobamate . Meperidine Mestinon ...... Mesantoin Metandren ..... Metahydrin Metatensin ..... Mesantoin Metatensin ..... Mestinon Metaxalone . Metolazone Methadone . Mephyton Methyldopa . Levodopa Metrazol ...... Mintezol Milontin ....... Miltown Minocin. ....... Niacin Mithramycin . Mitomycin Mobidin ....... Moban Modane ....... Matulane Modane ....... Mudrane Molindone ..... Mobidin Mutamycin ..... Mity-Mycin Myciguent ..... Mycitracin Mylanta ....... Milontin Myleran ....... Mylaxen Myleran ....... Mylicon Mylicon ....... Modicon N Nardil oun Norinyl NeDOIN viuinrs Nebs Negatan ....... NegGram Nembutal ...... Myambutol Neosorb ....... Neosone 86 Nicobid ....... Nitro-Bid Niconyl ....... Nicolar NICO) ovis Niconyl Nilstat ........ Nitrostat Nilstat ........ Nystatin Nitroglycerin ... Nitroglyn Nitrospan ...... Ditropan Nitrospan ...... Nitrostat Nitrospan ...... Nystatin Nitrostat ....... Nystatin NOHINYl answeas Voranil Norlestrin ...... Novabhistine Norlutin ....... Norlutate Norpramin ..... Imipramine 0 Omnipen ...... Unipen Orasone ....... Oracin Orenzyme ..... Avazyme Oretic ........ Oreton Orinase ....... Orenzyme Orinase ....... Ornade Ormex ........ Orexin Ornex ........ Orinase Ornex ........ Ornacol Ornex ........ Ornade Orthoxicol ..... Ornacol Otobiotic ...... Otobione Otobiotic ...... Urobiotic OVIal Lovomomse Uval ovulen .....uen Ovral Oxymetholone ..Oxymorphone P Pabalate ...... Robalate Pamelor ....... Dymelor Pantholin ...... Pathilon Pantopon ...... Parafon Pantopon ...... Protopam Paral ......... Parasal Paregoric ...... Percogesic Parest ........ Trest Pathocil ....... Pathilon Pathocil ....... Pitocin Pathocll ....... Placidy! Pavabid ....... Paveril Pavased ....... Pavabid Pavulon ....... Povan Penicillin ...... Polycillin Pentothal ...... Pentritol Percodan ...... Percobarb Percodan ...... Percogesic Percodan ...... Percorten Periactin ...... Percodan Periactin ...... Persantine Periactin ...... Taractan Peristim ....... Persantine Peritrate ....... Lotusate Peritrate ....... Pentryate Peritrate ....... Periactin Persantine ..... Persistin Persantine ..... Trasentine Pertofrane ..... Persantine Phenaphen . Phenagesic Phenaphen . Phenergan Phenelzine ..... Phenylzin Phenergan ..... Theragran Phenobarbital . Pentobarbital Phenoxene . Phenolax Phentermine . Phentolamine Phospholine . Phosphaljel Piperazine ..... Piperacetazine Pitocin ........ Pitressin Ponstel ....... Pronestyl Pramoxine ..... Pralidoxime Prazepam ..... Prazocin Prednisone . Prednisolone Presamine ..... Apresoline Propoxyphene . Propadrine Protopam ...... Protamine Pyridium ...... Pyridoxine Q Quaalude ...... Quinidine Quarzan .......Questran Quinidine ...... Quinine Quintess ...... Quiess R Regroton ...... Hygroton Regroton ...... Regonol Reticulex ...... Reticulogen Rifadin ........ Ritalin Ritalin ........ Ismelin S SAN uuu ieee Saluron Sand! onan Tandearil Sanorex ....... Ser-Ap-Es Sansert ....... Singoserp Santyl ........ Sandril Sebical ....... Sebulex Sebutone ...... Sebulex Senokot ....... Senokap Ser-Ap-Es ..... Catapres Serax ......... Eurax Serax ......... Psorex SOAR - iain Xerac Serenium ...... Dyrenium Serentil ....... Saronil Serentil ....... Surital SHON * .voininnns Silain Simethicone .. Cimetidine Simron ........ Sintrom Sinarest ....... Allerest Singoserp ..... Sinequan Sinulin .... 000 Sonilyn Sinurex ....... Sinarest Somophyllin .. Slo-Phyllin Sorbitrate ..... Sorbutuss Sparine ....... Sterane Sporostacin .. Sporicidin Sterazolidin . Butazolidin Sterazolidin . Stelazine Sudafed ....... Sudolin Sulfamethazine . Sulfasalazine . Sulfamethizole . Sulfathalidine Surfak =... Surbex Synalar ....... Synasal T Tage ..eowsvens Tao Taractan «uc es Tinactin Tedral ous vs ve Teldrin Tegopen ...:.. Tagamet Tegopen ...... Tegretol Tegopen ...... Tegrin Tegretol ....... Tegrin Temaril ....... Demerol Tepanil ....... Demerol Tepant ....... Temaril Tepanil ....... Terfonyl Tepanil ......e Tofranil Terionyl ux. 0:: Tofranil Testolactone. . . .. Testosterone Theolix ....... Theolixir Thiamine ...... Thiomerin Thiamine ...... Thorazine Thyrar o.com Thyrolar Thyrar ooo en: Tryptar Thyrolar ....... Theolair Wigan ...c:nieis Ticar Tobramycin .... Trobicin Tolinase ..,..Toleron Torecan ....... Toleron Triamcinolone .. Triaminicin Triaminic ...... Triaminicin Triclos ........ Tricon THCONO! cones Tricofuron Tridemic ...... Triaminic Trophite ....... Troph-iron TONG] .ovunvos Tylenol uU Unipen ........ Unicap Unipen ........ Urispas Unitensen ..... Salutensin Uracel .u.asaes Uracil rach ....v0ses Uracid Urised ........ Uracel Urised ........ Urestrin Urispas ....... Urised Urispas ....... Uristat URSA inne Uristix i} mr Valium ...vvniia Nalline Valmid ........ Valpin Valmid ........ Velban Valpin ccoviins Valium Vasodilan ...... Vasocidin Velban ........ Valpin Versapen ...... Verstran Verstran ....... Vastran Vesprin ....... Vastran Vicodin ....... Hycomine Vigran ........ Wigraine Viomycin ...... Vibramycin VION cccvvsvn Viteron Vitron-C ....... Vicon-C Vontrol - ....0e Vastran VBSOl cuerwrnins Vontrol Vytone ........ Vitron Ww rE Wyamine ...... Wydase Wycillin + ....... V-Cillin 2 ees Le Zactane ....... Zactirin Zactinin ....v.n Zarontin Zactifin ....... Zentron Zarontin vues Zaroxolyn Zarontin ....... Zentron Zentinic ....... Zymatinic Zetone ........ Zentron © Pharmacy Times Attachment 3. DRUG Fo®fM TI (Coder A) OATE ComPLETcEh Rox NumBER CooeR A ¥O®W CLERK L®T Loo Wuwber Pakieny Record Enlil ! Number TXew \\ a Tew Wb Example A | 1 23345 1234499 oo 2 0953S Example 8 A 0168S +—~ 2 [| 02%05 _ ERAN CL {.00350 | 23535 | okAasS fw JExavwg\e D eqadsSoe | \1\eS 20190 Exawmgle EI 13265 0% 140 ] 20545 2% 445 Example TF. \2R\S ___ | 1 en LoS\Ye | 208595 Example GI. Forsirmmmmebioses rie iss oe 025715 \ 4 340 VT ———— ——y ——— = AG + p— Ari sommes S&F 87 Attachment 4. | DRUG FofM I (CoderB) CooeR B HUST | DATE ComP\LETED . BOX NUMBER crews LBY | Pavien\ Boley | Loa) Wowbex Record [umber | Txew VA | Them Wb | _ E A BAS | sprang] 095%. | ——— ef 4 An 3 A 9 68 Attachment 5. DRUG Fofva I (AI udrcaNov) Rox NUMBER VATE ComeLETED ApsvuDrvcavoa DDE cteak_VW83I “oo PATIENT [Ey TYEM | a TYTeEM WW), Tn NUMBER | Record Nbr | Coder A [Coder B [Mivdica¥ion |coper A CooeR B |ndyudicalion. xawgle A | L q9999 | 444499 [99948 4 1 616%5 ]13%485 |016SS _ Exaile 2 [14340 [14360 [14340 - Appendix V. American Hospital Formulary Service classification system and therapeutic category codes AMERICAN HOSPITAL FORMULARY SERVICE CLASSIFICATION SYSTEM AND THERAPEUTIC CATEGORY CODES (AHFS#) (Classifications in parentheses are provisional but may be used in DPIF) AMERICAN HOSPITAL FORMULARY SERVICE CLASSIFICATION SYSTEM 04:00 ANTIHISTAMINE DRUGS 08:00 ANTI-INFECTIVE AGENTS 08:04 Amebacides 08:08 Anthelmintics 08:12 Antibiotics 08:12.02 Aminoglycosides 08:12.04 Antifungal Antibiotics 08:12.06 Cephalosporins 08:12.08 Chloramphenicol 08:12.12 Erythromycins 08:12.16 Penicillins 08:12.24 Tetracyclines 08:12.24 Other Antibiotics 08:16 Antituberculosis Agents 08:18 Antivirals 08:20 Plasmodicides 08:24 Sulfonamides 08:26 Sulfones 08:28 Treponemicides 08:32 Trichomonacides 08:36 Urinary Germicides 08:40 Other Anti-Infective 10:00 ANTINEOPLASTIC AGENTS 12:00 AUTONOMIC DRUGS 12:04 Parasympathomimetic Agents 12:08 Parasympatholytic Agents 12:12 Sympathomimetic Agents 12:16 Sympatholytic Agents 12:20 Skeletal Muscle Relaxants 16:00 BLOOD DERIVATIVES 20:00 BLOOD FORMATION AND COAGU- LATION 20:04 Antianemia Drugs 20:04.04 Iron Preparations 20:04.08 Liver and Stomach Preparations 20:12 Coagulants and Anticoagulants 20:12.04 Anticoagulants 20:12.08 Antiheparin Agents 20:12.12 Coagulants 20:12.16 Hemostatics 20:40 Thrombolytic Agents 24:00 CARDIOVASCULAR DRUGS 24:04 Cardiac Drugs 24:06 Antilipemic Agents 24:08 Hypotensive Agents 24:12 Vasodilating Agents 24:16 Sclerosing Agents 28:00 CENTRAL NERVOUS SYSTEM DRUGS 28:04 General Anesthetics 28:08 Analgesics and Antipyretics 28:10 Narcotic Antagonists 28:12 Anticonvulsants 28:16 Psychotherapeutic Agents 28:16.04 Antidepressants 28:16.08 Tranquilizers 28:16.12 Other Psychotherapeutic Agents 28:20 Respiratory and Cerebral Stimulants 28:24 Sedatives and Hypnotics 36:00 DIAGNOSTIC AGENTS 36:04 Adrenocortical Insufficiency 36:08 Amyloidosis 36:12 Blood Volume 36:16 Brucellosis 36:18 Cardiac Function 36:24 Circulation Time 36:25 (Cystic Fibrosis) 36:26 Diabetes Mellitus 36:28 Diphtheria 36:30 Drug Hypersensitivity 36:32 Fungi 36:34 Gallbladder Function 36:36 Gastric Function 36:38 Intestinal Absorption 36:40 Kidney Function 36:44 Liver Function 36:48 Lymphogranuloma Venereum 36:52 Mumps 36:56 Myasthenia Gravis 36:60 Myxedema 36:61 Pancreatic Function 36:62 Phenylketonuria 36:64 Pheochromocytoma 36:66 Pituitary Function 36:68 Roentgenography 36:72 Scarlet Fever 36:76 Sweating 36:78 (Thyroid Function) 36:80 Trichinosis 36:84 Tuberculosis 36:88 Urine Contents 40:00 ELECTROLYTIC, CALORIC, AND WATER BALANCE 40:04 Acidifying Agents 40:08 Alkalinizing Agents 40:10 Ammonia Detoxicants 40:12 Replacement Solutions 40:16 Sodium-Removing Resins 40:18 Potassium-Removing Resins 40:20 Caloric Agents 40:24 Salt and Sugar Substitutes 40:28 Diuretics 40:36 Irrigating Solutions 40:40 Uricosuric Agents 44:00 ENZYMES 48:00 EXPECTORANTS AND COUGH PREPARATIONS 52:00 EYE, EAR, NOSE AND THROAT PREPARATIONS 52:04 Anti-Infectives 52:04.04 Antibiotics 52:04.06 Antivirals 52:04.08 Sulfonamides 52:04.12 Misc. Anti-Infectives 52:08 Anti-Inflammatory Agents 52:10 Carbonic Anhydrase Inhibitors 52:12 Contact Lens Solutions 52:16 Local Anesthetics 52:20 Miotics 52:24 Mydriatics 52:28 Mouth Washes and Gargles 52:32 Vasoconstrictors 52:36 Unclassified Agents 56:00 GASTROINTESTINAL DRUGS 56:04 Antacids and Adsorbents 56:08 Anti-Diarrhea Agents 56:10 Antiflatulents 56:12 Cathartics and Laxatives 56:16 Digestants 56:20 Emetics and Anti-Emetics 56:24 Lipotropic Agents 56:40 Misc. GI Drugs 60:00 GOLD COMPOUNDS 64:00 HEAVY METAL ANTAGONISTS 68:00 HORMONES AND SYNTHETIC SUBSTITUTES 68:04 Adrenals 68:08 Androgens 68:12 Contraceptives 68:16 Estrogens 68:18 Gonadotropins 68:20 Insulins and Anti-Diabetic Agents 68:20.08 Insulins 68:24 Parathyroid 68:28 Pituitary 68:32 Progestogens 68:34 Other Corpus Luteum Hormones 68:36 Thyroid and Antithyroid 72:00 LOCAL ANESTHETICS 76:00 OXYTOCICS 78:00 RADIOACTIVE AGENTS 80:00 SERUMS, TOXOIDS AND VACCINES 80:04 Serums 80:08 Toxoids 80:12 Vaccines 84:00 SKIN AND MUCOUS MEMBRANE PREPARATIONS 84:04 Anti-Infectives 84:04.04 Antibiotics 84:04.08 Fungicides 84:04.12 Scabicides and Pediculicides 84:04.16 Misc. Local Anti-Infectives 84:06 Anti-Inflammatory Agents 84:08 Antipruritics and Local Anesthetics 84:12 Astringents 84:16 Cell Stimulants and Proliferants 84:20 Detergents 84:24 Emollients, Demulcents and Protectants 84:24.04 Basic Lotions and Liniments 84:24.08 Basic Oils and Other Solvents 84:24.12 Basic Ointments and Protectants 84:24.16 Basic Powders and Demulcents 84:28 Keratolytic Agents 84:32 Keratoplastic Agents 84:36 Miscellaneous Agents 84:50 Pigmenting & Depigmenting Agents 84:50.04 Depigmenting Agents 84:50.06 Pigmenting Agents 84:80 Sunscreen Agents 86:00 SPASMOLYTIC AGENTS 88:00 VITAMINS 88:04 Vitamin A 88:08 Vitamin B Complex 88:12 Vitamin C 88:16 Vitamin D 88:20 Vitamin E 88:24 Vitamin K Activity 88:28 Multivitamin Preparations 92:00 UNCLASSIFIED THERAPEUTIC AGENTS 94:00 (DEVICES) 96:00 (PHARMACEUTIC AIDS) Reprinted with permission of the American Society of Hospital Pharmacists, Inc., copyright holder DPIF. U.S. GOVERNMENT PRINTING OFFICE: 1982-361-161:510 Vital and Health Statistics series descriptions SERIES 1. SERIES 2. SERIES 3. SERIES 4. SERIES 10. SERIES 11. SERIES 12. SERIES 13. Programs and Collection Procedures.—Reports describing the general programs of the National Center for Health Statistics and its offices and divisions and the data col- lection methods used. They also include definitions and other material necessary for understanding the data. Data Evaluation and Methods Research.—Studies of new statistical methodology including experimental tests of new survey methods, studies of vital statistics collection methods, new analytical techniques, objective evaluations of reliability of collected data, and contributions to sta- tistical theory. Analytical and Epidemiological Studies.—Reports pre- senting analytical or interpretive studies based on vital and health statistics, carrying the analysis further than the expository types of reports in the other series. Documents and Committee Reports.—Final reports of major committees concerned with vital and health sta- tistics and documents such as recommended model vital registration laws and revised birth and death certificates. Data from the National Health Interview Survey.—Statis- tics on illness, accidental injuries, disability, use of hos- pital, medical, dental, and other services, and other health-related topics, all based on data collected in the continuing national household interview survey. Data From the National Health Examination Survey and the National Health and Nutrition Examination Survey.— Data from direct examination, testing, and measurement of national samples of the civilian noninstitutionalized population provide the basis for (1) estimates of the medically defined prevalence of specific diseases in the United States and the distributions of the population with respect to physical, physiological, and psychological characteristics and (2) analysis of relationships among the various measurements without reference to an explicit finite universe of persons. Data From the Institutionalized Population Surveys.—Dis- continued in 1975. Reports from these surveys are in- cluded in Series 13. Data on Health Resources Utilization.—Statistics on the utilization of health manpower and facilities providing SERIES 14. SERIES 15. SERIES 20. SERIES 21. SERIES 22. SERIES 23. long-term care, ambulatory care, hospital care, and family planning services. Data on Health Resources: Manpower and Facilities.— Statistics on the numbers, geographic distribution, and characteristics of health resources including physicians, dentists, nurses, other health occupations, hospitals, nursing homes, and outpatient facilities. Data From Special Surveys.—Statistics on health and health-related topics collected in special surveys that are not a part of the continuing data systems of the National Center for Health Statistics. Data on Mortality.— Various statistics on mortality other than as included in regular annual or monthly reports. Special analyses by cause of death, age, and other demo- graphic variables; geographic and time series analyses; and statistics on characteristics of deaths not available from the vital records based on sample surveys of those records. Data on Natality, Marriage, and Divorce.—Various sta- tistics on natality, marriage, and divorce other than as included in regular annual or monthly reports. Special analyses by demographic variables; geographic and time series analyses; studies of fertility; and statistics on characteristics of births not available from the vital records based on sample surveys of those records. Data From the National Mortality and Natality Surveys.— Discontinued in 1975. Reports from these sample surveys based on vital records are included in Series 20 and 21, respectively. Data From the National Survey of Family Growth.— Statistics on fertility, family formation and dissolution, family planning, and related maternal and infant health topics derived from a periodic survey of a nationwide probability sample of ever-married women 15-44 years of age. For a list of titles of reports published in these series, write to: Scientific and Technical Information Branch National Center for Health Statistics Public Health Service Hyattsville, Md. 20782 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES POSTAGE AND FEES PAID Public Health Service U.S. DEPARTMENT OF HHS Office of Health Research, Statistics, and Technology HHS 396 National Center for Health Statistics 3700 East-West Highway Hyattsville, Maryland 20782 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 THIRD CLASS E——— U.S.MAIL EE HRST From the Office of Health Research, Statistics, and Technology DHHS Publication No. (PHS) 82-1364, Series 2, No. 90 For listings of publications in the VITAL AND HEALTH STATISTICS series, call 301-436-NCHS U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES e Public Health Service ® National Center for Health Statistics ® Series 2, No. 9 Consent and Privacy in the National Survey of Family Growth: A Report on the Pilot Study for Cycle lll Data Evaluation and Methods Research Series 2, No. 91 SUGGESTED CITATION National Center for Health Statistics, K. Tanfer, W. Grady, and C. Bachrach: Consent and privacy in the National Survey of Family Growth: A report on the pilot study for Cycle Ill. Vital and Health Statistics. Series 2, No. 91. DHHS Pub. No. (PHS) 82-1365. Public Health Service. Washington. U.S. Government Printing Office, March 1982. Library of Congress Cataloging in Publication Data Tanfer, Koray. Consent and privacy in the national survey of family growth. (Vital and health statistics. Series 2, Data evaluation and methods research ; no. 91) (DHHS publication ; no. (PHS) 82-1365) Prepared by Koray Tanfer, William Grady, and Christine Bachrach. Bibliography: p. 1. Family size—United States—Statistical methods. 2. Sampling (Statistics) 1. Grady, William R. Il. Bachrach, Christine. Ill. Title. IV. Series. V. Series: DHHS publication ; no. (PHS) 82-1365. RA409.U45 no. 91 [HQ766.5.U5] 312'.0723s 81-607132 ISBN 0-8406-0240-5 [304.6'3] AACR2 1 For sale hy the Superintendent of Documents, U.S. Government Printing Office. Washington, D.C. 20402 Consent and Privacy in the National Survey of Family Growth: A Report on the Pilot Study for Cycle Ili This report describes the results of a pilot study for Cycle Ill of the National Survey of Family Growth. The report compares the effects on interview response and data quality of three pairs of alternative data collection procedures. Data Evaluation and Methods Research Series 2, No. 91 DHHS Publication No. (PHS) 82-1365 U.S. Department of Health and Human Services Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics Hyattsville, Md. March 1982 National Center for Health Statistics DOROTHY P. RICE, Director ROBERT A. ISRAEL, Deputy Director JACOB J. FELDMAN, Ph.D., Associate Director for Analysis and Epidemiology GAIL F. FISHER, Ph.D., Associate Director for the Cooperative Health Statistics System GARRIE J. LOSEE, Associate Director for Data Processing and Services ALVAN O. ZARATE, Ph.D., Assistant Director for International Statistics E. EARL BRYANT, Associate Director for Interview and Examination Statistics ROBERT C. HUBER, Associate Director for Management MONROE G. SIRKEN, Ph.D., Associate Director for Research and Methodology PETER L. HURLEY, Associate Director for Vital and Health Care Statistics ALICE HAYWOOD, Information Officer Vital and Health Care Statistics Program PETER L. HURLEY, Associate Director GLORIA KAPANTALIS, Assistant to the Director for Data Policy, Planning, and Analysis Division of Vital Statistics JOHN E. PATTERSON, Director ALICE M. HETZEL, Deputy Director WILLIAM F. PRATT, Ph.D., Chief, Family Growth Survey Branch MABEL G. SMITH, Chief, Statistical Resources Branch JOSEPH D. FARRELL, Chief, Computer Applications Staff Preface This report describes the results of a pilot study for Cycle III of the National Survey of Family Growth. It compares the effects of three alternative data collection procedures on inter- view response and data quality. The survey was designed and conducted by the Institute for Survey Research of Temple University, Philadelphia, Pa., under a contractual agreement with the National Center for Health Statistics. The alternative data collection procedures were designed by Koray Tanfer of the Institute for Survey Research in cooperation with William F. Pratt and Gerry E. Hendershot of the National Center for Health Statistics. Much of the report is based on the final re- port submitted by the Institute, and many of the tabulations in the report were prepared by Lee Robeson of the Institute. iii i wh = koneh NF i i LE » = i pt tT Yop o aims wd Shr = B - «oa ~ = n ® pn dl a I CT ir - - i 5 we 2 so. i Lv vo. 2 - : .l Dts rh ho. oo A ar nT Ep dl 4%, 6 Bs fm Td Spe - J igh 1, AERA ve narod a Jie gia . wed nll Afr ig SH ade ey er 23. lip a Ea r=) ii | week als i i tn a 8 i dink ; PERN AL hae Sr ea Contents Proface . . .... «op Co BB i 2 8 BEE Hawa ee fl mia 3 BRI RS BME er BSE IN AE 6 en nh pA eh TR 6 iii VALEORUCEIONS: oni ls ok HERES 0% BEET 02 SESH ns Sammie 2 0 EGRET S38 BERD EE L210 wg x wis ewe gol Boar Re 8 1 Purpose of the Cycle IN PHOLSIURY = vs ssanm ds vmnsns s sonar ss mms 4 v3 FEAWE ES BELEWS 8 0 wwndisy vomy 1 Data collection ProcBOures BSIBd. ..... . cobvm ds rumms ws Sammy s+ 3 wan s +s mand ss ERB HB APB HERENE 2 2 Summary of principal FINAINGS. . . cov v sti naai dss sums sas smmis ss swibmat is sinmEi ss ERTS Vr b MET ES» 3b 4 Source and limitations of the data . . . . . «i ee 7 SAMI dosiOn. « . cs vesmi i is TREES t 8 HEHE E Aly 5 REE rs FARE SEEN SIAMESE ER SEE Ee es HARE ee 7 Sample disposition BRA SUIVEY FESPONSE . + lv a5iv & As 5 & 26m iv 5 5 80 ben w 5 5 a 5 i 6 00m 0 ww bb BFE GRE EEE Bw 7 ASSIGNMENT LO TrEAIMBNT GIOUDS. + « + + » « + + «vs 2 + vt vss tv sv sessresrbnrsnrsnorrnerensnressersnnsssss 8 DAO IHTIEOLIONS + & 20 9is 5% 0% m 3% 5% 3 wa 5 ae % EE a EE RHEE EY EER AES EEE EE ew a 8 BOCUESUE his 3355s. i un aries a i co 59 1 on Ge on om PT) 0 0 5 [5 090 oF 15 0 9 Toe rn oe Lew 4 rt wg, ht oF 0 FORE (MEE 10 AOE, ACE, BN SUIVEY GIBB. + + + + + + vv vv vv vv vd vv v vw wwe vosiossssssvsnssersvnsevevesnnernrrwersn 10 Amount of prior INformMation. . . LL eee 10 Parent QUESTIONNAITE . . . oo ot 11 Tyne of INMErvIeW adImINISTration ; » os ww = v5 Liners ow dw ww oF ono ms ov 5 ow WR ES 5 EE HEANE DEES SBE Ew GA 12 BEIBIBIACERI CS : + mon wie +n ww eves om om wn eden me #0 40 Be 0 0 GE Be Jo 0 8 J S20 HEE) 1 0 02 wh we 1 ct ans om wt fo Fon IB OF Uriel 3 15 List of etallod 1ables. . . « vo noiv ss vs umm s ss vu tme ss wo wiwin s ¥ 2B PEE + +3 HHS RE $$ 8 THEE 2 BFE WL 2 ww. 16 Appendixes COBIBNES 0 +s vt EF 5 0 5 6. TERR + 6 § SBE EEF HH es 5s A&E Es HEE ES AMEE I] HERE MEE sd ME Cs 28 I. Definitions of certain terms used IN this TEDOTL + 2% « + « 5 simm sw + ¢ + swan es s» siuonms s LETH HS) FMFE Ls n@dwe 2a 29 Il. Advance letter, pamphlet, and flip charts constituting the prior information provided to respondents . . . ........... 31 HL. Parent QUESTIONNAITE . oo ott tt tt tt te tt ee ee 41 IV. Selected questions from the self-administered questionnaire (SAQ) and the interviewer-administered questionnaire (IAG... isos spr asia da ARR E ES NEHER 4 BEETS ES anew nimi Es haa ERT REET EE SESE 46 List of text figures 1. Interview refusal rates by amount of prior information and age . . . . . .. . «i iit iii 4 2. Interview refusal rates for minor women 15-17 years of age by source of refusal, whether a parent questionnaire was ad- ministered, and race. . ...... AB 2 TE TR EE YR RE i BEE TEER EE ERA EAE een a wh 5 3. Percent of respondents 15-17 years of age and parents who provided neither an exact amount nor a range in response to questions on family INCOME. . . LL. i tt tt te te ee ee 5 4. Interview nonresponse rates by reason for NONresponse and race. . . . . . «cv it ii ii ee 6 5. Interview nonresponse rates by reason for nonresponse and age . . . . LL... ee ee 6 6. Treatment and control groups case assignment design. . . . ov iit tt it tt ee 8 vi Symbols Data not available Category not applicable Quantity zero Quantity more than zero but less than 0.05 Quantity more than zero but less than 500 where numbers are rounded to thousands Figure does not meet standards of reliability or precision Consent and Privacy in the National Survey of Family Growth: A Report on the Pilot Study for Cycle lil by Koray Tanfer, Ph.D., Institute for Survey Research, Temple University, and William Grady, M. A., and Christine Bachrach, Ph.D., Division of Vital Statistics Introduction The primary mission of the National Center for Health Statistics is to collect and publish data relating to the health of the population of the United States. In carrying out this mission, the Center collects data on vital events registered in the United States, con- ducts inventories of health facilities and manpower, and conducts probability sample surveys based on household interviews, health examinations, and med- ical records. Data collection programs are supple- mented by research projects to investigate new tech- niques of data collection and evaluate operating programs. In response to the need for current information on the interrelated topics of fertility, family planning, and their effects on population growth, the National Survey of Family Growth was established as an inte- gral part of the Center program in 1971. The National Survey of Family Growth is a cyclic survey; that is, data are collected every few years by means of a sam- ple survey. The first cycle of the survey was con- ducted in 1973, the second was conducted in 1976, and Cycle III is being conducted in 1982. The sample design and data collection for Cycle I of the National Survey of Family Growth were con- tracted to the National Opinion Research Corpora- tion of the University of Chicago. Interviews were completed with 9,797 women from July 1973 through February 1974. For Cycle II of the National Survey of Family Growth, the sample design and data collection were contracted to Westat, Inc., of Rock- ville, Md. The Cycle II sample consisted of 8,611 women with whom interviews were completed from January 1976 through September 1976. The target population of Cycle I and Cycle II was the civilian household population of women 15-44 years of age living in the conterminous United States who were currently or previously married or were never-married mothers with offspring living in the household at the time of the interview. Data were collected by means of personal interviews with prob- ability samples of these women. The interviews fur- nished information for determining trends and differ- entials in fertility, family planning practices, sources of family planning advice and services, effectiveness and acceptability of various methods of family plan- ning, and aspects of maternal and child health that are related closely to family planning and child- bearing. Purpose of the Cycle III Pilot Study Cycle III of the National Survey of Family Growth will be the first cycle to include a sample of women of reproductive age (defined to be 15-44 years) regardless of marital status. All never-married women will be eligible for inclusion in the sample, rather than only those with offspring living in the household at the time of interview (as in previous cycles). The potential sensitivity of interviews with women who have never married (especially women who are minors) on the topics covered in the survey raised the question whether it is feasible for the Fed- eral Government to conduct such interviews. If so, special procedures to minimize the sensitivity of the interview and to maximize survey response and data quality needed to be tested. The feasibility of interviewing adolescents who had never married was demonstrated in three national surveys of young women conducted by researchers at Johns Hopkins University,!-2:3 as well as in other studies of adolescents based on more selective sam- ples. However, methodological issues in interviewing never-married women have received little attention in the literature. A notable exception is DeLamater and MacCorquodale, who examined the effects of ques- tion location and type of interview administration on the reporting of sexual behaviors; however, their study was based on a sample of young, white men and women in a single Midwestern city. The pilot study for Cycle III of the National Sur- 1 vey of Family Growth (NSFG) was designed to test the feasibility of conducting interviews under the auspices of the Federal Government on topics such as fertility, family planning practices, and maternal and child health with never-married women 15-44 years of age. A major objective of the pilot study was to compare three alternative procedures for ob- taining optimal response rates and ensuring data quality. The pilot study was conducted under con- tract by the Institute for Survey Research of Temple University. This report details the results of the pilot study. Definitions of terms used in this report are found in appendix I. Data collection procedures tested Three pairs of alternative data collection proce- dures were tested in the pilot study for Cycle III. One pair of the procedures tested the effect on the intci- view refusal rate of the amount of prior information provided to the respondent as a basis for informed consent to the interview. A second pair of procedures tested the effectiveness of administering a parent questionnaire in obtaining parental consent to inter- view a minor (15-17 years of age). The third pair of procedures tested the relative efficacy of two forms of interview administration (interviewer-administered compared with self-administered) in obtaining infor- mation on sensitive topics from the respondent. Amount of prior information.—The National Sur- vey of Family Growth is required to provide enough prior information to each respondent to obtain an “informed consent” to the interview. The informa- tion provided should allow the respondent to make a decision about participation that is based on knowl- edge of the nature of the survey and the right to refuse to participate. The amount of prior information supplied to the respondent may affect the survey response rate and the quality of data collected in several ways. Supply- ing complete and detailed information about the sur- vey may reduce the likelihood of refusal by increasing the respondent’s interest and curiosity and creating an atmosphere of trust. It also may reduce the likelihood of misreporting and nonresponse on sensitive questions by providing assurances of confi- dentiality and uses of the data obtained. On the other hand, it may be that the more infor- mation the respondent is given, the greater the likeli- hood that the respondent would find some aspect of the survey threatening, that interest would be dimin- ished by the lengthy explanation, or that she would feel she did not know enough to participate in the survey. For the pilot study for Cycle III, all women in the sample were mailed a letter that contained general in- formation about the NSFG, the sample selection process, confidentiality of responses, the purpose of 2 the survey, and the voluntary nature of participation. The women also received a second introduction to the survey from the interviewer that included a pamphlet and a short, standard verbal presentation. In addition to this basic information, half of the women in the sample were given supplemental infor- mation by the interviewer. The supplemental infor- mation consisted of a flip-chart containing 10 graphs depicting the types and uses of the data sought in the interview (appendix II). In both instances, the infor- mation was supplied before attempting to conduct an extended (main) interview. The research question addressed by this procedure was whether the addi- tional amount of prior information provided to the respondent affected the interview refusal rate. Providing information about the interview serves as a basis for informed consent as well as a means of obtaining respondent cooperation. The pilot study also explored the question of how much information is necessary before the respondent feels adequately informed. All respondents were asked at the end of the interview whether they had been ‘told enough about what the interview would be like.” The re- sponses of women who had received only the basic in- formation were then compared with the responses of those who had been given both the basic and the sup- plemental information. Parent questionnaire. —Whenever the eligible respondent was a never-married minor, signed paren- tal consent was requested in addition to the verbal consent of the respondent. The necessity of obtaining the consent of a parent (or guardian) may increase the likelihood of an interview refusal for two reasons: (1) two persons must agree to the interview rather than one, and (2) parents may be reluctant to expose an adolescent daughter to any interview, or to an interview about fertility-related behaviors. High rates of interview refusal, in turn, increase the likelihood of a selection bias, that is, bias result- ing from differences between the total group of eligible women and the subset of women who com- plete the interview. The pilot study was designed to test a strategy to reduce the likelihood of interview refusal among never-married women and their parents. The strategy tested in this study was administration of a short interview with a parent (the mother whenever possi- ble) before parental consent was requested. This brief interview elicited information on the mother’s child- bearing and on socioeconomic characteristics such as education and family income. The parental interview may reduce refusal rates for two reasons: (1) the parent becomes a participant in the survey, thus increasing his or her psychological stake in its outcome; and (2) it provides a mechanism to develop rapport between the parent and the inter- viewer. On the other hand, it is possible that the con- tent of the questionnaire, such as questions on family income, would be considered too sensitive and have an adverse effect on the parent’s willingness to pro- vide consent, thus increasing refusal rates. The parent questionnaire is shown in appendix III. The parent questionnaire was administered (after the prior information was given and before consent was requested) in half of the households in which a never-married minor was identified as the eligible re- spondent. In the remaining such households, consent was requested immediately after the prior informa- tion was given. Interview refusal rates then were com- pared for these groups. The parent questionnaire treatment also served another research function. Because data on the socio- economic characteristics of the parents were collected in both the parental and respondent interviews, a crude indication of accuracy of family background information reported by minor respondents could also be obtained by comparison. Interview administration. —Questions about sexual activity and other fertility-related topics may be espe- cially sensitive for never-married women. Because verbalizing responses to sensitive questions may be embarrassing or threatening to these respondents, the likelihood of item nonresponse and misreporting may be great in interviews requiring oral responses. Al- though the use of “answer cards,” which require only letter or number responses, may alleviate this prob- lem, the number of cards that may be used is limited. In an attempt to partially avoid these problems, some surveys have used a self-administered questionnaire to elicit information on sensitive topics. This approach offers the respondent greater privacy than when oral answers are required and may be associated with more candid and complete responses. However, the additional privacy afforded by the self-administered questionnaire also may affect data quality. This method does not allow as much com- plexity in the design of the questionnaire as question- naires for oral responses do (that is, it requires less complex skip patterns) and also does not permit interviewer intervention for missing, incomplete, or inappropriate responses. Furthermore, the quality of data obtained from a self-administered questionnaire depends on the literacy and educational level of the respondent. In the pilot study, half of the respondents received interviewer-administered questions only, and half received a combination of interviewer- administered questions and self-administered ques- tions. The self-administered portion of the interview, which covered potentially threatening or sensitive questions, was given after approximately 20 interviewer-administered questions and was followed by 40 to 85 additional interviewer-administered ques- tions. The content and design of both interview pro- cedures were similar, with minor format changes to facilitate self-administration. Selected questions from the self-administered questionnaire and interviewer- administered questionnaire are shown in appendix IV. Responses to the sensitive questions were compared for the two groups with respect to (1) frequency of item nonresponse, and (2) aggregate distribution of responses to each item. At the end of the interview, the respondents from each group were asked whether any of the questions had been “hard or uncomfortable to answer” and whether they thought they might have preferred the form of interview administration that they had not received. Responses to these questions provided an indication of the effect of type of interview adminis- tration on the respondent’s comfort -with the inter- view. Summary of principal findings Three data collection strategies were tested in the pilot study. Two strategies, provision of supplemental information about the nature and uses of the survey and administration of a short interview with a parent of minor respondents, were tested to determine their efficacy in reducing interview refusal rates. The third strategy, the self-administered questionnaire, was tested to determine its effects on data quality. Provision of supplemental information was asso- ciated with a reduction in the refusal rate for women 18-44 years of age but not for minor women 15-17 years of age, who had the highest refusal rate of any age group (figure 1). It also had little effect among black women but resulted in a reduction of more than 3 percentage points among women of other races. Thus supplemental information about the sur- vey yielded a small reduction in refusal rates but was not effective among all women. Administration of a parent questionnaire reduced refusal rates by more than 5 percentage points among minor women. Although interviewing a parent had almost no effect among black women, the refusal rate among women of other races was reduced almost 7 percentage points when a parent questionnaire was administered (figure 2). This reduction is particu- larly important because without a parental interview the refusal rate for women of other races was 19.9 percent compared with only 5.1 percent for black women; thus the procedure was most effective in the racial group for which refusals were greatest. The parental interview also had an important effect on obtaining information on family character- istics. Only approximately 46 percent of minor women provided any information on family income, but 84 percent of parents provided this information in response to questions asked during the parental interview (figure 3). Parents were also more likely to provide data on the educational attainment of the father than minor women were. A parental interview thus provides an effective strategy to improve survey response and availability of background information for never-married minor women. RN Basic information only 1 Basic and supplemental information 20 I 15 = 14.2 135 = \ 11.1 11.0 c S 10 | \ [1] a. 5.6 3.4 | 7 15-17 years 18-19 years 20-44 years Age NOTE: See appendix | for definitions of terms. Figure 1. Interview refusal rates by amount of prior information and age The major strategy for improving data quality that was tested in the pilot study was the use of a self-administered questionnaire to obtain information on sensitive topics. It was thought that self- administration might reduce response distortion for sensitive questions by providing the respondent with greater privacy than is afforded by interviewer admin- istration and by reducing the risk of “courtesy re- sponses’ (answers the respondent believes conform to the interviewer’s or society’s values). However, com- [1] Respondent refusal NR Parent refusal 25 ee 19.9 8.3 15 w 13.1 c 2 a Ny 4.6 10 \ 5.1 11.6 \ 5 fs 8.5 3.4 AN Nu \ 0 ZN AN A NN No parent Parent No parent Parent question- question- question- question- naire naire naire naire Black Other races NOTE: See appendix | for definitions of terms. Percent Parents Respondents Figure 2. Interview refusal rates for minor women 15-17 years of age by source of refusal, whether a parent questionnaire was administered, and race parison of responses obtained from self-administered questionnaires with responses obtained from interviewer-administered questionnaires did not sup- port this expectation; the distributions of responses were similar for both questionnaire forms. However, greater item nonresponse was found in the self- administered questionnaire, especially for open-ended questions. The pilot study results thus provided no evidence that response distortion is reduced when sensitive questions are asked using a self-administered form, but the results did indicate that greater item nonre- sponse rates are associated with this procedure. Given that the results obtained from a self-administered questionnaire are to some extent dependent on the complexity of the questionnaire design and the lit- eracy of the respondents, this questionnaire form ap- pears to entail several costs with no apparent gains in data quality. The combined response rate for the pilot study of 70.4 percent is the product of a screening response rate of 88.2 percent and an interview response rate of 79.8 percent. Much of the nonresponse may be attrib- uted to two factors: the timing and the duration of . the field period. August and September, when the fieldwork was carried out, are associated with high Figure 3. Percent of respondents 15-17 years of age and parents who provided neither an exact amount nor a range in response to ques- tions on family income population mobility, which reduces the probability of finding respondents at home. This problem was compounded by the characteristics of the study pop- ulation (young, never-married women are highly mobile) and by the short field period of 4 weeks, which reduced the number of possible calls. The ef- fects of these factors were evident in high screener and interview nonresponse rates due to reasons other than refusal. Figures 4 and 5 show that interview refusal rates varied by race and age. Black women identified as eligible for the study were less likely to refuse the in- terview than eligible women of other races (figure 4), resulting in a lower overall interview nonresponse rate among black women. Interview refusal and overall in- terview nonresponse rates were greater among eligible women 15-17 years of age, for whom parental con- sent for the interview was required, than among older women (18-44 years of age), for whom parental con- sent was not necessary (figure 5). Rates of nonre- sponse for reasons other than refusal varied little by race and age. Item nonresponse rates for sensitive questions about pregnancy and family planning were generally very low; among respondents given the interviewer- administered questionnaire, nonresponse was zero for [1 Interview refusal NJ Interview nonresponse for reasons SN other than refusal 25 ir 216 20.2 20 p= 15 p= 11.3 14.1 13.2 c 8 o Qa 53 10 = 1 \ \\ Black Other races Race NOTE: All races includes women for whom race was not stated. See appendix | for definitions of terms. 1] Interview refusal [nterview nonresponse for reasons NN other than refusal 2 225 20 = la 13.8 Re 155 § 3 8.6 7. 10 NEN 4 MN A NR 8. \ NN 18-19 20-44 years years Age NOTE: See appendix | for definitions of terms. Figure 4. Interview nonresponse rates by reason for nonresponse and race most items and never exceeded 2 percent for any item. Furthermore, approximately 71 percent of respondents given the interviewer-administered ques- tionnaire found none of the questions hard or uncom- fortable to answer, indicating that response distortion due to question sensitivity is probably not large. The results of the pilot study demonstrated the feasibility of including never-married women in the NSFG and of asking them potentially sensitive ques- tions about topics such as fertility, family planning, and maternal and child health. The survey response rate was acceptable given the timing and duration of Figure 5. Interview nonresponse rates by reason for nonresponse and age the field period, and the item response rate for sensi- tive questions was very high. The study also showed that the parental interview is an effective procedure for reducing nonresponse and enhancing data quality for never-married minor women, who are an impor- tant target population of the Cycle III survey. The results further indicate that survey refusals can be . reduced among never-married adult women (18-44 years of age) and among women of other races by giving them supplemental information about the sur- vey before attempting an interview. Source and limitations of the data The sample design and fieldwork for the pilot study of Cycle III were contracted to the Institute for Survey Research of Temple University, Philadelphia, Pa. The sample consisted of 759 eligible women, of whom 606 (79.8 percent) were interviewed; of the 606 interviewed women, 347 were 15-17 years of age, and 259 were 18-44 years of age. All interviews were conducted during August and September 1979. Sample design The sample was designed to broadly represent the civilian noninstitutional population of never-married women 15-44 years of age living in households and group quarters in the conterminous United States. The sample was selected using a five-stage design but, because the study was not intended to obtain na- tional estimates of population characteristics, it was not a strict probability sample. The first stage of the sampling process resulted in selection of four primary sampling units. The four areas were purposely chosen to provide variation in geographic region, level of urbanism, and racial com- position, as well as some variation in age structure and income level. The sample areas comprised the central city and suburban portions of a large North- eastern standard metropolitan statistical area, the urban portion of a small Southern standard metro- politan statistical area, and a rural Southern area (composed of two rural counties). When aggregated, the population of the four areas was similar to that of the national population with respect to the charac- teristics on which they were chosen. Within each of the first-stage sample areas, strict probability sampling rules were observed. The second and third stages of the sampling process resulted in selection of 48 small geographic areas (listing areas), 12 from each primary sampling unit. Selections at both stages were made with probabilities propor- tionate to size (number of dwelling units). In addi- tion, the second-stage selection of census tracts and enumeration districts used stratification by race and income to ensure that the sample remained broadly representative by those characteristics. The fourth stage of sampling consisted of the selection of dwelling units within listing areas. Be- cause more treatments applied to minor women (15- 17 years of age) than to adult women (18-44 years of age) (see section on ‘‘Assignment to treatment groups”), and because minor women were an impor- tant target population for the study, the study design specified that two-thirds of the approximately 600 interviews were to be completed with minor women and the remaining interviews with adult women. Thus because only about one-third of never-married women 15-44 years of age are minors, minor women had to be sampled at a greater rate than adult women. These different sampling rates were achieved during the fourth stage of sampling by randomly designating a portion of the dwelling units in the sample listing areas as subsample units. In these units (identified for the interviewer by a pink screener interview form), interviews were to be conducted only with an eligible minor. In the remaining households (assigned blue screener interview forms), any eligible woman, either minor or adult, could be interviewed. When more than one eligible woman was identi- fied in a household, all eligible women were listed on the screener interview form, and one woman was selected randomly. This constituted the fifth stage of the selection process. In subsample units, only minor women were eligible for this operation. Sample disposition and survey response Table 1 shows the final disposition by survey area of dwelling units assigned for listing during the fourth stage of sampling. Examination of the table shows that of the 8,442 dwelling units assigned, 703 were either vacant, were not dwelling units as defined by the NSFG, or were outside the listing areas. Of the remaining 7,739 units, 6,826 were successfully 7 screened, yielding a screener response rate of 88.2 percent (table 2). Only about 20 percent of the dwell- ing units not successfully screened were missed be- cause of refusals; the remaining portion of screener nonresponse was primarily a result of unsuccessful attempts to locate anyone eligible for the screener interview at home during the study period. Screening identified 759 women eligible for the extended interview (excluding adult women in sub- sample units, for which only minor women were eligible to be interviewed, and excluding women liv- ing in multiple-eligible households who were eligible but not selected). Among the eligible women, 606 completed an interview, producing an interview re- sponse rate of 79.8 percent (a discussion of inter- view nonresponse appears in a later section of this report) and an overall response rate (the product of the screener and interview response rates divided by 100) of 70.4 percent. The overall response rate varied by survey area, ranging from a low of 65.2 percent in the urban South to a high of 78.8 percent in the rural South. Although refusal to participate in the survey was a factor in producing the low overall response rates, three other factors were also very important: (1) timing of the survey (during the summer months when seasonal mobility is high), (2) composition of the study population (predominantly young, never- married women, who are highly mobile), and (3) short duration of the field period. Assignment to treatment groups The major objective of the pilot study was to ex- amine the effects of the alternative interviewing pro- cedures on response rates and data quality. There- fore, it was important that the characteristics of the respondents in each treatment cell (figure 6) be equal within the limits of random sampling error. This was necessary to limit the possibility that the effects of the treatments would be confounded with the effects of the characteristics of the respondents. Respondent assignment to treatment groups was accomplished after the fourth-stage selection of addresses was completed. Starting with a randomly selected address in each listing area, addresses sys- tematically were assigned to one of the eight treat- ment cell combinations. This assignment of cases en- sured a random distribution of respondents among treatment combinations and avoided spot assignment by the interviewers. Because cases were assigned to treatment cells before contact was made with the sample households, households containing eligible women 15-44 years of age were designated to receive the parent questionnaire. However, this treatment was carried out only when the selected respondent was 15-17 years of age, as a part of the procedure for obtaining parental consent. The outcome of the assignment of women to treatment groups is shown in tables 3, 4, and 5. Table 3 shows numbers of eligible and responding women by amount of prior information received, according to age, race, and survey area; table 4 shows numbers of eligible and responding minor women by whether a parent was interviewed, according to race and survey area; table 5 shows the number of responding women by type of interview administered, according to age, race, and survey area. Data limitations The pilot study was to provide information about the effectiveness of various survey procedures that would be applicable to a survey of the national popu- lation. For reasons of cost and efficiency, however, the sample design employed to select pilot study re- spondents was not a national probability sample. Therefore, strictly speaking, the results of the study cannot be generalized for the national population. However, the four areas selected as sites for the pilot study were chosen to be broadly representative of the national population; that is, the distribution of the study populations as a whole by characteristics such as age, race, and income was similar to that of the Nation (according to 1970 census data). Therefore, the results of the study, although not precisely gen- eral for the national population, will provide informa- tion of value in planning a national survey. Although the four pilot study sites were chosen Basic information only Basic and supplemental information Type of interview administration No parent questionnaire Parent questionnaire (minor women only) Parent questionnaire (minor women only) No parent questionnaire Interviewer-administered ® ® ® ® Self-administered @ ® ® Figure 6. Treatment and control groups case assignment design to be broadly representative of the national popula- tion, the respondents in the pilot study differ from single American women of reproductive age in their distribution by age and race. According to data col- lected in the March 1979 Current Population Survey, approximately 17 percent of never-married women 14-44 years of age were black, and approximately 40 percent were under 18 years of age.> Among pilot study respondents (15-44 years of age), these figures are 29 percent and 57 percent. In interpreting study results, overrepresentation of minor women and black women should be taken into account. Therefore, wherever the number of cases allows, results are shown separately by age and racial group. Most results shown in this report are given in the form of percent distributions and simple cross tabula- tions. Multiple classification analysis also was used to statistically adjust the report findings for age, race, and survey area but, because the adjusted results were virtually identical to the unadjusted findings, these data are not presented. Interactions between treat- ments also were explored by observing whether the effects of one treatment were similar within cate- gories of other treatments. The analysis yielded no evidence of such interaction effects. Because a strict probability sample was not used in the pilot study, no statistical tests of group differ- ences in rates or percents are reported in the analysis of results. Statistical tests based on an assumption of simple random sampling were calculated for use as a rough guide to the analysis. Results Age, race, and survey area Interview nonresponse rates, refusal rates, and rates of nonresponse for reasons other than refusal are shown in table 6, according to survey area, race, and age. Nonresponse rates ranged from 15.4 percent in the rural South to 26.1 percent in the urban South. Interview refusal was more common and constituted a greater proportion of total interview nonresponse in the South than in the Northeast. The high levels of nonresponse for other reasons in the two North- eastern areas sampled may reflect some disguised refusal,” as, for example, respondents not keeping appointments or respondents deliberately staying away from home. Another factor that may contribute to geographic differences in nonresponse for other reasons is variation in seasonal mobility by area, which would result in differing proportions of eligible women not at home. Interview nonresponse rates were lower among eligible black women than among eligible women of other races, primarily because black women were less likely to refuse the interview. Rates of nonresponse for other reasons are similar for the two racial cate- gories. Age variations in interview nonresponse rates are in part a result of the requirement for written paren- tal consent for interviews with minor respondents. Interview nonresponse rates ranged from 22.5 percent among women 15-17 years of age to 15.5 percent among women 20-44 years of age. The refusal rates among women in these age groups were 13.8 percent and 7.1 percent. However, because of the require- ment for parental consent, each interview with an eligible minor had two potential sources of refusal— the parent and the minor. When the refusal rate for women 15-17 years of age is broken into its two components, parental and respondent refusals (8.5 percent and 5.4 percent), the resulting rates of respondent refusal are similar to those observed in the older age groups. Although parental consent was not required for 10 respondents 18 years of age and over, a small number of parents did intervene and refuse to allow their daughters to participate. Three parental refusals occurred among women 18 or 19 years of age, but none occurred among women 20-44 years of age. After accounting for the effect of parental refusal on response rates, age made little or no difference in the willingness of eligible women to participate in the study. Amount of prior information An examination of table 7 reveals that refusal rates among women 18-44 years of age were lower for those who received basic and supplementary informa- tion about the survey (3.9 percent) than among those receiving only basic information (11.0 percent). Among women 15-17 years of age, however, provi- sion of supplementary information had virtually no effect on refusal rates (14.2 compared with 13.5). The absence of a difference among women 15-17 years of age results from the different effect of the supplemental information on minor women than on their parents; although the supplemental information reduced parental refusals from 9.2 percent to 7.8 per- cent, respondent refusals increased from 4.4 percent to 6.4 percent at the same time. Provision of supplementary information also re- duced the refusal rate among women of other races by 3.3 percentage points. This difference probably is understated because of the overrepresentation in the sample of women 15-17 years of age for whom the supplementary information had no effect. Legal and ethical considerations require that respondents be given enough information about an interview to allow them to make an informed choice about participation in the study. However, the amount of information needed as a basis for informed consent is difficult to determine. In an effort to address this issue, pilot study respondents were asked at the end of the interview whether they thought they had been told enough about what the interview would be like. Table 8 shows the percents of respond- ents who answered “yes,” “no,” and “not sure” or “don’t know” to this question. More than four-fifths (82.8 percent) of the re- spondents felt they had been told enough about the interview. Among those who did not answer yes, nearly two-thirds were not sure. Approximately 6 percent of the respondents felt they had not been given enough information. Table 9 shows the percent of respondents who answered yes to this question according to the amount of prior information given. This percent is similar for respondents who received the supple- mental information before the interview (84.3 per- cent) and for respondents who were given the basic information only (81.4 percent). Similar results were obtained when the relationship between amount of prior information and the likelihood of respondents reporting they had been told enough about the inter- view was examined in each survey area and race and age group shown in table 9; in most cases, the differ- ences are small, and none are larger than might be expected by chance in samples of this size. Another issue addressed in the pilot study was whether the provision of supplemental information about the nature of the questions to be asked would more adequately prepare the respondent for sensitive topics in the interview and make these topics less threatening or embarrassing to the respondent. To gather information on this issue, all respondents were asked at the end of the interview if any of the ques- tions had been “hard” or “uncomfortable” to answer. Table 10 shows that about a quarter (25.7 percent) of the pilot study respondents answered yes to this question, and that there was little variation in this percent by the amount of prior information received. When the relationship between the amount of prior information and the percent answering yes was ex- amined within categories of race, age, and survey area, the only substantial difference occurred among residents of the urban Southern area (table 10). Parent questionnaire When a designated respondent was under 18 years of age, interviewers were instructed to obtain written consent of the parent to interview the daughter. In approximately one-half of the cases, a brief interview with the mother concerning her own childbearing and socioeconomic characteristics was to be conducted before her consent to interview the daughter was re- quested. The main objective of this procedure was to test its effect on the likelihood of parental refusal. The procedure also allowed the comparison of infor- mation on family characteristics given by minor re- spondents with that obtained from their parents. Table 11 shows interview refusal rates by whether a parent questionnaire was used, according to survey area and race. Because the parent questionnaire was used only for eligible women under 18 years of age, this table excludes women 18-44 years of age. The results in table 11 indicate that a smaller proportion of parents and minor respondents refused to participate when a parent questionnaire was administered than when it was not used. Approxi- mately 1 in 6 (16.3 percent) of the respondents in the “no parent questionnaire” group refused to be interviewed (or parental consent was denied), com- pared with approximately 1 in 9 ( 11.1 percent) of the respondents or parents in the “parent question- naire” group. Furthermore, although the parent questionnaire was designed to reduce refusals among parents, daughters of parents who were given the questionnaire were only about half as likely to refuse the interview as their counterparts in the “no parent questionnaire’ group (3.4 percent compared with 7.1 percent). Table 11 also shows that the effects of the parental interview on the refusal rate varied by race. The parent questionnaire had little effect among black women, but among women of other races it was associated with a reduction of about 7 percentage points. A substantial reduction (8.7 percentage points) also was found among residents of the suburban Northeast. However, the observed differ- ences among residents of other areas are too small (given the small sample size) to support any state- ments that the procedure was effective in reducing refusals in those areas. The parent questionnaire procedure also allowed a rough assessment of whether complete and accurate information on family characteristics could be ob- tained from minor respondents. During the interview, respondents were asked two questions about family income, one question about the education of their fathers, and one question about the education of their mothers. The same questions were asked of the parent as part of the parent questionnaire. The percents of parents and minor respondents giving answers to the questions and the distributions of responses given by parents and respondents then were compared. Minor respondents may have difficulty providing accurate answers to questions on family characteris- tics for several reasons. A minor’s knowledge of family income and parental education often depends on what he or she is told by the parents. Some minor respondents may be unable to answer the questions because they never were told the information. Fur- thermore, because the information may be less impor- tant or meaningful to minor respondents than to their parents, they may not recall what they have been told or may remember it incorrectly. Table 12 shows the percent of minor respondents and their parents who answered ‘don’t know,” did not answer, or refused to answer questions on family 11 income, mother’s education, and father’s education. Two questions on family income were asked. The first asked for the exact dollar amount. If a response that could be coded was not given to the first question, the respondent was asked to identify a range within which her family income fell. Table 12 shows the percent not answering each of these questions as well as the overall proportion answering neither question. Minor respondents were more than 3 times as likely as their parents to provide no information on family income; approximately 54 percent of the respondents compared with 16 percent of the parents did not report either an exact amount or a range for income. Respondents were also more than twice as likely as their parents to provide no information on fathers’ education (approximately 18 percent com- pared with approximately 8 percent). However, there was little difference between parents and respondents in the likelihood of reporting mothers’ education. In table 13, the distribution of responses to these questions on family characteristics given by parents is compared with that given by their daughters. Differ- ences in the distributions may be the result of several factors—misreporting by minor respondents, misre- porting by parents, and bias resulting from the exclusion of persons who did not answer the ques- tions. Thus the comparisons in table 13 provide only a crude indication of the level of misreporting by minor respondents. Table 13 shows that the distributions of responses given by minor respondents are similar to the distributions of parental responses. Minor respond- ents were somewhat more likely than their parents to report an exact family income of $25,000 or more (38.5 percent compared with 29.4 percent) but have the same distribution when reporting in either exact amounts or categories. Minor respondents were more likely to report 12 years of education for mothers (49.1 percent compared with 42.8 percent), but these differences are not large and are based on small numbers of women. The data thus suggest that the major problem in collecting family background infor- mation from minor respondents is the large propor- tion of women who are unable or unwilling to answer the questions. The pilot study also addressed the question of whether use of the parent questionnaire affects the cost of data collection. Because the questionnaire required only a short time to administer, the amount of interviewing time was not expected to differ substantially between the “parent questionnaire’ and “no parent questionnaire’ treatments. However, if additional visits to households in the “parent ques- tionnaire” group were needed to find the parent at home at a convenient time for conducting the interview, data collection costs might be affected. As table 14 shows, 64 percent of interviews with minor respondents were completed within two calls, and 92 12 percent were completed within four calls, when a parent questionnaire was not used. When a parent questionnaire was administered, these percents were only slightly less (59 percent and 88 percent). Type of interview administration For about half of the pilot study respondents, the entire questionnaire was administered by an inter- viewer. In the alternate procedure, a portion of the questionnaire containing the most sensitive questions was given in a self-administered schedule. It was thought that the greater privacy afforded by the self-administered questionnaire (SAQ) might result in less response distortion (misreporting) and greater comfort with the sensitive questions asked. On the other hand, the greater control over the interview situation in the interviewer-administered question- naire (IAQ) group was expected to result in lower levels of item nonresponse. The respondent did not know the type of interview administration at the time of the initial contact; therefore, the interview refusal rate was not used to compare the two procedures. In the absence of accurate measures on the sensitive topics included in the pilot study, the presence or absence of response distortion cannot be determined directly. An indirect test is possible, however. Because respondents were assigned ran- domly to SAQ and IAQ treatment groups, the two groups may be expected to be similar with respect to most characteristics. If the lack of privacy in the IAQ leads respondents to alter answers to sensitive ques- tions, the distributions of responses to these ques- tions should differ between the SAQ and IAQ groups. Specifically, IAQ respondents would be expected to give ‘“‘courtesy’” responses more frequently than respondents answering the SAQ. “Courtesy” re- sponses are answers given to conform with percep- tions of the interviewer’s or society’s values or expectations rather than with the respondent’s actual behavior. Table 15 compares the responses of SAQ and IAQ respondents to selected sensitive items in the ques- tionnaire. The distributions of responses for women in the two treatment groups were similar. In only two instances are the differences as large as 8 percentage points: among ever-pregnant women, the SAQ re- spondents were 8.2 percentage points less likely than IAQ respondents to report the use of contraception at first intercourse and 8.3 percentage points less likely to report that their first pregnancy ended in abortion. However, these differences remain well within the range of sampling variability. Therefore, the overall pattern of similar responses by IAQ and SAQ respondents gave little evidence of response distortion resulting from the interviewer-administered questionnaire. To examine the effect of type of interview administration on the extent of nonresponse to sensitive items, the proportions of respondents giving no answer to selected items were compared for the “ SAQ and TAQ groups. These data are shown in table 16. For respondents given the IAQ, percents of respondents giving no answer are zero for 10 of the 12 items shown in the table. For the remaining two items, only a small percent of respondents, 1.0 percent for use of contraception at first intercourse and 2.0 percent for age at first intercourse, gave no answer to the question. However, among those given the SAQ, percents giving no answer were zero for only 2 of the 12 items and ranged from 0.7 percent to 18.2 percent for the remaining 10 items. Table 16 shows percents of respondents giving “no answer” responses only, because “don’t know” is a valid answer to the items shown. However, in some cases, “don’t know” may be a disguised refusal, that is, a way to avoid answering the question. No “don’t know” responses were given by SAQ respondents to the items in the table; four were given by the IAQ group (two to whether currently pregnant and two to age at first intercourse for ever-pregnant women). No refusals were given by either group to the items in the table. Even when these types of item nonresponse are taken into account, rates of nonresponse were sub- stantially greater in the self-administered question- naire than in the interviewer-administered question- naire. Open-ended questions were particularly liable to nonresponse in the SAQ. As table 17 shows, approxi- mately 1 in 5 SAQ respondents gave no answer to questions on the reason for not using a method of contraception at first intercourse (20.0 percent) and on the reasons for currently skipping use of contra- ceptives (17.8 percent). The proportion of IAQ respondents who gave no answer to these questions was 1 percent or less in both cases. IAQ respondents were more likely to respond to the two preceding questions that they did not know the reason or that there was no reason; such responses were given by 11.2 percent and 3.4 percent of IAQ respondents, compared with 4.7 percent and 2.2 percent of SAQ respondents. Nevertheless, the proportion giving spe- cific answers that could be coded was 12 and 17 percentage points greater in the interviewer- administered questionnaire than in the self- administered questionnaire. In view of the privacy afforded by the SAQ, respondents given it might be expected to be more comfortable in answering questions on sensitive topics than those given the IAQ. Table 18 shows the proportions of SAQ and IAQ respondents who, at the end of the interview, reported finding any of the questions hard or uncomfortable to answer. Of the IAQ respondents, about 29 percent reported finding questions hard or uncomfortable, compared with approximately 23 percent of the SAQ respondents. This small difference was maintained for each geographic area and for each race and age group. However, the magnitude of the difference varied considerably, the greatest being in the suburban Northeast and among women 20-44 years of age. In these two subgroups, the proportion of respondents finding questions hard or uncomfortable to answer were about 15 and 11 percentage points greater among those given the IAQ than in the SAQ group. Another question was asked at the end of the interview to determine whether respondents in the IAQ and SAQ groups would have preferred the method of interview administration they had not received. The SAQ respondents were asked, “Would you have preferred if. an interviewer asked those questions to you, instead of filling out the question- naire yourself?” The IAQ respondents were asked, “Would you have preferred to answer some of the questions by filling out a questionnaire yourself?” Tables 19 and 20 show the answers to these questions according to survey area, race, and age. In both groups, most respondents answered no to the question, indicating a preference for the method of interview administration they had received; about 66 percent of the SAQ respondents and about 51 percent of the IAQ respondents answered no. Also, almost 3 times as many [IAQ respondents stated a preference for the SAQ (39.4 percent) compared with SAQ respondents who preferred the IAQ (13.7 percent). However, in interpreting these results, two factors should be noted. First, of the two groups of women, only the SAQ respondents were interviewed with the alternative questionnaire form. Second, a large proportion of women in each treatment group responded “don’t know” or “not sure” to these questions. A final consideration in evaluating the alternate types of interview administration is the length of time needed to complete an interview. Interview length affects respondent burden as well as the average cost of an interview. It was expected that use of the self-administered questionnaire would result in a longer average interview time, due to respondent difficulties in reading and understanding questions and in following instructions. Table 21 shows mean interview length by type of interview administration, survey area, race, and age. As table 21 shows, the mean length of interview was almost identical for the two procedures. Inter- views that included the SAQ averaged 39.4 minutes; interviews administered entirely by an interviewer required an average of 39.6 minutes to complete. Differences in mean interview length for SAQ and IAQ also were small for each of the survey areas and racial groups shown in table 21. When differences in interview length were examined by age, however, some differences were observed. Among the youngest respondents (15-17 years of age), SAQ interviews averaged approximately 3 minutes longer to complete than [AQ interviews. Among those 18-44 years of age, however, mean interview length was approxi- 13 mately 4 minutes longer for the IAQ group than for the SAQ group. Several factors may explain the longer average length of IAQ compared with SAQ among respond- ents 18-44 years of age. Respondent sophistication and experience with self-administered questionnaires may have alleviated the problems of using an SAQ. Also, as discussed previously, item nonresponse was 14 substantially greater for the self-administered ques- tionnaire. Giving no answer shortens the length of the interview, particularly when open-ended questions are asked. Finally, when questions are answered, absence of interviewer intervention when an inappropriate or incomplete answer is given probably contributes to shortening interview time. References IK antner, J. F., and Zelnik, M.: Sexual experience of young unmarried women in the United States. Fam. Plann. Perspect. 4(4): 9-18, Oct. 1972. 2Zelnik, M., and Kantner, J. F.: Sexual and contraceptive experience of young unmarried women in the United States, 1976 and 1971. Fam. Plann. Perspect. 9(2): 55-71, Apr. 1977. 3 Zelnik, M., and Kantner, J. F.: Sexual activity, contraceptive use and pregnancy among metropolitan-area teenagers: 1971-1979. Fam. Plann. Perspect. 12(5): 230-237, Sept.-Oct. 1980. 4DeLamater, J., and MacCorquodale, P.: The effects of interview schedule variations on reported sexual behavior. Sociol. Methods Res. 4(2): 215-236, Nov. 1975. 5U.S. Bureau of the Census: Marital status and living arrange- ments, March 1979. Current Population Reports. Series P-20, No. 349. Washington. U.S. Government Printing Office, Feb. 1980. 15 List of detailed tables 10. 11. 12. 16 Number of dwelling units, by survey area and sample dispo- BILION . ov cme» 5 5 0% Waa 3 4» ww win vee a § Response, nonresponse, and refusal rates by survey area and type of rate Number of eligible and responding women, by amount of prior information, age, race, and survey area Number of eligible and responding minor women 15-17 years of age, by whether a parent questionnaire was administered, race, and SUIVEY area . . . . « « «vv vv vee ee Number of respondents by type of interview administration, age, race, and survey area Number of eligible women and interview nonresponse rates, by reason for nonresponse, age, race, and survey area Interview refusal rates by amount of prior information, age, race, BNC SUIVBY Brea « 4 «x + + 5 5 5 8 EWE 5 & 5 2 Ba @ + % 3 Number and percent distribution of respondents, by response to the question “Do you think that the letter and the pamphlet we gave you told you enough about what the interview would be like?" Percent of respondents who answered yes to the question “Do you think that the letter and the pamphlet we gave you told you enough about what the interview would be like?" by amount of prior information, age, race, and survey area . . Percent of respondents who answered yes to the question “Did you find any of the questions hard or uncomfortable to answer?’’, by amount of prior information, age, race, and survey area Interview refusal rates of eligible minor women 15-17 years of age, by whether parent questionnaire was administered, source of refusal, race, and survey area Number of responses given by minor respondents 15-17 years of age and their parents to selected questions on family char- 17 17 18 18 19 20 20 21 21 22 13. 14. 15. 16. 17. 18. 19. 20. 21. acteristics and percent ‘‘no answer,” ‘don’t know,” or refusal, by type of respondent and question Number and percent distribution of answers given by minor respondents 15-17 years of age and their parents to selected questions on family characteristics by question, according to typeofrespondent . . . .. .... Number of interviews with minor respondents 15-17 years of age and cumulative percent distribution by number of calls necessary to complete interview, according to whether parent questionnaire was administered . . . . . ............. Number of respondents asked about selected sensitive charac- teristics and percent reporting characteristic, by type of in- 1erVIeW adMINISTEatION +. ov 2 4 v 5 iv 5 3 3% 4 5 56 @ #4 * & Number of respondents asked about selected sensitive charac- teristics and percent giving no answer, by type of interview administration Number and percent distribution of responses to selected open-ended items by type of response, according to type of interview adminISIration . .. . « + & + vw es 2 «cs 5 ww wie 4 & 4 Percent of respondents who answered yes to the question "Did you find any of the questions hard or uncomfortable to answer?”’, by type of interview administration, age, race, ANI SUIVBY BBE . « «vv ov 5 4 54 SH Wi BB § 5 5 WW BEE ee Percent distribution of respondents given the self- administered questionnaire by preference for interviewer- administered questionnaire, according to age, race, and sur- vey area Percent distribution of respondents given the interviewer- administered questionnaire by preference for self- administered questionnaire, according to age, race, and sur- NEY BIB. vv + + % # MBH iH 4 2 3» WTAE * 5 0» wm wth Mean length of interview in minutes of respondents, by type of interview administration, age, race, and survey area. . . . . 22 23 23 24 25 25 26 26 27 Table 1. Number of dwelling units, by survey area and sample disposition [See appendix | for definitions of terms] All survey Survey area Sample disposition areas Suburban Genial Urban Rural Northeast ory. South South Northeast Number of dwelling units Total dwellingunits assigned .... cc icnssims sans sass snes 8,442 1,379 2,358 2,583 2,122 Vacant, not a dwelling unit, outside listingarea . . .............. 703 89 198 195 221 Dwelling units eligible for screening... .................... 7,739 1,290 2,160 2,388 1,901 Refused screening : : «ws sass ss FFs BE RMT 3 AREF BET I mn. 183 26 93 35 29 Other Screening NONYBSPONSE: . + + + wiv 4 ws 3s EF sR £93 046 1 38% 53 730 127 255 246 102 Number of dwelling units for which a screener was completed . . . ... 6,826 1137 1,812 2,107 1.770 Number of dwelling units with an eligiblewoman .............. 759 196 204 203 156 Interview refused cai inE rT TARR TREE TIARA TART IRE FE 86 17 15 34 20 Other Interview NONTESPONSE ... & + ix «vx iss % wis 2% B16 os 3a 6m 55 DW 6 67 19 25 19 4 INISIVIBWICOMPIBIOT. oo v0 vs wie ion nia wis i minim wii an orcas a: 0 606 160 164 150 132 Table 2. Response, nonresponse, and refusal rates by survey area and type of rate [See appendix | for definitions of terms] Survey area All survey Type of rate areas Suburban Central Urban Rural Northeast ny South South Northeast Percent SCresnING rESPONSE . vv + vse sows saws suv Fro ss S80 SE HES E63 88.2 88.1 83.9 88.2 93.1 Screening refusal . . . .... Lee 24 2.0 4.3 1.6 1.5 Other Screening NONIeSPONSE . . «vv vv vv vt eee eee eee ees 9.4 9.8 11.8 10.3 5.4 INLErVIBW YBSPONSB . ovis sia sis sma sw iH as dais FEE HT HWY 28 Was ans 79.8 81.6 80.4 73.9 84.6 IRSIVIBWITBTUSA , vv vc sv nims sms sma sme sims some AEs 3 HEY 11.3 8.7 7.4 16.7 12.8 Other interview NONIeSPONSE . . . «vv vv vv ieee eee ee eens 8.8 9.7 12.3 9.4 2.6 Overall response ©... 70.4 71.9 67.5 65.2 78.8 1The combined response rate is the product of the screening response rate and the interview response rate divided by 100. 17 Table 3. Number of eligible and responding women, by amount of prior information, age, race, and survey area [See appendix | for definitions of terms] Amount of prior information AR, 1308; ariel Stvey area Total Basic information Basic and supplemen tal only information Eligible Responding Eligible Responding Eligible Responding Number of women Allwomen! ........................ 759 606 385 307 374 299 Age ABA7°YBAIS ws su un smms sams sw mi ima od 448 347 229 184 219 163 VEAL YBAIS: os viv mis vv nis sw vd mas smss op 307 259 154 123 1563 136 1819vyears ..........ovinnnnnn. 81 68 45 35 36 33 20-44 Yeals vv nus a na snes snws 226 191 109 88 117 103 Race Black wos emwss mms sdne swans ee sw mone s 206 177 107 92 99 85 Otherraces ............uuuuuunnnnnn 547 429 277 215 270 214 Survey area Suburban NOPIHEESE . cv cox ss ss nwt nmm si 196 160 96 81 100 79 Central city Northeast ................. 204 164 98 77 106 87 Urban South ....................... 203 150 108 79 95 7 RUPBESOUII. «vo. vm wn ww 5 oni S03 3090 00 156 132 83 70 73 62 Tincludes 4 women for whom age was not ascertained and 6 women for whom race was not ascertained. Table 4. Number of eligible and responding minor women 15-17 years of age, by whether a parent questionnaire was administered, race, and survey area [See appendix | for definitions of terms] Total No parent questionnaire Parent questionnaire Race and survey area Eligible Responding Eligible Responding Eligible Responding Number of women Allminorwomen .................... 448 347 240 178 208 169 Race Black ius :niv iss vies ign sinew ras 114 94 59 48 55 46 ORCL YBORS: 15 aim vv ww 0 or 0 wes ot 2 tet 4 0c 0 334 253 181 130 1563 123 Survey area Suburban Northeast .................. 124 100 65 50 59 50 Central city Northeast ................. 91 72 49 36 42 36 Arbon South uel swms swe mma swatsis pins 135 95 70 47 65 48 RuralSouth ........................ 98 80 56 45 42 35 18 Table 5. Number of respondents by type of interview administration, age, race, and survey area [See appendix | for definitions of terms] Interview administration Age, race, and survey area Total ! Interviewer- Self- administered administered Number of respondents ANTOSPONABMS + ,u5: avs sns tne ssnssnnnsami|emms vues spe omns 606 310 293 Age TET VY EBIS & «15: 5 onion 7 2s 4 5050 To 3 50 2620 4 A oh So 50 vo so oe om ime 3) wn mn #2 347 172 174 VBRAYBAIS « vs sit ss Hrs hE IAT RR IARI EAN RNS AER TEE EEE 259 138 119 1B-TOVBAS ss ss ns cn ss Tr INE I SMT ss mals wwe s@ ws es mw 68 33 35 DO-AAYOAIE + 205: n vim meson iors ow ow 0 rw owe 0 ow a on 0 a ae ae 191 105 84 Race Black ..... mG AR HA AEE RE BAIR RN AB RIAA REINA IRAE IEE 177 91 86 Otherraces ....: cums savor sans samssmus envy iwsls smuenonsowrymws 429 219 207 Survey area Suburban Northeast . . . . ..... iit ee eee 160 79 80 Cantral City NOTHOasT .. . . vo oo viv saws smn sms duals dims s 409 .6® Es 200s 164 92 7 LUrbanSouth . . .. cess smaisvinss sms vans ams ewe ews ama Ee ew 150 78 71 Burgi South : wus twos ims ram ds emms emmy swme rainls rowafs hms wuuis mms 132 61 Fi) Tincludes 3 women assigned to self-administered questionnaire but given interviewer-administered questionnaire. Table 6. Number of eligible women and interview nonresponse rates, by reason for nonresponse, age, race, and survey area [See appendix | for definitions of terms] Reason for nonresponse Number Age, race, and survey area of women Total Refusal Other reasons Percent All eligible women LLL 759 20.2 11.3 8.8 Age VBI GUE + vos 2050s 2k Bh 5 60H 4% 90 6 8 HPA HAE 4 80 6 WH BE NE EE 448 22.5 213.8 8.7 ABIAANBAIE + 5 v4 45 5% 3.550% 5:8 0 9 3 51 91306 580 0 oe 3 307 15.6 37.5 8.1 EVI OBEE! ro. mn sb i 9 om 0 te scm A mt oes nimi Fn 81 16.0 48.6 7.4 Er Ppp 226 15.5 71 8.4 Race BIBgK © visi os i ei Eni nie ea ni AE SEY AP ALRBI LET FARE REL SHED 206 14.1 5.3 8.7 OtNBrraces ......s cn vvsmresiassnmassnssmergansemmens ns ssmsssmess 547 21.6 13.2 8.4 Survey area Suburban Northeast . ...ccs vss waen sss EL BE PH AH ANS BABE BHR BENE SH 0 196 18.4 8.7 9.7 Central city NOrthBast . . co su vi vs sas sams $4 bo CRE HTS £00 TFET EHS $ 6008 ¥ 204 19.6 7.4 12.3 IPDIOR SOUT ois wim: we 81015 0 90 on ico voncws os wa cms ci wns iw ow Senn it me 0 mm ow ot 203 26.1 16.7 9.4 BRUPBI SOU &. 210: a: 5: wim cn enim ives lh Aus 9 a, 50s seo, 5 fs 0 ap 0 THE 1 0 ok a hee wc RA 9 8 156 15.4 12.8 2.6 Tincludes 4 women for whom age was not ascertained and 6 women for whom race was not ascertained. 2Respondent refusal = 5.4 percent; parent refusal = 8.5 percent. 3Respondent refusal = 6.5 percent; parent refusal = 1.0 percent. 4Respondent refusal = 4.9 percent; parent refusal = 3.7 percent. 19 Table 7. Interview refusal rates by amount of prior information, age, race, and survey area [See appendix | for definitions of terms] Amount of prior information Age, race, and survey area Basic Basic and Total information supplemental only information Percent AIL CHOIDIO WOIMBIY 4 4 xm uw uw oi wn 80 0% 0% 5 400 Bw 9) 0 18 5,5 Wi 0 08 RW 6% 5% 11.3 12.5 10.2 Age TERT YRS 55 055 ws. 1 00 50 00 0 2 A 5010 se in 1 3, ms set 13.8 113.5 214.2 IB AR YAS. ov vv viv vn iw a we EE WE RE EE EE RR WE SY RE 7.5 11.0 3.9 18 1OYBAIS .... ouvir ti rir raat ra srr aa 8.6 1.1 5.6 BOARNYONG 0 5 50 85 58 0 5 Bl ei 3 RI BR 010 0 0 0 0) Ss 0b 0 0 so i 1 on 7.1 11.0 3.4 Race Black Ls cvs vs da ESE RHEE ASE DEAE ARSENE PORE FER Ey RRA GE PE 5.3 5.6 5.1 OTST TABS co oe a ion uns ese fiat or odh tune tet tla re we Pek hg es wh Rao htt Fy Fh yt pt St ce soe 13.2 14.8 11.5 SUBUIDAT NOTIHBASE o.oo v vo inh 30 5a 210 fk 40 BE $F 05 05 50 405 E 200 0d 0m 5 000 0k es 8.7 8.3 9.0 Contral city NOFIHEaSE . vo iva sr me rues si E6 EH RE DERE SHANE RMT 2A0 53 WEE EER 90 7.4 10.2 4.7 LIrhan SOUT ou wv vn ov mm oes 3 bm oa 90 as £59 00 500 0 was 0m vw aE RE 5 RE $8 16.7 18.5 14.7 BO A SOR Le i ol vineienr ce. ikon rods: ioigk be eh ontario ands hme Ra i ro hoe ves aL rit rg Bahan so PH ede os uendhsr 12.8 12.0 13.7 TRespondent refusal = 4.4 percent; parent refusal = 9.2 percent. 2Respondent refusal = 6.4 percent; parent refusal = 7.8 percent. Table 8. Number and percent distribution of respondents, by response to the question “Do you think that the letter and the pamphlet we gave you told you enough about what the interview would be like?’ Response Neher Percent distribution respondents All reSPONAENTS . Ltt tt tt ee ee ee ee ee 606 100.0 eran sioubo mints: cose bests i 8 mie 03 5 AF AHF Ba 560508 5 tn wri Sg ou do oo or om: so md Fo non i 502 82.8 NO: sms sn FR AE IRR AE NINERS ERE LAR EHR RAS AREER SHAE Res AR EA EE EY Ee 37 6.1 Don’t KNOW, Not SUre, OF NOANSWEY + & ow wv o% «8 4s 95. 5 48 50% 0% 5 ES 85 57 $85 0588 2068 67 11a 20 Table 9. Percent of respondents who answered yes to the question “Do you think that the letter and the pamphlet we gave you told you enough about what the interview would be like?’’, by amount of prior information, age, race, and survey area [See appendix | for definitions of terms] Amount of prior information Age, race, and survey area Basic Basic and Total information supplemental only information Percent AI TOSDOIIIOITIS 5 4. 055 5 05 & 2 we Suir win: nies 929) sim cos cos om oko coo Lom ut ws ns et wt warms i 82.8 81.4 84.3 Age VBA TVBOES: 5 7 5 on 550 5 5 Bs 2 erst 5 Fld 09500 0 Beth vr mt EL cf cam vn 3) me 580 0 1 tn 82.1 81.5 82.7 VESAAYBAIE ov von vs btm db R ERT I AAI I HS CHE EEE IA GSB E TIRES FREI REE 3 FOE 84.2 81.3 86.8 TS TOYEAIS: vv vile wai S p02 30 TF FA #0 WT 3505 10% re 00 #1 5 ES A FCN 95.6 94.3 97.0 RO-ABYOOIS ov. 5 vio 4: 510s 0 £55 3 mw Fv m0 5 07050 0 To sem [com ow 80.1 76.1 83.5 Race BIagK: ssa as snes ims i Saini as san sshd AREA mA BEBE noms RSF AA © ER DE 83.1 84.8 81.2 OINBY FACRS ov owns pws ssn sR as PERF IRE SEER IRBs ERT SESE CEES E048 w 82.9 80.0 85.9 Survey area Suburban Northeast . ......... ee 78.1 79.0 77.2 Centra) City NOFEREaSE ... cov: smus soos sainsimmi des PEs SARE Emmys FED me sme s 86.0 83.1 88.5 IrDaR SON . ovis cvms sums sant dws s mms hE SHAT tems sRAE FW AE LHD EAB EE BEY 8 81.2 78.5 84.3 BUTBNSOUIIY 5. ; cv iinis sei ms oislos #5 a om ana comm sion m wife ore ww mea wim imin win x oa x G0 8 87.1 85.7 88.7 Table 10. Percent of respondents who answered yes to the question ‘Did you find any of the questions hard or uncomfortable to answer?"’, by amount of prior information, age, race, and survey area [See appendix | for definitions of terms] Amount of prior information Age, race, and survey area Basic Basic and Total information supplemental only information Percent All respondents . ..... 5 Aosk EE De THE BREE ERP 2 Be wees fe 0: OR rama Wed WAR BR 25.7 27.7 23.7 Age BAT YBBIS scr iamsm mr ish Rs STAR EESS EF HARES SHE LOH IRS SP IARNT IES IWF 5 25.1 27.7 22.4 VBA YBHES . ovviv iin vs mimi $0 05 RoE 5 5 0 5E SWE 508 ®t ¥ R 26.6 21.6 25.7 USITDIYBBES i oem iv 5o iw 5 29050 5 580 lor 97 08h 3 00 coc cc 0 0 0 0 RS 9 rw 20.6 22.9 18.2 DORA NY BANE ov in 5 505 05 9 in 8 3 om BL Br BE 58 BRR A BE AEE NRE EIB RARE 28.8 29.5 28.2 Race BIACK : vm sw men mmae mk B05 80 005055 EE 5 50 T ¢ gw aw wn or 3 ww rw a a 28.2 31.5 24.7 ORNOFTBBHE = im v1 550 6: 5.2% 1 6c 3 v5 58 3% 3 10 ot #0; = Fat 52 01 0 To ion 0 4 500 308 0 0 ob ks 10 8 4 0s 0s she 0 of oi 3 os 00 8 2 0 0 00 24.7 26.0 23.4 Survey area Suburban Northeast .... oh. .eeiiws: izmsssms saa ides Imms IFRS ERNE EEDIF TS 4 21.9 19.8 24.1 Central CItY NOrtheast . wiv snes +9 0s SHEL SHS FEES MNT LEFT DEAS 2 HRS B® 8 00000 8 24.4 26.0 23.0 LITBBNISOUHY. os snus sums ss 2 2555 c@a smh 5 mn alae ervions wus rune ames sues a 34.0 40.5 26.8 RUTBLSOUEN ccs omc ome nmisis msi smb ss wid smind Salis amis SHE HMme Smal A B55 4 22.7 24.3 21.0 21 Table 11. Interview refusal rates of eligible minor women 15-17 years of age, by whether parent questionnaire was administered, source of refusal, race, and survey area [See appendix | for definitions of terms] Race and survey area No parent questionnaire Parent questionnaire Source of refusal Source of refusal Total Parent £lig ible Total Parent £1ig id minor minor Percent All eligible minor WOMBN . ...:casvsvrsnms smn sans onns 16.3 9.2 7.1 111 7.7 3.4 Race BIBI: + uv wv: wiv 000 ar wien wr m0 wom ree i a ww 5.1 1.7 3.4 5.5 5.5 - CIBER IAEES onini 4 vuns 2408 5 83 $5 TS BE 3 HRs 5 Bd: Mid 8 19.9 11.6 8.3 13.4 8.5 4.6 Survey area Suburban Northeast, . . . cess sas nasi avws sve en es sous © 4 13.8 9.2 4.6 5.1 5.1 - Central City Northeast . .« cs vs ss snsr seas sos scans sve os 4.1 20 2.0 7:1 74 - Urban South... .... i eee 24.3 11.4 12.9 16.9 10.8 6.2 Bua SOU «ulus 5.6.0 4 30d ns 2a WI LBD S 580s Sie & 5d & Ble 2 4 19.6 12.8 7.9 14.3 79 71 Table 12. Number of responses given by minor respondents 15-17 years of age and their parents to selected questions on family characteristics and percent ‘no answer,” ‘don’t know,’ or refusal, by type of respondent and question [See appendix | for definitions of terms] Question Type of respondent Minor respondent Parent Minor respondent Parent Family income Number of responses Percent no answer, don’t know, or refusal RONDe OF XACT AMIOUINE & iv av inwiw s ww vor rowie 0 #% 9108 8 00% ¥ 9508 59 08 508 8 WF FE 169 EXACLRINOUBY. 0:51 on0 518% 502 00 015: 0 500000 Gh: 2 0 0hw. BAER 2: Bh alee Wh nmrelsl oc oan vo aoc hms Sco0) 169 BEN0R : + viv sm ss nh Ss BOE RARE BREE TIFT TRIE BERBER RAGE ENE AD RRAY 143 Education Mother's educational tIBINMEBNL ccs vas sass tus d sas Rass sr 0k 08a A006? 169 Father's educational talTNBIYE « oo vo wwe a aoe wow ene em we em owe ew 5 169 169 169 84 169 169 53.8 84.6 63.6 2.4 17.8 16.0 49.7 32.1 22 Table 13. Number and percent distribution of answers’ given by minor respondents 15-17 years of age and their parents to selected questions on family characteristics by question, according to type of respondent [See appendix | for definitions of terms] Type of respondent Question " a Minor Parent Minor Parent respondent respondent Family income Number of answers Percent distribution Total GIVING @XaCT AMOUNT . «ov vv vv et tees eae tet nae enn ne naan 26 85 100.0 100.0 Loss thaniSTB,000 ues vi sis vi 7 SHEE LHR SHAE EEA ® BEE Dw ww 9 31 34.6 36.5 BIB.000-20.:000 «cz oon vm icin 5 vi 9 we ivin ow 47 wis nwa ww win ik aL a 7 29 26.9 34.1 $25,000 OF NOTE. uv civ ov wvin iw wns 8 315 2500 hth @ SAW ATE SREE ART EHEC SEE 10 25 38.5 29.4 Total giving range Or BXECTAMOUNT uv ; seis & oe sie &% Hi # 4% 0% 5.0308 5 THA 5% 200 78 142 100.0 100.0 Less than BIB.000 = 05 ols «vibe <5 asim && 0 4 5 Sols Be SHAE TEAS LSE 8 30 54 38.5 38.0 SIB,000-20. 900 . coon vivas vias od 56 550 EG EEE BE AWE EE REE EEE 24 46 30.8 32.4 BB O0DIO TINOIE ain ve: wi reat 75003 1 0500 3 i 55 cw i ee i 0 30 91 a i sis 24 42 30.8 29.6 Mother's educational attainment Total GIVING ANSWEr 4: cuss sams smut sna TonE os L815 EE EE RE GE 165 166 100.0 100.0 LOSS tNBN T2YOBIS «vv sna vans SHRI AAD FC HRI NHBRC Tor BT PT ERE HEE ET 52 62 31.5 37.3 VDYBORE oo + iv et 000 00 4 0 9% 3 G0 EAR CAE SE A LH RR RRO 81 7M 49.1 42.8 TS YEAS OF MOMC . ov vv ttt tt vs neat annie assess snes ens snsnnnns 32 33 19.4 19.9 Father's educational attainment TOI GIVINGONSWOF 4. sos sa Rs FRA AT II TEAR BFR CRI SHE RHE SEES 139 155 100.0 100.0 LOSS ANGI VDIVOBES vi 0: 3 iwi amn @ #10 0 #1 is 0107 whi “5 wi ov 3 16 0 wc Sob Nw, 0 0 Cp 0 O06 4 G60 OH 1 51 60 36.7 38.7 V2 VRAIS oi iia 5 350 ow ok 3050 ot 04 530 B97 8 062 a8 81 3. fms cok to ros sod vm cw 0 fo ut re so 51 56 36.7 36.1 TS YBOISOrINOIS as inns smn sani FATIH E RE 5 ARE 8 anf Tod ow 37 39 26.6 25.2 TExcludes “no answer,” “don’t know,’’ and refusal responses. Table 14. Number of interviews with minor respondents 15-17 years of age and cumulative percent distribution by number of calls necessary to complete interview, according to whether parent questionnaire was administered Parent questionnaire Number of calls Not administered Administered Number ANINIEIVIBWS iu sn sims imma mois iRe ims Bar Am AE sma Ios Ss ms noms 178 169 Cumulative percent distribution Vums isms smmicmas sons ins i RTE THI NHI IR GS HEE s FDI ER IAB GEA ETI RTE 27.5 21.9 | et gf Sm 64.0 58.6 Bans sme re a mms G0 on 5h 8 0 8 eS 3 8 TD oe B02 Bo 4100 Te AIRE on 8 Bm 1 9 0 80.9 77.5 Yt sm SEPIA ES AIRES IR UAB RANE VNR AS EIS HEE RAR EERIE 92.1 87.6 BB i ei 3m wm de BE WR TRB ee eB WR 8 eal TR WR EE RE 8 94.9 91.7 BOT TIVOTT 4 0 50 8 5070 9: 5.00 1 0 3 90 1 8.000 00 9c 310 10 91 1 Je Je 00 og ifs Bo pi 0 os 2 aes 0 4 ano he 4 100.0 100.0 23 Table 15. Number of respondents! asked about selected sensitive characteristics and percent reporting characteristic, by type of interview administration [See appendix | for definitions of terms] Interview administration Charscteristic Self- Interviewer- Self- Interviewer- administered administered administered administered All respondents Number of respondents Percent with characteristic Everhad intercourse = ss s was es nmi sw ws soem ais sms emma + 293 310 49.5 49.0 Ever pregnant . .. . oii iii eee 293 © 310 16.7 171 CUrreNtIY PIOGRANT. . oo sc tins sams sp Hi SRN ER HERBS SMWs #9 293 310 243 1.9 Ever pregnant Used contraception: AL TSE IBICOUNISR oy mmx «ui wis io iw wie twow us i ir 90 0 okt me 00 tn 0 om 49 53 23.9 321 Between first intercourse and first pregnancy ............. 49 53 32.6 35.8 Between first and second pregnancy ..............0u... 19 23 52.6 52.2 Number of pregnancies: i I Lr LL ai a ee pe fr Se 49 B3 61.2 56.6 i ow wi win ow wine x om to wr ow mw ae i 0 8 cm 63 Be BoB een FUR Be FB 49 53 24.5 24.5 BOI MOIR. svn sv mai a BF INH EERE R DTS CHEE RFR ERE. 49 53 14.3 18.9 Outcome of first pregnancy: LIVE BIEN 5 ivr or % 0: 5s 9rw 502590 fon-we ws once, sot i on wi so 92 se vaso ie Semen 49 53 75.0 66.7 ABOETION + oo ain in wien mo, 928 8000 30k i620 3.050 30 3 7600 3 6 60 BE 10 900 4 0 49 53 16.7 25.0 Miscarriage or StlbDITth + cs sc ssvs mwas was sess mm es ome + 49 53 8.3 8.4 Sexually active, never pregnant Used contraception: At First intercourse . o.oo vue 92 98 46.7 46.4 Since first INIBTCOUEES +5 vs v 5s 3 ERT VE Ts STF 2 ERAT 6 HET» 47 53 66.7 67.9 BVBr wvms sans s amet a Bas SHAE S0ss eH Be 80 SLES E 9 Bee 92 98 80.5 82.7 TExcludes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. 2Excludes those not having intercourse after menstruation began. 3Excludes users at first intercourse, respondents having only 1 intercourse, and those not having intercourse after menstruation began. 24 Table 16. Number of respondents! asked about selected sensitive characteristics and percent giving no answer, by type of interview administration [See appendix | for definitions of terms] Interview administration Chargeieristic Self- Interviewer- Self- Interviewer- , administered administered administered administered All respondents Number of respondents Percent giving no answer Everhad Intercourse: «x ows » wwe oS ws smmis some Mums sues 8 293 310 2.0 - Ever pregnant ....... i rari auc 8 4 Ed Bw Bn 8 od BH S00 8 4 Bom ag 293 310 1.4 - CUrTentiY PYEORANT ... os vw ws swims sms sms sams slams a mms 293 310 0.7 - Ever pregnant Used contraception: AL EIISLINLOTCOMISE| ov vis 5 wwes Finis 8 [8 io oo io 30% 00 200 0 0 0 0g 49 B53 6.1 - Between first intercourse and first pregnancy ............. 49 B53 6.1 - Between first and second pregnancy . . ........ 0... 19 23 - - NUMber of Pregnancies « « ««s swiss sous snus imme spss s@y +» 49 53 - - Outcome of fist PrEgRBNCY « » « «ss shiv s smms taivs slres sins 49 53 18.2 - Age At TIS INIBICOUISE « » vs vi mns vinina swims sons spins vom is 49 83 2.0 - Sexually active, never pregnant Used contraception: At first intercourse? . . o.oo vite 92 98 2.9 1.0 Since first intercourse3 . . . ie 47 53 10.6 - Age at First Intercourse . .. vss ws snmesmor sofa sms ois 93 99 3.2 2.0 1TExcludes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. 2Excludes those not having intercourse after menstruation began. BExcludes users at first intercourse, respondents having only 1 intercourse, and those not having intercourse after menstruation began. Table 17. Number and percent distribution of responses’ to selected open-ended items by type of response, according to type of interview administration [See appendix | for definitions of terms] Interview administration Item and type of response Self- Interviewer- Self- Interviewer- administered administered administered administered Reason for nonuse of contraception at first intercourse Number of responses Percent distribution PALVCSDONSEE ili oo 0. 10 1 os 3.000 0 5 5 50 ot ot wh or 8 wnat ac 2 Bo ree 85 89 100.0 100.0 SPeCHICIEBBON GIVEN +» vv viv sma vim os avon mins ons in 64 78 75.3 87.6 NG reasonior ORL KNOW, «vw a 9 50 0 x 8 0 9m a ® 00 8 0k wo a4 w 4 10 4.7 11.2 INGIARSWEY wo vivo 0 0 900 bein 3 30 500 8 Brel) 0 Foiod & BU WFR: 009 0% 208 17 1 20.0 1.1 Reason for skipping use of contraception AHTESPONSAS . . sss vos srr mm as RAE * TEPER Pad Ba bw 45 29 100.0 100.0 SPECIFIC IOBBOM GIVEN «iv» civ 5x 5 aw wih 3i% 5 5% ow J 5 we hs ) 36 28 80.0 96.6 ING reason OF GONE KNOW « + vw w 4 v5 ws 30 ww «mat + 45 wk Sw» 5 1 1 2.2 34 INGEARSWRE vn 4 mms: 3700 ole 3 2m 8 BE TR S&S 8 - 17.8 - TExcludes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. 25 Table 18. Percent of respondents who answered yes to the question “Did you find any of the questions hard or uncomfortable to answer?’, by type of interview administration, age, race, and survey area [See appendix | for definitions of terms] Interview administration Age, race, and survey area 1 Self- Interviewer- Vora! administered administered Percent AILTESOONTBINS «vv vn wns ar 0 0 a ari ic 0 00 000 5 0) 90100 0 3 008 fot WLR 1 0 ns ci or 2 90 @ 25.7 225 28.7 Age (LG eT LL I TIT TIT Ir 25.1 23.0 27.3 VBBL YEAS viv vv 5/0 514 #08 © wws v0 0% aw 008 AE ww we aE 26.6 21.8 30.4 IS TOVOAIS. vv 0:5: 0 wie iv 00 madd suc az ones sata wns 0 coed Whip 0 0 9 wt lez ls wi 0 4d 20.6 20.0 21.2 OAR YO ois iin sd Bos mR SAT RRRE ERTS TREE BERS BRIE BRE me 28.8 22.6 33.3 Race BIECK «55.02 4mm 20 EF SWE EHR EE GEE EG RE EEA A Ww a We 28.2 25.6 30.8 OLIVOF TBOBE. « «viv v0 0 09 ae wy car io awn are en 0s 0) 00 0 1 5 24.7 21.3 27.9 Survey area SUDUEGEYNOTINBASE + vc. +00 wv sise: bon a0 did ih Belen £7008 400 lol: 20 0 3 AH ae 21.9 13.8 29.1 Central city NOrthBast ... scvs so ss ans sp @e uae pos LE REE RAR ETE 24.4 225 26.1 Urban SOME 5 ars 55.56 40090900 55 10 0 000008 2000 5 400 3 300 00 A 34.0 324 35.9 BUI SOUL ov wv sirm ow wimune wre wm wom vows io ake wie ow www $ 20h 2 22.7 225 23.0 1includes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. Table 19. Percent distribution of respondents? given the self-administered questionnaire by preference for interviewer-administered questionnaire, according to age, race, and survey area [See appendix | for definitions of terms] Preference for interviewer-administered questionnaire Age, race, and survey area , 2, ” y Not Don’t know Total Preferred preferred or not sure Percent distribution AUTESPONCRINS. + + cnn vv sep ms mmmsmm ss Rais same sas se 100.0 13.7 65.6 20.6 Age B17 VODES ov nvr s on 0s im i009 we, 0» bot wt 0 100.0 15.0 62.4 22.5 ABABA YBALS 1... vv « iv 00s ios wm co wn ow 1h #08 0 as 010 ch fo 80 ol 00 0 5 00 00 ca 100.0 11.9 70.3 17.8 1B IDYORIS oo sans ia RI RARE ARPT HTD ANB ARR RE SES 100.0 5.7 85.7 8.6 D088 YOES & v5.50 5: 5.5 50) 8 000 0 090 EE AE ae 100.0 14.5 63.9 21.7 Race BACK isir5: vr 1 02 he vot raver et et anes be Ranigzca: zen ios os eb sane: Rens et IAL RET er RMR 100.0 23.3 58.1 18.6 OhBPT80ES + vs cvs mrs A FER EF FETA A ARERR BESS B ER Te 100.0 9.8 68.8 25 Survey area Suburban NOrtheast = uw ws sors sr ss rms Foe de wale wv ws Feo 100.0 7.5 81.3 11.3 Control oily. NOrNBESE xv vv vow vd vw sin is ww wid ww win wise wom a 100.0 18.6 54.3 279 KIPBRIVSOUI veo 0 oncom cw orbs wcwone imi 0. 55a) 800 0 ALLER Hite ES 100.0 14.3 62.9 22.9 Burl SOUth «cs ss sss nsw ada ss +8 va NEL CEPR 0D 80 100.0 15.5 62.0 225 1Excludes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. 26 Table 20. Percent distribution of respondents! given the interviewer-administered questionnaire by preference for self-administered questionnaire, according to age, race, and survey area [See appendix | for definitions of terms] Preference for self-administered questionnaire survey area 4 Age, race, and y Not Don’t know Total Preferred preferred or not sure Percent distribution ANTESPONCOIMES «ov vv vv vs sas sens sm mssiesssss sieve 100.0 39.4 51.3 9.4 Age VE 17IY0OYE 0 vv 00 wit 0 16 5 00 0 ARE 8 30 8 RR WE EEE EE 100.0 43.0 50.6 6.4 VS-A4YBNS. + oli vi th ws ub Bs SMBs ade s CRBS SRLS RT He 5 T 00 100.0 34.8 52.2 13.0 IS IOVOBIS ois v0 oun oom a Sod 319 40 S05 50 0a a AE Se 100.0 48.5 42.4 9.1 BOAAYOANS ovo viv ss 5 5iv Bi 0% 2% 08 5.55% 5 55 w Ta ole vee 100.0 30.5 55.2 14.3 Race BACK: cin vv ow © pwd wh nln Hie @e 45 TT Bias Sibded Sul sins 100.0 34.1 87.1 8.8 OIhCETACES v5 + laws nme b 2.0 ME ANTS SHIH 46 AF SEES 2 E #0 100.0 41.6 48.9 9.6 SUbUIBEN NOPINOASE + 4 + 30 5 # 5+ wn wi Si Baw Few am 00 aw ww 100.0 41.8 50.6 7.6 Contral City NOTTABASY. «wow iv wv a0 wn wo wri ari w a 0 aces wwin te 100.0 38.0 51.1 10.9 HBNSOUIN 4 oo wiv hav vo ws Bah oT Hh ou lie vider HE Res 3 100.0 44.9 44.9 10.3 ROrSISOULR «ines «os sms $06E 200 5F 8-5 $RDS OH #058 100.0 31.1 60.7 8.2 TExcludes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. Table 21. Mean length of interview in minutes of respondents, ! by type of interview administration, age, race, and survey area [See appendix | for definitions of terms] Interview administration Age, race, and survey area Self- Interviewer- Yoel administered administered Mean length of interview in minutes AU AGSPONTBIMS oa v 210% & de @ io 55 SE 3 HOWE FETE LR ® 000 EVE 0F o 10.94 4 39.5 39.4 39.6 Age AGV7 YAS cvs vain sv dp pi mE rR Am AR BERS Sea Ra OAT ow EEE piney 36.2 37.8 34.6 ABBA NOOSE ovis vv 55 vw win i 5 05 000 Jo 0 0 308 T0050 0 ew 0 he ww 44.0 41.8 46.0 IE TDYVOAS ov viv ve mas wile sens 510d ow aS 31 © hE Hw EE TERE 37.1 34.2 40.3 ROMA YBBUS ci vv 200% 2 008 5 418 07 3.88 5 D100 5 00 F +580 we wlan PRE Re 46.5 45.0 47.7 Race BIACIL. i rina tae 3% 0 6 5050) as 00 T0000 0 99 LR 5 EEE TER 0 We BR ER ee 47.9 48.0 47.7 ONBr TAGS cv +s ms 50m a sass e660 50 ®ve sus md es sds rsvs ss sss wee 36.3 36.0 36.5 Survey area Suburoan Northeast uve coms sams sma dba v aE ms 50 qv vey ums es kE se 33.9 34.6 33.1 Central CIty NOFINBAST +» & suv ssa sa shams Bink 8 ® a a: 410 4 0 41500 0 5 3 030 wd . 42.4 41.1 43.4 LIPDBNY SOUTH © 5 2 vv wo iw oi 550 91 5 00 955) 08050 8 6 55 50% 9 31 90 § 0 190 808 185 ai ek (win) no 0 00 wie 39.0 39.6 38.5 ROULBLSOUIN' o100.0 5: 000000 5 010 003 0 09 £10 90 Tk tr Bien 0) 0 903 20 S00 3 0 0 0 0 lw 6 0 900 43.5 43.2 43.7 1Excludes 3 respondents assigned to self-administered questionnaire group but receiving interviewer-administered questionnaire. 27 Appendixes Contents 28 Definitions of certain terms used INTIS TEPOrt . « vi «c+ ivi vss vs pans vs s SRB ws SNS RBE sn BEs be Ewa 29 Advance letter, pamphlet, and flip charts constituting the prior information provided to respondents ............. 31 POVENCEBIBEIBE ium vv v0 % 0 00 0% 5.5 0 ore miahs 3, 50 3 0 0 i A 00 70 oo 7 or a 4 00 A 0 3) 6 tet 0 0 0 es 31 PAMBIIEE. vv»: 55 5 wen 0 55 00 0 oon. 0 50 0 ft 1 0 Be 5 2000 0 Lp I 000 1 0 1 0 ie BLO BB tl 0s 2 5 35 I cit 32 FUHDCHAES wieivs vinnie tres as see Rome ter a REE vs BESET ES ERAT 7 Sra Se NL 5a IE Ber ee 5 mie oir a ee 36 Paront GUESLIONNEINE: © vivo vs bmw sss HAFEREES SR ar EEs ESSE ROI TE HEBER RRR EERIE § oh ESTE 41 . Selected questions from the self-administered questionnaire (SAQ) and the interviewer-administered questionnaire UANCI © coe 5mm 3000 oh S105 18 RE 1 GLRSR 4 RFR 0 RR eR 46 Appendix I. Definitions of certain terms used in this report Combined response rate. —Product of the screener and interview response rates divided by 100. Conterminous United States.—Land area consist- ing of the District of Columbia and all States except Alaska and Hawaii. Dwelling unit. —A single room, or group of rooms, intended for separate living quarters in which the people must live and eat separately from everyone else in the building (or apartment), and the room or group of rooms must have either: a. A separate entrance directly from the outside of the building or through a common hall, or b. Complete kitchen facilities for the use of this household only including: ® A range or cooking stove. ® A sink with piped water. ® A mechanical refrigerator. Education. —The highest grade of regular school completed. Family income. —Total combined income during 1978 for all family members living in the household, including income from all sources such as wages, salaries, Social Security or retirement benefits, help from relatives, and so forth. Geographic region. —U.S. Bureau of the Census groups the 50 States and the District of Columbia into four regions as follows: Region States included Northeast . . . . Maine, New Hampshire, Vermont, Massa- chusetts, Rhode Island, Connecticut, New York, New Jersey, Pennsylvania North Central . . Michigan, Ohio, Indiana, lllinois, Wiscon- sin, Minnesota, lowa, Missouri, North Da- kota, South Dakota, Kansas, Nebraska South. . ..... Delaware, Maryland, District of Columbia, Virginia, West Virginia, North Carolina, South Carolina, Georgia, Florida, Ken- tucky, Texas, Tennessee, Alabama, Missis- sippi, Arkansas, Louisiana, Oklahoma States included West ....... Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, Nevada, Washing- ton, Alaska, Oregon, California, Hawaii Alaska and Hawaii are not included in the NSFG sample design. Household. —A family living together, or five or fewer unrelated individuals living together in a dwell- ing unit. Interview nonresponse rate. —Percent of women eligible to be interviewed who did not complete the interview because of refusal or other reasons. Interview refusal rate. —Percent of women eligible to be interviewed who refused to complete the interview. Interview response rate. —Percent of women eligi- ble to be interviewed for whom an interview was completed. Interviewer-administered questionnaire. —Ques- tionnaire form in which all questions are read to the respondent by the interviewer, and in which all responses are recorded by the interviewer. Item nonresponse rate. —Percent of women who provided no answer, refused to answer, or answered “don’t know” to a particular question in the inter- view. Race. —Classification as black or of “other races” according to interviewer observation at the time of the screener interview. Screener interview. —Preliminary interview at the household to collect information about the dwelling unit and to determine whether the household in- cludes one or more women who are eligible for the detailed interview. Screener response rate.—Percent of sample dwell- ing units for which a screener interview was com- pleted. Self-administered questionnaire. —Questionnaire form in which the respondent reads the interview 20 questions and records the answers without inter- viewer intervention. Standard metropolitan statistical area (SMSA). —A county or group of contiguous counties (except in New England) that contains at least one central city of 50,000 people or more, or “twin cities” with a combined population of at least 50,000. In addition, other contiguous counties are included in an SMSA if, 30 according to certain criteria, they are socially and economically integrated with the central city. Urban area.—As defined by the U.S. Bureau of the Census, all cities or “twin cities” with at least 50,000 population in 1970 together with the sur- rounding closely settled area and all other incorpo- rated or unincorporated population centers with 2,500 inhabitants or more. Appendix ll. Advance letter, pamphlet, and flip charts constituting the prior information provided to respondents Advance letter DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE PUBLIC HEALTH SERVICE OFFICE OF HEALTH POLICY, RESEARCH, AND STATISTICS HYATTSVILLE. MARYLAND 20782 NATIONAL CENTER FOR HEALTH STATISTICS Dear Friend: The United States Public Health Service is doing an important study about American families and childbearing. This study will show changes in our population and in the needs for medical care and other services, both public and private. It will also give scientific information on maternal care, teenage pregnancy, day care services, family planning, sterility, and other matters about childbearing--information which is needed for many public health service and medical programs. We have asked the Institute for Survey Research of Temple University--a nongovernment survey organization--to visit and talk with women in a sample of households around the Nation. Every household in the country had a chance to be chosen. Since we cannot visit every household in the Nation, the sample house- holds were scientifically selected from among all groups of our people. Your household is one of those chosen. Although your help in this study is completely voluntary, it is also very important. Each chosen house- hold, like yours, must represent many others that we cannot visit, and once a household is chosen, we are not permitted to substitute another. So, only you may answer for all those you represent. In the next few days, an interviewer from the Institute for Survey Research will call at your home. Please show this letter to the other members of your household, so that they will be expecting the inter- viewer, too. She will have an identification card and will carry a letter of introduction from the United States Public Health Service. When you talk with the interviewer, the information you give will be kept completely confidential as required under laws passed by the Congress of the United States. Your answers will be put together with the answers from other households to make totals, averages, and other statistics. The results will help us to understand better the growth and needs of American families. Your cooperation is a public service that will be very much appreciated. This study is called the National Survey of Family Growth and is author- ized by the Public Health Service Act (42 USC 242k). If you have other questions, we will be pleased to answer them. Sincerely yours, Serie #2 Director 31 Pamphlet Institute for Survey Research Temple University National Survey of Family Growth @elglo{SIei Clo R{o)S U.S. Public Health Service National Center for Health Statistics NATIONAL CENTER FOR HEALTH STATISTICS National Survey of Family Growth All your life you've been reading and hear- ing about national surveys, yet it is unlikely that you ever participated in any. Now your household has been chosen to take part in an important study called the NATIONAL SURVEY OF FAMILY GROWTH. In this pamphlet, we try to answer some of the questions people frequently ask us about the survey. WHAT IS THE NATIONAL SURVEY OF FAMILY GROWTH? It is a nationwide survey conducted by the National Center for Health Statistics, a part of the U.S. Public Health Service. Every few years, brief interviews are conducted in a sample of households across the nation, chosen to represent all groups in our population. More detailed interviews are conducted with about 10,000 women in the childbearing years who live in these sample households. The survey is authorized in Section 306 (b)(1)(h) of the Public Health Service Act (42 USC 242k). From the National Survey of Family Growth, we learn many medical and social facts about pregnancy and childbirth among American women. In the interviews, we talk with women about their knowledge of pregnancy and childbearing, about marriage or plans to marry, about their physical and sexual development, about the babies they have had orexpecttohave, about theirplanningofbirths or getting help to have babies, and about health problems and health care before, during and after pregnancy. There are other questions in the survey which ask about some related family facts such as schooling, work experience, day care, and present employment. HOW WAS | CHOSEN? In doing this survey we cannot talk to every woman--that would be far too expensive. So we scientifically select a “cross section” of households. We begin by choosing certain counties or cities. Then, in each of the selected areas, we choose small areas such as blocks or tracts of land. Finally, we choose certain households within the smaller areas. We do not know who lives in the chosen households before we get to the door. But the people who live in this select group of households make a sample of the people in the counties and cities chosen. Since the survey is about pregnancy and childbearing, only women in the childbearing years (15-44 years of age) will be interviewed, and only one eligible woman will be interviewed in a household. If there is more than one eligible woman in the household, one of them is randomly chosen to be sure that the sample is representative of all women in the childbearing years. Thus, eachwomanwho is chosen to be interviewed represents many others of the same age, education, medical history and so forth. If you are chosen in your household and cannot participate in the survey, for any reason, then all the other women you represent will also be missing from the totals. The results may be misleading. 33 34 HOW DO | KNOW MY ANSWERS WILL BE KEPT CONFIDENTIAL? Confidentiality of all the information you give is protected by public law, Section 308(d) of the Public Health Service Act (42 USC 242m) and the Privacy Act of 1974 (5 USC 552a). Any information which will allow the questionnaire to be identified with an individual is kept separately from the actual questionnaire. Your answers will be used by research project staff working on this survey. Each of them has signed an affidavit to keep confidential all information provided by respondents. Finally, all personal identifying information such as names, addresses, local community and other selected information which might readily identify an individual is removed before data from this survey are made available to others for bona-fide research purposes. The answers you give willbe combined with those from thousands of other households and the results will be reported in percentages and totals in such a way that no one’s answers can be identified. WHY IS THE PUBLIC HEALTH SERVICE DOING THIS SURVEY? The U.S. Public Health Service uses the survey results to better carry out its responsibilities for the health of the nation. From the survey we can better understand how muchthe populationis likely to growinthe next few years. This information is needed for planning public facilities—such as schools, housing, hospitals and facilities for older citizens. The survey information is a vital part of health research to provide better health services and health education--programs which help people in need such as couples unable to have babies of their own, pregnant “teenagers trying to solve their problems, couples looking for a safe and acceptableway to space their children, women concerned about cancer of the reproductive organs, and working mothers who need reliable day care services for their children. Many other public and private organizations also need the statistics from this survey. Since surveys like this one are expensive, and each organization cannot afford one of its own, the government makes the results available in statistical summaries and reports, and in other data forms for research purposes. DO | HAVE TO ANSWER THE QUESTIONS? No! Your participation is completely voluntary and confidential, and your choice will have no effect on any services, privileges, or benefits to which you are entitled. However, each chosen household represents many others that were not chosen, and it is very important that we get your answers so that others like you will be represented. Once your householdis chosen, we are not permitted to substitute another household for yours, so only you may answer for all those other households you represent. WHAT GOOD ARE SURVEYS, ANYWAY? HOW WILL | RECOGNIZE THE FAMILY A survey is conducted when information is GROWTH SURVEY INTERVIEWER? needed about a larger group of people, but The interviewer who calls on you is the time and money make it impossible to talk to Institute for Survey Research representative everyone. A sample of the total group is in your area. She will be carrying identification carefully selected and used to estimate the which looks like the card shown below. answers that would have been given by all. Surveys are not a new idea. In earlier days, survey methods tended to be poor and unscientific. But in recent years, researchers oS, VARS INVIRSITY IBENTICATION CARS have developed far better methods of ISR ann! conducting surveys, so that it is now possible S This is tw identify: to make very good estimates about the 7 population from a carefully drawn sample. A ID# AS AN INT! WEI FOR THE INSTITUTE FOR Suv RESEARCH £& WHO IS THE INSTITUTE FOR SURVEY ELLIN SPECTOR RESEARCH? ASSISTANT DIRECTOR The Institute for Survey Research is an independent research organization which is a part of Temple University in Philadelphia, Pennsylvania. It conducts surveys on many different subjects. It has been chosen by the National Center for Health Statistics to conduct this phase of the National Survey of Family Growth. Institute for Survey Research Temple University 1601 N. Broad St. Philadelphia, PA 19122 215-787-8351 35 36 Flip charts YES = 86% HAVE YOU HAD A BABY BORN TO YOU AT ANY TIME? BREASTFED 36 percent OID NOT BREASTFEED 64 percent ABOUT ONE OF EVERY THREE WOMEN INTERVIEWED SAID THAT THEY BREASTFED THEIR CHILDREN AT INFANCY. YES + ONE OF EVERY FIVE WOMEN INTERVIEWED HAD BEEN HOSPITALIZED BECAUSE OF PREGNANCY COMPLICATIONS, THAT IS, FOR REASONS OTHER THAN NORMAL DELIVERY OR FALSE LABOR. 20 percent —& pave HAD TWO : ™y OR MORE Cn L HAVE HAD ONE 75% HAD NONE WHILE THREE QUARTERS OF THE WOMEN REPORTED NO PREGNANCY LOSS, ONE OUT OF FOUR HAS HAD AN ABORTION, A MISCARRIAGE OR A STILLBIRTH. 37 38 WANTED 65 of every I00 births MISTIMED 24 of every 100 births UNWANTED MORE THAN ONE THIRD OF THE BABIES BORN TO AMERICANS EITHER ARRIVED AT THE WRONG TIME OR WERE NOT WANTED AT ALL. 39 percent ABOUT 25 PERCENT OF MARRIED WOMEN 39 PERCENT OF WOMEN WITH MEDICAL HAVE A MEDICAL PROBLEM THAT MAKES PROBLEMS IN HAVING BABIES WOULD IT DIFFICULT OR IMPOSSIBLE TO HAVE LIKE TO HAVE A CHILD IN THE FUTURE. ANY BABIES IN THE FUTURE. percent 22 20 19 % 13% PILL IUD CONDOM DIAPHRAGM FOAM, RHYTHM CREAM, JELLY CONTRACEPTIVE METHODS DIFFER IN TERMS OF THEIR EFFECTIVENESS. FORINSTANCE, THE COUPLES USING THE CONDOM ARE FIVE TIMES MORE LIKELY TO FAIL THAN THOSE USING THE PILL; AND THE DIAPHRAGM THREE TIMES MORE THAN THE 1UD. IN OWN HOME BY RELATIVE IN RELATIVE'S HOME IN NON-RELATIVE'S HOME DAY CARE OR OTHER SPECIAL ORGANIZED FACILITY ise | IN OWN HOME BY NON-RELATIVES OR 2S) OTHER KINDS OF ARRANGEMENTS MORE THAN ONE-HALF OF WORKING MOTHERS HAVE RELATIVES TAKE CARE OF CHILDREN, AND ABOUT 12 PERCENT USE ORGANIZED FACILITIES LIKE DAY CARE 39 40 or 60 59% 50 - 40 30 percent sexually experienced AGES 15-17 AGES 18- 19 PROPORTION OF SEXUALLY EXPERIENCED SINGLE WOMEN HAS INCREASED BETWEEN 1971 AND (976 60 Ts 50 Q < > © N QU 40} on, 8 s 30F g o N Q I 20F L) e $ 11% 0 3 pom] bd NEVER SOME- ALWAYS ALWAYS USED TIMES USED USED A USED MEDICAL METHOD PREMARITAL PREGNANCY IS CONSIDERABLY MORE FREQUENT AMONG ADOLESCENTS WHO NEVER USE A CONTRACEPTIVE METHOD THAN AMONG THOSE WHO ALWAYS DO. Appendix lll. Parent questionnaire STUDY #518-225-01 SUMMER 1979 LA#: Information released to INTERVIEWER'S NAME: DATE: selected to be interviewed in this household. | see that she is INSTITUTE FOR SURVEY RESEARCH TEMPLE UNIVERSITY -0f The Commorwealth System Of Higher Education- 1601 NORTH BROAD STREET PHILADELPHIA, PENNSYLVANIA 19122 OMB No.: 68-578056 Expires: December 1980 PARENT QUESTIONNAIRE ONLY FOR MINORS IN GROUPS 3, 4, 7, 8 HU# : —- TREATMENT #: I Assurance of Confidentiality contained on this form which would permit iden- tification of any individual or establishment has been collected with a guarantee that it will be held in strict confidence by the contractor and NCHS, will be used only for purposes stated in this study, and will not be disclosed or anyone other than authorized staff of NCHS with- out the consent of the individual or the establishment in accordance with Section 308(d) of the Public Health Service Act (42 U.S. C. 242m). Introduction My sampling rules show that is the person (NAME OF RESPONDENT) (AGE) years old. Before we can include her in the survey, we are required to have parental consent and we need to talk with her mother about some of the questions that she can answer best. 1D#: CASE #: 41 42 (ONLY ADMINISTER IF IN TREATMENT #'S 3, 4, 7, OR 8) Altogether, how many babies have you given birth to, including any who died very young? (SKIP TO Q. 5) None 000 (NUMBER OF LIVE BIRTHS) Now I'd like to get some information about (your baby/each of your babies). (ASK QQ. 2-4 FOR EACH LIVE BIRTH) When was your (Ist, 2nd, etc.) child born? (RECORD IN COLUMN 1) What did you name the baby? (RECORD IN COLUMN 2) Was a boy or a girl? (CIRCLE CODE IN COLUMN 3) (NAME OF CHILD) COLUMN 1 COLUMN 2 COLUMN 3 SEX BIRTH DATE NAME Boy | Girl First child 1 2 Second child 1 2 Third child 1 2 Fourth child 1 2 Fifth child 1 2 When were you born? Loy / OR (MONTH) (DAY) (YEAR) AGE (IF DOES NOT KNOW DATE OF BIRTH, ASK): How old were you on your last birthday? 6. What is the highest grade or year of regular school or college you have completed? No formal schooling 00 Elementary School Ol 02 03 O04 05 06 07 O08 High School 09 10 1 12 $l lage and Graduate/Professional BW CAE YEN ee Other (SPECIFY): 96 7. Are you Protestant, Roman Catholic, Jewish, or something else? (Go TO Q. 8) Protestant 01 Roman Catholic 20 (skip Jewish 30 Other (SPECIFY): TO Lo Q. 9) None 50 Don't know 98 8. What denomination is that? Baptist 21] Lutheran 22 Methodist 23 Presbyterian 24 Episcopalian 25 No specific denomination 28 Other Protestant (SPECIFY): " Don't know 98 43 9. What is the highest grade or year of regular school or college (RESPONDENT'S FATHER) has completed? No formal schooling 00 Elementary School Ol 02 03 Oo4 05 06 07 O08 High School 09 10 11 12 College and Graduate/Professional 13 14 15 16 17 18+ School Other (SPECIFY): 96 10. Is (RESPONDENT'S FATHER) Protestant, Roman Catholic, Jewish, or something Sop RARE else (Go 10 Q. 11) Protestant ol Roman Catholic 20 (skip Jewish 30 Other (SPECIFY): TO Lo Q. 12) None 50 Don't know 98 11. What denomination is that? Baptist 21 Lutheran 22 Methodist 23 Presbyterian 24 Episcopalian 25 No specific denomination 28 Other Protestant (SPECIFY): 29 Don't know 98 12. When was (RESPONDENT'S FATHER) born? 7 Lo. OR TMONTH) (DAY) (YEAR) AGE (1F DOES NOT KNOW DATE OF BIRTH, ASK): How old was he on his last birthday? 13. Last year--that is, in 1978--what was your total combined family income, that is yours and any other family member living here now? Include income from all sources such as wages, salaries, Social Security or retirement benefits, help from relatives, rent from property, and so forth. $ —p (SKIP TO Q. 15) pm (TOTAL FAMILY INCOME) (Go Refused 97 TO Q. 14) Don't know 98 14. (HAND R CARD 10) Here is a card showing amounts of weekly and yearly income. Next to each amount is a letter. Would you tell me what letter represents the income of your family during the past 12 months? (RECORD LETTER) (LETTER) 15. Thank you for talking with me. Now | need to talk to : (RESPONDENT) (ASK FOR PARENT'S CONSENT TO INTERVIEW MINOR) 45 Appendix IV. Selected questions from the self- administered questionnaire (SAQ) and the interviewer- administered questionnaire (IAQ) Topic Whether currently pregnant Whether ever pregnant Number of pregnancies Outcome of first pregnancy Whether ever had intercourse Age at first intercourse Whether contraception used at first intercourse 46 SAQ (ASKED OF RESPONDENTS WHOSE LAST MENSTRUAL PERIOD WAS NOT WITHIN THE LAST 31 DAYS): What do you think is the reason why your period is delayed? (Q.1) Have you ever been pregnant? (Q.2) How many times have you been pregnant including your current preg- nancy if you are pregnant or think you may be pregnant now? How did this first pregnancy end? (Q.C22) (ASKED OF NEVER-PREGNANT RESPONDENTS): Have you ever missed a period when you thought you might be pregnant? (Q.3) (IF NO): Have you had sexual intercourse at any time in your life? (Q.4) (ASKED OF EVER-PREGNANT RESPONDENTS): How old were you when you had sexual intercourse for the first time in your life? (Q.9) (ASKED OF NEVER-PREGNANT RESPONDENTS): How old were you when you had sexual intercourse for the first time ever? (Q.36) (ASKED OF EVER-PREGNANT RESPONDENTS): The first time you had sexual intercourse after your IAQ (ASKED OF RESPONDENTS WHOSE LAST MENSTRUAL PERIOD WAS NOT WITHIN THE LAST 31 DAYS): Are you pregnant now? (Q.23) (IF NO): What do you think is the reason that your period is delayed? (Q.29) Have you ever been pregnant (before)? (Q.30) How many times have you been pregnant altogether (including this one)? (Q.31) Did your first pregnancy end in a live birth, an abortion, a miscarriage, or a stillbirth? (Q.44) (ASKED OF NEVER-PREGNANT RESPONDENTS): Have you ever missed a period and thought you might be pregnant? (Q.69) (IF NO): Have you had sexual intercourse at any time in your life? (Q.70) (ASKED OF EVER-PREGNANT RESPONDENTS): How old were you when you had sexual intercourse for the first time in your life? (Q.32) (ASKED OF NEVER-PREGNANT RESPONDENTS): How old were you when you had sexual intercourse for the first time ever? (Q.71) (ASKED OF EVER-PREGNANT RESPONDENTS): The first time you had sexual intercourse (after Topic Whether contraception used since first intercourse Whether contraception used between first intercourse and first pregnancy Whether contraception used between first and second pregnancies NOTE: Copies of the pilot study questionnaires are available upon request. SAQ monthly periods began, did you or your partner use any method of birth control to prevent pregnancy? (Q.12) (ASKED OF NEVER-PREGNANT RESPONDENTS): The first time you ‘had intercourse after your monthly periods began, did you or your partner use any method of birth control so you would not get pregnant? (Q.39) (ASKED OF NEVER-PREGNANT RESPONDENTS WHO DID NOT USE CONTRACEPTION AT FIRST INTERCOURSE): Have you or your partner ever used a method of birth control since the first time you had intercourse? (Q.41) Between the first time you had sexual intercourse and the time you first be- came pregnant, did you or your partner use any methods of birth control? (Q.13) Between your first pregnancy and your second pregnancy, did you or your partner use any method of birth con- trol? (Q.13) %U.S. GOVERNMENT PRINTING OFFICE: 1982-361-161:506 IAQ your monthly periods began), did you or your partner use any method of birth control to prevent preg- nancy? (Q.34) (ASKED OF NEVER-PREGNANT RESPONDENTS): The first time you had intercourse (after your monthly periods began), did you or your partner use any method of birth control so you would not get pregnant? (Q.73) (ASKED OF NEVER-PREGNANT RESPONDENTS WHO DID NOT USE CONTRACEPTION AT FIRST INTERCOURSE): Have you or your partner ever used a method of birth control? (Q.77) Between the first time you had sex- ual intercourse and the time you first became pregnant, did you or your partner use any method of birth con- trol? (Q.37) Between your first pregnancy and second pregnancy, did you or your partner use any method of birth con- trol? (Q.37) 47 grein rr ae 1 Th i =p . a ul = ) ua Eh wi oy a : N Fahl = Fa Re = re zn i i = Eo TT iT Eg k 1 BE 5 er P= TE we a LF shy . a A= = A . EEE . = - ~ . = LE m - = Te = 5 Fat 2 . ) R a . y B B 5, " * a a - - =" mia ad . - B , - u yg a = “Em . Bi Eo - en 3 y B= Aik ol ic Rm ory oo a A Ea . or i, Ll = En n oe . Is * PEE 2 % i 3 a : i . wat J re - : Eo —- OF = } = LA ol rt - . k . Wn wel Aim = . oo Ng am » = B . . - oo -* f 2 - . B ) j - i Ey - - . : fe a = : at oo re oo Fe a ) "A i k Ek Fa. CF : re og po 5 Se i a - als = EE E oe : £ 3 = Boedg Rh a E oo emer er — BE = - we = . B 3 = I: fi wo ¥ Su 5 oo A “ B ) oo 5 F . - ~ k 1, ¥ - k mek annd, WE CPR dre GL it Sa Ra. ah a, We, a od + Vital and Health Statistics series descriptions SERIES 1. SERIES 2. SERIES 3. SERIES 4. SERIES 10. SERIES 11. SERIES 12. SERIES 13. Programs and Collection Procedures.—Reports describing the general programs of the National Center for Health Statistics and its offices and divisions and the data col- lection methods used. They also include definitions and other material necessary for understanding the data. Data Evaluation and Methods Research.—Studies of new statistical methodology including experimental tests of new survey methods, studies of vital statistics collection methods, new analytical techniques, objective evaluations of reliability of collected data, and contributions to sta- tistical theory. Analytical and Epidemiological Studies.—Reports pre- senting analytical or interpretive studies based on vital and health statistics, carrying the analysis further than the expository types of reports in the other series. Documents and Committee Reports.—Final reports of major committees concerned with vital and health sta- tistics and documents such as recommended model vital registration laws and revised birth and death certificates. Data from the National Health Interview Survey.—Statis- tics on illness, accidental injuries, disability, use of hos- pital, medical, dental, and other services, and other health-related topics, all based on data collected in the continuing national household interview survey. Data From the National Health Examination Survey and the National Health and Nutrition Examination Survey.— Data from direct examination, testing, and measurement of national samples of the civilian noninstitutionalized population provide the basis for (1) estimates of the medically defined prevalence of specific diseases in the United States and the distributions of the population with respect to physical, physiological, and psychological characteristics and (2) analysis of relationships among the various measurements without reference to an explicit finite universe of persons. Data From the Institutionalized Population Surveys.—Dis- continued in 1975. Reports from these surveys are in- cluded in Series 13. Data on Health Resources Utilization.—Statistics on the utilization of health manpower and facilities providing SERIES 14. SERIES 15. SERIES 20. SERIES 21. SERIES 22. SERIES 23. long-term care, ambulatory care, hospital care, and family planning services. Data on Health Resources: Manpower and Facilities.— Statistics on the numbers, geographic distribution, and characteristics of health resources including physicians, dentists, nurses, other health occupations, hospitals, nursing homes, and outpatient facilities. Data From Special Surveys.—Statistics on health and health-related topics collected in special surveys that are not a part of the continuing data systems of the National Center for Health Statistics. Data on Mortality.— Various statistics on mortality other than as included in regular annual or monthly reports. Special analyses by cause of death, age, and other demo- graphic variables; geographic and time series analyses; and statistics on characteristics of deaths not available from the vital records based on sample surveys of those records. Data on Natality, Marriage, and Divorce.—Various sta- tistics on natality, marriage, and divorce other than as included in regular annual or monthly reports. Special analyses by demographic variables; geographic and time series analyses; studies of fertility; and statistics on characteristics of births not available from the vital records based on sample surveys of those records. Data From the National Mortality and Natality Surveys.— Discontinued in 1975. Reports from these sample surveys based on vital records are included in Series 20 and 21, respectively. Data From the National Survey of Family Growth.— Statistics on fertility, family formation and dissolution, family planning, and related maternal and infant health topics derived from a periodic survey of a nationwide probability sample of ever-married women 15-44 years of age. For a list of titles of reports published in these series, write to: Scientific and Technical Information Branch National Center for Health Statistics Public Health Service Hyattsville, Md. 20782 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES POSTAGE AND FEES PAID Public Health Service U.S. DEPARTMENT OF HHS Office of Health Research, Statistics, and Technology HHS 396 National Center for Health Statistics RR EE 3700 East-West Highway : THIRD CLASS —— Hyattsville, Maryland 20782 . U.S.MAIL ES OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 HRST From the Office of Health Research, Statistics, and Technology DHHS Publication No. (PHS) 82-1365, Series 2, No. 91 For listings of publications in the VITAL AND HEALTH STATISTICS series, call 301-436-NCHS U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES e Public Health Service ® National Center for Health Statistics ® Series 2, No. 92 WET LA] Methodology for Analyzing Data from a Complex Survey: The first National Health and Nutrition Examination Survey Data from the National Health Survey, Series 2, No. 92 COPYRIGHT INFORMATION All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated. SUGGESTED CITATION National Center for Health Statistics, J. Landis, J. Lepkowski, S. Eklund, and S. Stehouwer: A statistical methodology for analyzing data from a complex survey, the first National Health and Nutrition Examination Survey. Vital and Health Statistics. Series 2-No. 92. DHHS Pub. No. 82-1366. Public Health Service. Washington. U.S. Government Printing Office, Sept. 1982. Library of Congress Cataloging in Publication Data A Statistical methodology for analyzing data from a complex survey. (Vital and health statistics. Series 2, Data evaluation on methods research; no. 92) (DHHS publication; (PHS) 82-1366) Authors: J. Richard Landis ... [et al.]. 1. Health surveys— Statistical methods. 2. Nutrition surveys— Statistical methods. |. Landis, J. Richard. Il. National Center for Health Statistics (U.S.). lll. Series. IV. Series: DHHS publication; (PHS) 82-1366. RA409.U45 no. 92 [RA408.5] 312'.0723s 82-600240 ISBN 0-8406-0264-2 [614.4'2'0723] For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 VIAL HEALTH olAl of A Statistical Methodology for Analyzing Data from a Complex Survey: The first National Health and Nutrition Examination Survey This report presents an approach to analyzing data from a survey with a complex sample design. The data used to illustrate the approach are from the first National Health and Nutrition Examination Survey, a national probability sample survey that was conducted in 1971-74 with an augmentation survey in 1974-75. Data are examined using regression techniques, analysis of variance, and categorical data analysis. Data from the National Health Survey Series 2, No. 92 DHHS Publication No. (PHS) 82-1366 U.S. Department of Health and Human Services Public Health Service Office of Health Research, Statistics, and Technology National Center for Health Statistics Hyattsville, Md. September 1982 National Center for Health Statistics ROBERT A. ISRAEL, Acting Director JACOB J. FELDMAN, Ph.D., Associate Director for Analysis and Epidemiology GAIL F. FISHER, Ph.D., Associate Director for the Cooperative Health Statistics System GARRIE J. LOSEE, Associate Director for Data Processing and Services ALVAN O. ZARATE, Ph.D., Assistant Director for International Statistics E. EARL BRYANT, Associate Director for Interview and Examination Statistics ROBERT L. QUAVE, Acting Associate Director for Management MONROE G. SIRKEN, Ph.D., Associate Director for Research and Methodology PETER L. HURLEY, Associate Director for Vital and Health Care Statistics ALICE HAYWOOD, Information Officer Interview and Examination Statistics Program E. EARL BRYANT, Associate Director MARY GRACE KOVAR, Special Assistant for Data Policy and Analysis Division of Health Examination Statistics ROBERT S. MURPHY, Director Foreword and acknowledgments This report is a contribution to the literature on sta- tistical methodology designed to improve the analysis of data from surveys with complex sample designs. Specifically, it is designed to help the users of the data tapes from the National Health and Nutrition Exam- ination Surveys. Many of these users are analyzing data from a survey with a complex sample design for the first time. Some statistical guidance is required to utilize the data available from such studies properly. This report should provide such guidance so that they can proceed with confidence and caution. To make this report possible, the Division of Health Examination Statistics extended the contract (no. 233-79-2092) with the School of Public Health at the University of Michigan. Dwight Brock of the National Institute on Aging, NIH, Ron Forthofer of the School of Public Health, University of Texas, and Robert Casady of the Na- tional Center for Health Statistics, reviewed drafts, dis- cussed statistical methodology, and made suggestions for changes. Their professional involvement helped us all to learn from this project. Mary Grace Kovar Special Assistant for Data Policy and Analysis Office of Interview and Examination Statistics i 0 En ae Sl a UE {1 Es i i " " re [sy Tk ig Se TP tama: Contents ’ Forewtd andiacknOWIBOOMBILE . ....ci. x. mn ivih vuuivn wie wine isis vine sein ns nmiios sels Sule sae sls 450% #3 na sina iii AVEPOAUCHION: «vv x nv oh visio wis wis im ras 0 40 0009 L905 00 0 EWEN Be 0 #3 0 00 0 3 6.00 es 4 9 oh a ehalis aloe sha nt 1 CS UINBY BBO: i + oie haw ioral age gona oat bm ie Sia ain oni maith gd ae To a pari Fy no ad Aba a AR a eT ry ST RA) 2 CS AMPICESBIOTTION 1. vr smite vhosts. Rarsh +50m paniet dts ino ie] tel A 0 OE oT Lf Sha gE Ra te held aio rh wip 2 IN OBS DOINGS trv ven. vs ihe wor wiienwran io gbm nsw he wt elon 1 0 0 RR TART 400 rT Eire Ee a phd rs 4 MISSING data Bd AMPUIBTION . «sc + vii b irs siakinie £00 B00 BERT WR SI 5000 Wid dw i dia ola 0m aie et web bacon le ve Sr 4h 5 Design considerations for examined PBISONS «1... cute ins sive nis nine nests ysnesnn sss asin salu ves wy elses 5 Analytical strategies............. Ein re te le on HER AT SA St se a a a BEER noe ole 8 ContinUOUS VaRablaS: IMBBINS. . . +5 suis «oie sows aie tsa mam vides 508 55 wale 4s w%0 41s ww sinim wlio nln alain os afin la oat Biavising 9 SUDOIOUR COMPALISONSIMBANS : so i6 sx statin fis wis 5s Foain a askin bin inion aw win) wis) Ros wig Bie im 7 wd) 68 dias be Fw die icp Tora gine sis 9 Continuous variables: Multiple regression Models ..........uiiii in intn nit titties 10 Continuous variables: Analysis OF VBIHANGCE .. qv .cos run trusses subs dammans ess sss ssssassissisees sions sre 12 CHIOQOTICAVANIANIES css vi wilson siamo sind ms in smn (00 0% 00 98 008 S04 1 00 00 (010,810 wie 6 0 1 2 im wpa ml a pai. whe Sonia 16 CUTRINIRIY +o. he slr sural mimo wie fe 4 Bm os AH or 30 of wi He pk ins mn Tn es et 4 GS ARS Hom 02 hia hd te 4 18 hd Ho oH 3 20 RO BRON CES ssw 10 ww bate ah ln a) a won 4 We eo 0 0 Fi Fo 4 E00 3 Ue 3 EET 40 06 ot 08 8 8 SR £00 i giao Selene 22 LiSt OF AIAG BABIES. +. «on cs =: + wie sie win in vids aim Hw 5 550: 0 55 900 4% 5 5 55, 910i 356» (aEin i 8 i ale mk aie 9010) + Tce mee al Win mp 4 408 24 Appendixes I. :Definitions OF tors ANA VAIIBDIGE . cu . vv vo visas ain sins io 6+ 9 wane bo 00m oi 25ers is 21ae0e Wie ie ¥ia-u,0 ale #200 sera i0 ars os 39 I. COMPULING CONION COIS TllBS . + «ova v sis v5 50 Hin sina sits abn e ws mins ® s Kale 56 w Rin 4 0 nino in malate Wa © Viale os was ieis #9 0 w tws 40 fg iT pen HN 0 ik WL 1 3 as = 5 of PE Fl _ Hall =e lat oy 8 i Hel - ype ER k wi : i! > i - . SE i LE Rd i i li i a a tes o . Ly PAs = Hee o pei ) gl oy TR y i #4 (I i ; y : Ie R= - Eri A Statistical Methodology for Analyzing Data from a Complex Survey: The first National Health and Nutrition Examination Survey J. Richard Landis, Ph.D.; James M. Lepkowski, Ph.D.; Stephen A. Eklund, D.D.S., M.H.S.A., Dr.P.H.; and Sharon A. Stehouwer, University of Michigan Introduction For many large-scale surveys like those conducted by the Bureau of the Census and the National Center for Health Statistics in the United States, and the World Fertility Survey coordinated by the International Statistical Institute in the Netherlands, data are ob- tained through multi-stage sampling designs involving clustering and stratification, as well as estimation tech- niques that include post-stratification and non-response adjustments.1-3 Consequently, the direct application of standard statistical analytic methods may be mis- leading for such survey data. The inappropriateness of standard methods in this context is due to the complexi- ~ ties in the sample design which induce a non-standard covariance structure among the sample quantities under investigation. Although the modification of statistical analytic procedures to incorporate the effects of complex sam- ple designs is an important area of research, the methodologies appropriate for such data have not been made readily available to general users of statistical software packages. Exceptions to this are the software packages developed for the analysis of survey data, in- cluding OSIRIS IV (University of Michigan) and SUPERCARP (Iowa State University),> which are available for purchase and documented for outside users. In addition, the programs SESUDAAN and SURREGR, which are accessed through SAS,6 can be obtained from Babu Shah of the Research Triangle Institute. The methods and results presented in this mono- graph were developed in the process of producing the extensive analyses reported in three NCHS publica- tions7-9 based on data from the first National Health and Nutrition Examination Survey. Although the gen- eral methodology outlined herein is not new or unique, some of these procedures are not well known to users of standard statistical software packages. In particular, several computing stages involving separate algorithms are required to generate the analysis of variance and contingency table analyses. This document is intended to provide a representa- tive set of analyses illustrated by data from the first National Health and Nutrition Examination Survey. These data are available on public use tapes and can be purchased from the National Technical Information Service. They permit analyses by researchers with varied statistical approaches and available computing software. Even though other users may not have access to the same computing packages used for this report, parallels with other software will be similar. These results have been computed under various assumptions ignoring the weights, ignoring the sample design, or ig- noring neither the weights nor the sample design. The importance of the design on estimates of variance, and consequently, of test statistics, is highlighted through- out, both as a research finding of interest for this survey and as an illustration of the critical importance of incor- porating these design effects into any analyses of data from the first National Health and Nutrition Examina- tion Survey or from other complex surveys. Survey design The sample design for the first National Health and Nutrition Examination Survey (NHANES I) is basi- cally a three-stage, stratified probability sample of clusters of persons in area-based segments. The sample was designed to represent the civilian noninstitutional- ized population ages 1-74 years in the coterminous United States, excluding persons residing on lands set aside for the use of American Indians. Successive sam- pling units used in the sampling were the primary sam- pling unit (a county or group of counties denoted as a PSU), census enumeration district (ED), segment (a cluster of households), household, eligible person, and sample person. For the April 1971 through June 1974 period, the design provided for selection of a representative sample of the target population 1-74 years of age. The entire sample was given the nutrition-related interview and examination; a subsample of adults 25-74 years of age received a more detailed examination focused on other aspects of health and health care needs. To increase the size of the subsample of adults and, consequently, the usefulness of the data obtained, the design further provided for selection of an additional national sample of adults 25-74 years of age. This sample was given a detailed examination in July 1974-September 1975. The extension of NHANES 1 is referred to as the “Augmentation Survey.”’10 The estimated civilian noninstitutionalized U.S. population ages 1-74 years at the time of examination is shown in table 1 by sex, race, and age. Because cer- tain analyses must be done on the basis of age at exam- ination, for the sake of consistency the population estimates also have been based upon age at examina- tion rather than the age at interview. Sample selection The first stage of the sample selection began with the 1960 decennial census lists of addresses and nearly 1,900 primary sampling units (PSU’s) into which the entire United States had been divided. The 1960 decen- nial census information was used in the selection of first stage units because the 1970 census information was not available. The 1970 census information was used at subsequent stages of selection as it became available, although it was not used for later stages of selection within all primary selections. Each PSU is either a standard metropolitan statistical area (SMSA), a single county, or a group of two or three contiguous counties. The PSU’s were grouped into 357 strata, as for the National Health Interview Survey during 1963-72, and subsequently collapsed into 40 superstrata for use in NHANES I. For the April 1971-June 1974 period, 15 of the 40 “superstrata’ which contained a single large metro- politan area of more than 2 million population were chosen in the sample with certainty. The remaining 25 noncertainty strata were classified into four broad geo- graphic regions of approximately equal population (when the large metropolitan areas selected with cer- tainty were included) and cross-classified into four broad population density groups in each region. A controlled-selection technique!!! was used to select two PSU’s from each of the 25 noncertainty superstrata. The probability of selection of a PSU was proportional to its 1960 population. Representation of specified state groups and rate of population change classes in the selections was controlled in the sample selection process. In this manner a total first stage sample of 65 PSU’s was selected, 15 large metropolitan certainty areas and (2)(25) = 50 paired selections from noncertainty areas. These 65 sample PSU’s are the areas within which clusters of sample persons were selected for examination. Although the 1970 census data were used as the frame for selecting the sample within the PSU’s when the data became available, the calendar of operations required that the 1960 census data be used for the first 44 locations in the sample. The 1970 census data were then used for the final 21 locations of the sample and for the Augmentation Survey. Beginning with the use of the 1970 census data, the segment size was changed from an expected 6 housing units selected from compact clusters of 18 housing units to an expected 8 housing units. This change was imple- mented because of operational advantages. Research by the U.S. Bureau of the Census indicated that pre- cision of estimates would not be appreciably affected by such a modification. For large enumeration districts (EDs), the segments were clusters of addresses from the 1960 Census Listing Books (later the correspond- ing books for 1970). For other ED’s, area sampling was employed and consequently some variation in the segment size occurred. To make the sample representa- tive of the then current population of the United States, the address or list segments were supplemented by a sample of housing units that had been constructed since 1960. i Within each selected PSU a systematic sample of clusters of housing units or segments was selected. The ED’s selected for the sample were coded into one of two economic classes. The first class, identified as the poverty stratum, was composed of current poverty areas that had been identified by the Bureau of the Census in 1970 based on information obtained prior to the 1970 Census plus other ED’s in the PSU with a mean income of less than $3,000 in 1959 (based on 1960 Census). The second economic class, the non- poverty stratum, included all ED’s not designated as belonging to the poverty stratum. All sample segments in ED’s classified as being in the poverty stratum were retained in the sample. For those sample segments in nonpoverty stratum ED’s, the selected segments were divided into eight random subgroups and one of the sub- groups remained in the NHANES I sample. Con- tinuing research during the NHANES 1 field collec- tion period indicated that efficiency of estimates could be increased by changing the ratio of poverty to non- poverty segments from 8:1 to 2:1. Therefore, in the later survey locations the selected segments in the nonpoverty stratum ED’s were divided into only two random sub- groups and one of the subgroups was chosen to remain in the sample. Adequate reliability for separate analyses of those classified as being below the poverty level and those classified as being above the poverty level was achieved through a disproportionate allocation of the sample among poverty and nonpoverty strata within selected PSU's. After identifying the sample segments, a list was made of all current addresses within the segment boundaries. A household member was interviewed to determine the age and sex of each household member, as well as other demographic and socioeconomic infor- mation required for the survey. If no one was at home after repeated calls, or if the household members re- fused to be interviewed, the interviewer tried to deter- mine the household composition by questioning neighbors. To select the persons in the sample segments to be examined in NHANES I, all household members ages 1-74 years in each segment were listed on a sample selection worksheet, with each household in the seg- ment listed on the worksheet in the order in which it had been listed by the interviewer. The number of household members in each of six age-sex groups (see table 2) were listed on the worksheet under the appro- priate age-sex group column. The sample selection worksheets were then arranged in segment-number - order. A systematic sample of persons in each age-sex group was selected to be examined using the sampling rates displayed in table 2. In general, this procedure resulted in only one per- son being selected from a household. However, in a few instances, more than one person was selected from a given household. This sampling strategy for the general sample of NHANES 1 resulted in the selection of 28,043 sample persons 1-74 years of age, a sample that can be regarded as representative of the target popu- lation displayed in table 1. In addition, a subsample of adults 25-74 years of age was designated to receive a detailed health exam- ination in addition to the general health examination given to all selected persons. This detailed sample was chosen systematically after a random start from the general sample of selected persons using sampling rates shown in table 3. For example, adults 45-64 years of age were subsampled for the detailed examination at a somewhat higher rate than 25-44 years of age from among all persons selected within cooperating house- holds. The mobile examination units were moved from one location to the next during a 39-month period (1971-74) to permit administering single-time exam- inations to the sample of persons participating in the study. These mobile units were moved throughout the North during the summer months and throughout the Southern areas in the winter months. Consequently, certain measures may reflect seasonal influences. The sample for the Augmentation Survey, adults 25-74 years of age selected for examination in 35 pri- mary units, also constitutes a national probability sample of the target population. Moreover, when con- sidered jointly with those selected for the NHANES I detailed examinations in the first 65 locations, the en- tire 100 location sample also represents the adult popu- lation at that time. The sampling frame for selecting the augmentation sample was the 1970 decennial census list of addresses and PSU’s. The methods for establishing the sample frame and selecting households were generally similar to those used in selecting the general sample. However, only 5 of the 15 superstrata (composed of only 1 very large metropolitan area of more than 2 million popula- tion) were drawn into the augmentation sample with certainty. The remaining 10 of these superstrata were collapsed into 5 groups of 2 each from which only 1 superstrata was selected. Thus, the probability of selec- tion for each of these 10 superstrata is 0.5, even though each of the 5 collapsed pairs is represented in the design with certainty. When these latter 5 locations are con- sidered a part of the 100 primary sampling unit design, they are selected with certainty. In this Augmentation Survey there was no eco- nomic stratification of enumeration districts and no oversampling among special age-sex groups. One of every two eligible persons within sample households (using a random start among those 25-74 years of age) was selected for participation in the survey. Nonresponse In a health examination survey, as well as any sur- vey involving volunteer participation, the survey meets one of its severe problems after the sample is identified and the sample persons are requested to participate in the examination. A sizable number of sample persons who initially are willing to complete the household in- formation, and possibly some of the medical history questionnaires (which are done in the household), usually will not participate in the examination. Full participation by individuals is determined by many factors, some of them uncontrollable by either the sample person or the survey personnel. For example, family health beliefs and practices, employment status, and access to transportation could affect participation in the survey. Because nonresponse is a potential source of bias, intensive efforts were made in NHANES I to develop and implement procedures and inducements to reduce the number of nonrespondents and thereby reduce the potential of bias due to nonresponse. These procedures are discussed in a Vital and Health Statis- tics series report.l Also during the early stages of NHANES I when it became apparent that the response rate for the examinations was lower than in the preceding health examination surveys, a study of the effect of remunera- tion upon response in NHANES I was undertaken. The findings, published by NCHS,!2 included remunera- tion as a routine procedure in NHANES I starting with the 21st and 22nd examination locations. Despite response rates of over 98 percent at the household interview stage and intensive efforts of per- suasion, only 20,749 or 74.0 percent of the sample persons from the first 65 survey locations were exam- ined. When adjustments are made for differential sam- pling for high-risk groups, the weighted response rate becomes 75.2 percent. Consequently, the potential for a sizable bias exists in the estimates from this survey. However, from what is known about the nonrespond- ents and the nature of the nonresponse, the likelihood of sizable bias is believed to be small. Using data from NHANES I and from an earlier survey, efforts have been made to examine possible health-related differences between examined and non- examined persons. An investigation of reasons for par- ticipation and nonparticipation in NHANES I was conducted by interviewing a sample of 406 people com- prised of 290 examined persons, 35 persons who had made appointments for the examination but who never came to the mobile examination center for the exam- ination, and 81 persons who refused to participate in the survey.!3 The sample persons for this study came from four survey locations: St. Louis, Monterey, New York, and Philadelphia. They were asked to indicate why they did not choose to be examined in NHANES I. The primary reasons given were that they had no need for a physical examination (48 percent), or that the examination times were inconvenient because of work schedules or other demands (15 percent). Only 6 per- cent of those persons who were not examined indicated that they refused the examination because of sickness, and 3 percent based their refusal on a fear of possible findings. Data on both examined and nonexamined (but interviewed) persons were analyzed by using informa- tion from the first 35 survey locations of NHANES 1.14 For the health characteristics compared, the two groups were quite similar. For example, 20 percent of the examined people reported that a doctor had told them they had arthritis, compared to 17 percent of the un- examined people. Similarly, 18 percent of both the examined and the nonexamined persons had been told by a doctor that they had high blood pressure. Twelve percent of both groups reported that they were on a special diet, and six percent of both groups said that they regularly used medication for nerves. In another study of factors relating to response in Cycle I of the Health Examination Survey, 36 percent of the nonexamined people viewed themselves as being in excellent health compared with 31 percent of the examined people.15 A self-appraisal of poor health was made by 5 percent of the nonexamined persons, and by 6 percent of those who were examined. In a different study of Cycle I findings, those who participated in the survey with no persuasion and those who participated only after a great deal of persuasion generally had few differences for numerous selected examination and questionnaire items.16 This was interpreted as evidence that no large bias exists between these two groups for the items investigated, and was offered as further sup- port for the belief that little bias is introduced to the findings because of differences in health characteristics between examined and nonexamined persons. Because of the nonparticipation of some sample persons in NHANES I, an adjustment procedure to account for nonresponse (similar to that used in pre- vious National Health Examination Surveys) was used. The reciprocal of the probability of selection of the sample persons is multiplied by a factor that brings estimates based on examined persons up to a level that would have been attained if all sample persons had been examined. This nonresponse adjustment factor was computed separately within relatively homogeneous classes defined by five income groups (under $3,000; $3,000-$6,999; $7,000-$9,999; $10,000-$14,999; and $15,000 or more) within each stand. The factor is the ratio of the sum of sample weights for all sample per- sons to the sum of sampling weights for all responding sample persons within the same homogeneous class. To the degree that groups can be defined which are homogeneous with respect to the characteristics under study, the nonresponse adjustment procedure can be effective in reducing the potential bias from nonresponse. In addition, a poststratified ratio adjustment procedure was employed to force agreement between the final sample estimates of the population and independent controls prepared by the U.S. Bureau of the Census for the noninstitutionalized population of the United States as of November 1, 1972 (the approximate midpoint of : the survey) for the cells shown in table 1. The combined adjustment factor for nonresponse and poststratification among the detailed examinees was 1.45 for the 65 PSU’s of the 1971-1974 period and 1.40 for the Augmentation Survey. For the 65-PSU sample of NHANES I, the percent: distribution of the adjustment factors used for the 325 cells (determined by the cross-classification of the five income groups by the 65 stands) is shown in table 4. Missing data and imputation Examination and other types of surveys in which multiple observations are made on the same person are subject to the loss of information not only through failure to examine all sample persons, but also from the failure to obtain and record all items of information for examined persons. When data for specific items are missing for some of the examinees, values for these items are often imputed to minimize the effect of such item nonresponse on population estimates. The issues relating to adjustments for missing data in surveys of this magnitude are complex and too numerous to discuss in this report. However, the adjust- - ments for relevant variables used in this research, par- ticularly the dental and blood pressure findings used as examples in the subsequent discussions, are of interest here. Dental findings were available for 20,218 of the 20,749 examinees in this NHANES I survey. Those 531 (2.6 percent) whose dental records were lost or not obtained through examinations were assigned im- puted values. Imputation of dental findings for an examinee was done by randomly selecting a record from among examinees of the same age in years, race, sex, and income group who had dental findings recorded. The values for this matched examinee were then im- puted for the missing items for the examinee with missing data. When data for income were not avail- able, the match was limited to age, race, and sex. These imputed values are included in all of the analyses in- volving the dental variables in this report. The age and sex distribution of the examinees with and without dental data from the survey is shown in table 5. Among the 13,671 examinees ages 18-74 years in the total, or nutrition, sample for 1971-1974, there were 76 (0.6 percent) examinees missing either the single measurement of systolic or diastolic blood pres- sure or both. Out of the 6,913 examinees ages 25-74 years in the detailed and augmentation sample, only 28 (0.4 percent) were missing measurements of either systolic or diastolic blood pressure or both in the first sitting position. For the recumbent position, 59 (0.9 percent) were missing measurements of either systolic or diastolic blood pressure or both, while for the second sitting position, 64 (0.9 percent) were missing measure- ments for either or both blood pressures. In no case was a diastolic measurement present without an accom- panying systolic measurement. For the statistical analysis of the blood pressure variables reported in Vital ‘and Health Statistics, Series 11-No. 203,17 imputed values for missing sys- tolic and diastolic blood pressures were assigned from the records of matched examinees with the same age, sex, and race, with similar arm girth, weight, and height. However, these missing value imputations are not recorded on the public use tapes; the imputation process would need to be repeated prior to statistical analysis of the data if identical analyses to those re- ported in the Series 11-No. 203 report were desired. Because there are so few of them, persons with missing blood pressures can be excluded from investigations of hypotheses involving these variables without seriously altering population inferences. Thus, to simplify the analyses in this report, records with missing data for blood pressure variables were excluded for estimates or hypothesis tests in which that required variable was missing. In general, missing data cannot be ignored in the analysis. For these analyses values were imputed for the missing dental variables and persons with missing blood pressures were excluded when the necessary value was missing. However, for variables with exces- sive rates of missing data (for example, greater than 1 percent), the data analyst must exercise caution in making estimates and drawing inferences from the survey findings. Design considerations for examined persons Although the sample design for this survey is de- scribed in extensive detail in the previous sections and in another document,! aspects of the design pertaining to data analysis considerations will be discussed further in this section. All 20,749 examined persons in the first 65 survey locations received a specifically designed nutrition-related examination. In addition, approxi- mately 20 percent of those ages 25-74 years (3,854 persons) received a more detailed examination con- cerning other aspects of health and health care needs. An additional 3,059 persons ages 25-74 years were examined in the 35-Jocation Augmentation Survey to increase the size of the detailed sample, and hence, the reliability of the estimates. The data collection forms for the entire sample, together with the additional forms for the detailed sample, are published elsewhere.18 Although the sample design for this survey was complex, the essential feature is the selection of pri- mary sampling units (PSU’s) consisting of counties or groups of counties from each of the defined strata. In particular, the NHANES I design involved the selec- tion with certainty of the 15 large standard metropolitan statistical areas with more than 2 million population. For data analysis purposes, several of the 15 cer- tainty strata were combined by NCHS to form only 10 strata. The data tapes from NCHS reflect this revised indexing of the certainty strata, although this recombi- nation of strata is not documented completely in previous NCHS publications. Each of these “certainty PSU’s” consists of a large number of enumeration districts which were treated as PSU’s. Each of the remaining 25 strata can be considered as being composed of exactly two PSU's. The Augmentation Survey discussed in Vital and Health Statistics, Series 1-No. 149 poses additional complications for analysis. The 3,059 examined per- sons selected for this Augmentation Survey represent a national probability sample of the target population when used as a separate 35-location sample. The Aug- mentation Survey can also, be combined with the 65 location detailed sample to form a 100-PSU national probability sample, in which the combined number of persons is 6,913. Of the PSU’s, 10 were included on both the Augmentation Survey and the initial survey. There was oversampling of the elderly in the initial detailed sample group (tables 2 and 3), but not in the Augmentation Survey. The number of PSU’s and the corresponding num- ber of examined persons in each of these strata for each of these survey components are summarized in table 6. Thus, for analytical purposes, this design can be char- acterized as having the following: 1) 10 strata with selection of segments as PSU’s and with multiple PSU’s for all survey components (survey locations 1-65); 2) 25 strata with a) paired selections of PSU’s for the general and detailed sample (survey locations 1-100); b) selection of a single PSU for survey locations 1-35 and for the augmentation sample (loca- tions 66-100). Throughout the remainder of this report, these paired or multiple selections will be referred to as sampling error computing units (SECU’s) indicating their role as basic units in variance calculations. For example, if all strata have exactly two SECU’s, a paired selection model involving squared differences of SECU totals within each stratum can be used to obtain Taylor series approximations to variances and covariances of sample estimators. Thus, if a particular design has exactly 2 PSU’s per stratum, these PSU’s play the role of SECU’s without further recombination. On the other hand, the NHANES I design summarized in table 6 requires that the multiple PSU’s in strata 1-10 be com- bined into two SECU’s each in order to have a paired design. Although the analyses of this report do not deal with 35-location design where only one PSU was selec- ted for the noncertainty strata, it should be noted that NCHS recommends that the 25-noncertainty strata be collapsed into 13-SECU?’s for variance computational purposes in the documentation available with the microdata tapes. Even though the overall number of examined per- sons in this survey is quite large, subclass analyses still can lead to estimators with unstable properties, particu- larly estimators of their variances based on Taylor series approximations for which the SECU sample sizes are small. For example, in the general sample the number of examined persons for the “other race’ cate- gory is extremely sparse in some of the strata as shown in table 7. Moreover, as shown in table 8, the number in some strata is quite sparse both for black people and those of other races in the detailed survey. Conse- quently, analyses by racial subclasses requires particu- lar attention to the coefficient of variation of the denom- inator for the estimators involving ratio means; for the detailed sample, certain analyses such as multiple re- gressions by racial subgroups may lead to serious com- putational difficulties or analyses of questionable relia- bility. This issue will be addressed further in subsequent sections. Another important aspect of the NHANES I design is the oversampling of the following subgroups thought to be at high risk of malnutrition: 1) Persons with low income; 2) Preschool children; 3) Women of childbearing age; and 4) Elderly persons. Adjusted sampling weights that reflect these unequal selection probabilities, in addition to the basic prob- ability of selection and the adjustments for nonresponse and poststratification, were computed and are on the public-use data tapes. An additional design complication arises because there was no oversampling of the subset of the sample persons ages 25-74 years who received the more detailed health examination. Women of childbearing age were not oversampled as they were for the major nutrition component of NHANES I. However, some oversam- pling remained among the elderly and poor people. There are separate adjusted sampling weights on the data tapes for the 3,854 persons given this detailed examination. Consequently, when computing estimates of ana- lytic statistics and their estimated variance-covariance structure, the appropriate sampling weights need to be utilized in the weighted analyses. Thus, in this report hypotheses involving variables from the initial detailed sample of persons ages 25-74 (survey locations 1-65) were investigated using the adjusted sampling weights associated with those sample persons. Analyses in- volving the augmentation detailed sample (survey loca- tions 66-100) used the adjusted sampling weights for this group. When hypotheses were investigated across the combined detailed sample groups (survey locations 1-100), a third adjusted sampling weight was used for the combined groups. Hypotheses involving variables from the entire nutrition-related initial sample (survey locations 1-65) utilized the adjusted sampling weights for that sample. Analytical strategies Because of the complexities in the sample design, an analysis could be performed in any one of at least three different ways depending on whether the sam- pling weights were used or whether the sample design features were incorporated in the estimation procedure. For simplicity, the following three options will be discussed: Use of sampling Option Weights Design features Vales suas nanvipere dees No No De os teak Meda a A Yes No oS LP gn Yes Yes Although the analyses could be performed under any of these options, it will be demonstrated that option 3 is more appropriate for making final inferences from these NHANES I data. However, as a practical matter, most hypotheses initially were investigated under op- tion 1, since the implementation of each option in suc- cessive order from 1-3 involved considerably more preparation and computing costs. Relationships found to be statistically significant under option 1 were sub- jected to more definitive analyses under option 3 utiliz- ing the sample weights and the survey design effects. Consequently, the estimated covariance structure for the sample estimators, based on the complexities of the survey design, was utilized in all final models and in- ferential conclusions. There is a certain risk associated with this sequen- tial strategy. Relationships found to be nonsignificant under option 1, the “screening stage,” may in fact be significant if the complex sample effects on the var- iances of the estimators actually reduce the estimated variances. Although this situation is rare in highly clustered data such as those obtained in the NHANES I, substantive relationships thought to be important should be investigated more rigorously under option 3, even if the statistical tests indicate the lack of significance under option 1. In survey research, the design effect is commonly defined as the ratio of the variance for a statistic from a complex sample to the corresponding variance from a simple random sample of the same size. These effects are used by survey designers and analysts for a variety of purposes. Frequently the design effect has been used to summarize conveniently the effects of a complex sample design on the precision of estimates from the survey data and to specify design features for new sur- veys. Increasingly, design effects are being used to adjust estimates and statistics computed under simple random sampling assumptions for the effects of the complexities in the sample design on measures of pre- cision. Given the importance of these effects to those designing and analyzing surveys, simple useful models have been sought for design effects. Such models are useful for deriving estimates of design effects for statis- tics for which they are not available and for suggesting methods to adjust estimates computed under the as- sumption of independent selections for complexities in the sample design. A review of these design effect con- siderations and analytical strategies for survey data from complex sample designs was presented by Lep- kowski.19 Throughout this publication, the estimated design effects will be shown to illustrate the importance of these effects in definitive hypothesis tests or model fitting calculations. All analyses under option 1 can be performed quite simply and relatively inexpensively using standard sta- tistical software packages. In this option sampling weights and design effects are totally ignored. Thus, the data are regarded as coming from a simple random sample with equal probability of selection for every element in the population. Analyses under option 2 in- corporate the sampling weights in estimating the analytic statistics, but simple random sampling computations are still utilized as under option 1 for the variance estimation. Analyses under option 3 utilize both the sampling weights and the complex sampling design in calculating the esti- mates and the estimated variance-covariance structure of analytic statistics. The calculations for options 2 and 3 were performed with the OSIRIS IV software pack- age developed by the Computer Support Group within the Survey Research Center of the Institute for Social Research at the University of Michigan.4 Alternatively, other statistical software packages could be used if they can incorporate the sampling weights and the design structure into the analysis. : In particular, for this report the computer program &PSALMS was used for estimating ratio means and the program &REPERR was utilized to fit regression models. For relatively simple statistics such as ratio means, differences of ratios, and totals, the &PSALMS routine approximates the complex sample variance of these estimators using a linearized Taylor series expan- sion. For more complex statistics, such as regression coefficients, either a balanced half sample (BHS) or a Jack-knife replicated variance estimation procedure is available. The BHS option within the &REPERR rou- tine was utilized to fit simple and multiple regression models to the NHANES I data. Both of these routines are available within the OSIRIS IV library, and are described in more detail by Vinter.20 Because of the multiple SECU’s within the certainty strata 1-10, the estimation procedure to implement op- tion 3 can be extremely time consuming and expensive, particularly if replication procedures are used to fit re- gression models. On the other hand, if each stratum has exactly 2 SECU’s, the BHS approach to fitting regres- sion models is straightforward and economical. To alleviate these cost and computing time diffi- culties, the multiple SECU identification codes in each of the certainty strata (i.e., 1-10) were randomly allo- cated into 2 pseudo-replicates within the stratum. Con- sequently, the paired selection computation procedures could be utilized across all 35 strata for all statistical analyses, not just those involving multiple regression. The effects of randomly assigning the multiple SECU’s to two paired pseudo-replicates was investigated by comparing standard errors and design effects for esti- mates of proportions and means within the age groups shown in table 10 for variables such as decayed, missing, and filled (DMF) teeth, systolic blood pres- sure (SBP), and calories. The means and standard errors were computed under the multiple SECU clas- sification scheme and under the paired SECU group- ings. For these variables, it is apparent that the random allocation of SECU’s in the certainty strata to form a complete paired design has not substantially altered the estimates of variances or the corresponding design ef- fects for overall means and subclass means. As a result of this pairing for the 10 certainty strata, all variance-covariance computations can be obtained directly as appropriate sums of squares and cross- products of differences between SECU or replicate totals across the 35 strata in the initial sample utilizing 70 paired SECU’s. Consequently, all the analyses under option 3 for the data from the 65 survey locations were performed assuming this paired selection design. On the other hand, analyses under option 3 for the com- bined data from the detailed and augmentation surveys (the 100-1ocation survey) require a multiple selection model for variance computations because the design cannot be paired for the 25 noncertainty strata; each of the strata 11-35 have 3 SECU’s in this combined design. Consequently, when combining the data from the detailed and augmentation survey, the user needs to utilize a variance-covariance estimation procedure that permits multiple SECU’s per stratum. For example, either the multiple selection model in the &PSALMS program of OSIRIS IV at the University of Michigan or the replication methodology discussed by Gurney?! can be used for these calculations. Continuous variables: Means Means and standard errors were estimated for sev- eral variables to investigate the relative effects of the sampling weights and the sampling design on the esti- mates. These results are displayed in table 11 for four variables—number of decayed, missing, and filled teeth, systolic blood pressure, calories consumed daily, and age. For the total sample, the unweighted and weighted analyses (options 1 and 2) for these variables are similar for the means and variances. However, the complex sample design introduces a considerable increase in the estimated variance of the mean (option 3). The ratio of the standard error of the mean under option 3 to that obtained under option 2 (shown in the last column in table 11) ranges from 1.71 to 2.73. Consequently, the design effects range from 2.92 to 7.43. One might expect the design effects to be smaller when stratifying into subclasses such as age groups. This expected reduction is due to the clustering effect which is both a function of subclass size as well as the homogeneity coefficient. The latter is the extent to which persons in the same subclass tend to have similar responses within clusters. Thus, unless the homogeneity coefficient increases for smaller subclasses, the design effect will be smaller for the age subclasses than for the overall sample. To investigate this possibility, means, standard deviations, and standard errors of the means of these variables were computed within age groups shown in table 12. Although the design effects are somewhat reduced, they are not negligible, ranging from 1.48 to 5.07. Subgroup comparisons: Means Many hypotheses involve the comparison of two subgroup means. Because of the clustered design and the sampling weights, the difference between the mean response for each subgroup was computed as the difference between two weighted ratio means within the context of the &PSALMS routine described by Vinter.20 To assess the effects of the sampling weights and the complex sample design on the magnitude of the ¢- statistics associated with the tests for these differences, a representative analysis was investigated under op- tions 1-3. In particular, the mean systolic blood pressure was compared for two subclasses determined by the lowest 15th percentile and highest 15th percentile of skinfold thickness in selected age by race subgroups. These results are shown in table 13 under each of the three analysis options. In all subgroups, the simple random sample estimates for the unweighted and weighted analyses are similar, both for the means and variances. However, the complex sample design intro- duces a considerable increase in the estimated variance of the difference in the means between the two sub- classes. Specifically, the ratio of the standard error of the difference of the mean under option 3 to that obtained under option 2 in the last column in table 13 ranges from 1.3 to 2.0. Thus, the design effects for estimated means range from 1.7 to 4.0. In other words, the t-statistic computed under option 2 is from 1.3 to 2.0 times larger than that computed under option 3 because the variances under option 3 are larger. Continuous variables: Multiple regression models The basic model used for assessing the joint effects of several predictor variables on the variation of a continuous response variable is the multiple regression model. The general super-population model is Y,=B, + B,X,; + B3X;, +. i +B JX, +E; (1) where Y; denotes the ith observation of the dependent variable, X,; denotes the ith observation of the kth independent or explanatory variable, and E; is the random variation of the ith observation of Y. The subscripts 2, 3, . . ., k identify the specific explanatory variables. B, is the intercept term, and B, is the change in the expected value of Y; corresponding to a unit change in the kth explanatory variable, holding all other explanatory variables constant. B,, Bs, . . ., B, are often referred to as the regression slopes or (partial) regression coefficients. Alternatively, multiple regression models can be developed in terms of standardized independent vari- ables. This approach leads to standardized estimators usually referred to as beta coefficients. The beta coef- ficients are the result of a linear regression in which each variable is “normalized” by subtracting its mean and dividing by its estimated standard deviation or sum of squares about the mean. In other words, the beta coefficient adjusts the estimated slope parameter by the ratio of the standard deviation of the independent variable to the standard deviation of the dependent variable. In this formulation, the model does not have a constant or intercept term. A beta coefficient of 0.3 may be interpreted to mean that a standard deviation 10 change of 1.0 in the independent variable will lead to a 0.3 standard deviation change in the dependent vari- able. Beta coefficients are also used to make statements about the relative importance of the X variables in the model. Assumptions of the multiple regression model The classical assumptions associated with the regression model are 1. The model specification is correct. 2. The X’s are nonstochastic. In addition, no exact linear relationship exists among two or more of the independent variables. 3. The E; are independent, identically distributed as N(0,02). Any set of real data is unlikely to meet all of these assumptions, particularly one utilizing a complex survey design such as the NHANES I. However, certain violations of these assumptions may not seriously affect statistical inferences. For example, under simple random sampling arguments, it is straightforward to show that the least squares estimators of the regression coef- ficients retain their desirable asymptotic properties (unbiased, consistent and efficient) when the X’s are stochastic (i.e., a violation of the second assumption) provided that the explanatory variables are each dis- tributed independently of the true errors in the model (see, for example, Kmenta22). More detailed discus- sions of the properties of regression model estimates from complex sample surveys can be found in Holt, Smith, and Winter.23 Specification error If any variables are omitted from the regression equation that are correlated with both the dependent variable and the independent variable(s) included in the model, the estimates of those regression coefficients will be biased. This particular problem is the reason a multivariable (rather than a series of bivariate) estima- tion procedure may be required when investigating a phenomenon that has multiple, interrelated causes. For example, in the relationship between dietary intake patterns and dental caries experience, if a variable such as age is omitted, biased estimates of that relationship emerge because there is a correlation between age and dental caries experience. In spite of the effort to include all of the theoretically important variables in the model, if some have been omitted, either because they were not part of the data collected or theory has not yet advanced sufficiently to implicate them, the estimators given by the model could be biased. Another concern with specification error is the actual mathematical relationship between the response variable and the joint distribution of the independent variables in the model. If the true relationship is, for example, logarithmic, the specification of the model as linear may lead to biased and inconsistent parameter estimates. Therefore, careful attention to all available theoretical knowledge concerning the relationships involved is essential. Some of the relationships studied might be better represented by a series of simultaneous interdependent equations. For example, the symptoms of periodontal disease could influence the frequency of dental visits, dental visits could influence toothbrushing behavior, and toothbrushing could affect periodontal disease. In such a circumstance, ordinary least squares estimation of individual equations can lead to biased and incon- sistent parameter estimates. While these forms of pos- sible misspecification probably do not pose a serious threat to the conclusions reached by Burt,” they do warrant future exploration to more precisely assess the underlying form of these relationships. Measurement error When variables are measured with error, they can affect the results of statistical procedures applied to them. In general, considerable effort was expended to ensure a minimum of observer error in the gathering of NHANES I data. Potential problems concerning some variables and the procedures employed to minimize some of these are described in Vital and Health Statistics, Series 11-No. 225.7 Consider, for example, the group of nutritional and dietary variables from the 24-hour recall record. There are short-term and long- term variations in what people eat. Therefore, the 24- hour recall record is an imperfect measure of long-term dietary patterns. This kind of random error in an independent variable in a regression equation will bias the estimate of the regression coefficient of that variable toward zero. Under simple random sampling argu- ments, it is possible to demonstrate that the form of the bias is B' =B/(1 +) where B’ isthe biased estimate of the regression param- eter as computed by ordinary least squares, is the unbiased estimator, and is the ratio of the true variance to the additional variance attributable to the measurement error. > (See for example, Snedecor and Cochran.24) The extent to which the bias in estimates of regression coefficients can be expressed in this formulation under the complexities of option 3, utilizing both the sampling weights and the survey design effects, requires further investigation. . Because some empirical work has provided esti- mates of the ratio of interindividual (true between subject variation in individual intake) to intraindividual (day-to-day variation in individual intake) variation, rough estimates of this bias are possible.25 These data suggest that values of A of at least one or two are not unreasonable. Based on this information, the relation- ships between dental caries experience and diet in one 24-hour period are, as estimates of the relationship between dental caries experience and lifetime dietary patterns, underestimates by a factor of 1/2 to 1/3. Stated another way, estimates based on lifetime data are likely to be two or three times larger than those provided by the 24-hour data. When a variable with this type of error is used as a dependent variable, as in the investigation of the effect of dentulous state on dietary patterns, the problem encountered is less severe. Standard errors will be overestimated, but estimators will be unbiased. There- fore the only real hazard is the failure to reject the null hypothesis when it should be rejected. Heteroscedasticity When assumption 3 is violated, standard errors estimated by ordinary least squares tend to be inef- ficient. Because the variance of variables such as DMF and PI measures tends to increase with age, the pos- sibility of this phenomenon influencing the results presented should be investigated. Weighted least squares procedures may be required when heteroscedasticity is a problem. The extent to which this correction procedure is sufficient under the complexities of option 3 requires further investigation. Nonnormality of random variation term Dependent variables such as DMF teeth and PI have distributions that are skewed toward zero in the younger age groups. The random variation term is, therefore, not normally distributed. In some instances, transformations may be employed to provide reasonable approximations of normality. In others, where trans- formations are of little value, it still may be possible to employ multiple regression models as though the dis- turbances are normally distributed, because the pro- cedure is considered to be relatively robust when sample sizes as large as occur in analysis of the NHANES I are used. Empirical results for regression models To investigate predictive relationships among con- tinuous variables, multiple regression models also can be fitted under either option 1, 2 or 3. Specifically, the effects of the sampling weights and complex design on the precision of regression coefficients were investigated 1" under options 1-3 for the number of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP) and calories consumed daily regressed on age within race-sex subclasses as summarized in table 14. First, it can be observed in the corresponding entries under options 1 and 2 that the results are similar, particularly for DMF teeth on age and SPB on age. These both have a strong linear relationship in all the race-sex subclasses. However, for calories on age, which has extremely small R? values for all subgroups, the estimate of the slope is quite different for some subclasses. For the “other males” there is a 12-fold increase in the slope under option 2 compared to option 1 and for the “other females” it differs by a factor of nearly three. Of course, in both of these subclasses the sample size is relatively small. The square root of the design effects for the slope parameter estimates in these simple linear regression models for the white race data are displayed in the last column of table 14. In particular, these quantities range from 0.98 to 1.85 for each of these three variables; the design effect is smaller for men than for women. In table 14, the results under option 3 are reported only for the white subgroups even though the number of black persons examined appears to be reasonably large. This omission is due to the failure of the balanced half sample routine in the weighted regression program when entire strata had no data in an early version of the &REPERR program in OSIRIS IV (this routine now has been modified to allow for empty strata). The problem of missing data for black persons in some SECU’s as shown in table 7 is even more pronounced within the more restrictive detailed examination, as displayed in table 8, and for persons of other races. Consequently, due to the sparse design across strata, only the data for the white and black races were used in many of the analyses. In addition to simple linear regression models, multiple regression models also can be fitted within this same framework. As discussed previously, the paired SECU’s for each of the 70 strata were utilized in the balanced half sample routine &REPERR to generate estimated variances for the estimated slope parameters. Table 15 summarizes the results for 6,349 persons ages 11-30 years of DMF regressed jointly on age (in single years), race (1 = white, 2 = black), sex (1 = male, 2 = female), and sweets, which is the sum of the reported frequencies for the ingestion of food from the three categories of desserts and sweets, candy, and beverages (sweetened, carbonated, and noncarbonated). In this model, the design effects for the regression coefficients range from 1.25 to 4.49. Similarly, the results of a multiple regression model for 13,573 people ages 18-74 years of systolic blood pressure regressed jointly on age, race, sex and Quetelet’s Index of body mass expressed in kg/cm? units are displayed in table 16. Here again, the design effects for the regression coefficients range from 2.69 to 3.61. 12 These empirical results, expressed in terms of estimated design effects, demonstrate the effects of incorporating the sampling weights and the survey design adjustments into multiple regression models. The decision of when to incorporate sampling weights and design features into the analysis depends on more than a recognition of the potential errors in inference that can arise because of such effects. Some analysts argue that when making a model-based inference from survey data about a super-population model, one may ignore the sampling design features, even in a design as complex as NHANES I. However, many survey prac- titioners argue that a design-based inference, as illus- trated here, is more appropriate for survey data, espe- cially when examining exploratory models for which the specification of the model is likely to be in error. Accounting for unequal probabilities of selection and other design features in the design-based approach recognizes that the model may be misspecified and that somewhat conservative inferences are desired. Further, the model of interest in the design-based approach is appropriately one in which the model refers directly to the finite population from which the sample was selected. In this and subsequent sections, the analytic perspective for survey data is the design-based view of inference for complex sample survey data. Continuous variables: Analysis of variance The familiar analysis of variance (ANOVA) situa- tion involves a set of factors X, X5, ...,X, each of which may have several levels. These factors are used to explain the variability in a response variable Y. In general, an appropriate measure of total variation for Y, such as the total corrected sum of squares, is partitioned into individual components each attributable to a factor or group of factors. The usual hypothesis tests require the assumption of equality of variances and zero covariances among subgroups and the assumption of simple random sampling. ANOVA for data from complex surveys such as NHANES I requires alterna- tive considerations. Because of unequal probabilities of selection, the clustered design, and the adjustment weights for nonresponse and poststratification, the mean response for each subclass (or domain), deter- mined by the cross-classification of the relevant factors, is computed as a weighted ratio mean. Consequently, the variance-covariance structure of these weighted ratio means must be incorporated into the ANOVA tests when attempting to identify the statistically im- portant sources of variation. One approach is the large sample methodology utilizing weighted least squares algorithms for the computation of Wald statistics originally described by Grizzle, Starmer and Koch26 for the analysis of multi- variate categorical data. This general methodology was modified and applied to data from complex sample surveys in a series of papers26-30 using data from another NCHS Survey, the National Health Interview Survey. A. brief outline of the application of this methodology to data from the NHANES 1 is presented in appendix I of the Vital and Health Statistics Series 11-No. 209,31 Koch and Stokes32 and Koch, Stokes, and Brock.33 In essence, this strategy involves a vector F of subclass or domain ratio means, together with an appropriate, valid, and consistent estimate of the co- variance matrix Vr of these means, and the framework of a general linear model. Consequently, the usual ANOVA hypotheses about which factors or combina- tions of factors make statistically significant contribu- tions to the variation among these domain means can be investigated by fitting linear models to the vector of means by the method of weighted least squares relative to the estimated covariance matrix. Quite a few different approaches can be used to estimate the covariance structure of the ratio means. One method is to use the balanced repeated replication (BRR) strategy described by McCarthy34 and Kish and Frankel.35 Several different variations of this replication approach were investigated empirically within the context of National Health Interview Survey (NHIS) data as reported in Freeman, Freeman, Brock and Koch.29 On the other hand, for paired designs in which there are exactly 2 SECU’s within each stratum such as the NHANES I design, as well as for other multistage sample designs, direct methods involving sums of squared differences and cross-products can be utilized to obtain estimates of the variances of the numerators and denominators of the ratio means. These variances and covariances can be incorporated directly into a linear Taylor series expansion for the estimated co- variance structure of the ratio means. This particular direct approach to the estimation of the covariance matrix is described for contingency table proportions expressed as ratio means in Lepkowski and Landis36 and Lepkowski.37 These calculations are directly anal- ogous to those for ratio means in general. The ANOVA results presented here were obtained in two stages of computing and data analysis. First, the vector of ratio means of the dependent variable, together with their estimated variances and covariances, were computed directly within the OSIRIS.IV package using the &PSALMS routine. Any computing algorithm de- signed to generate a vector of ratio means and a consistent estimate of its covariance structure under the complex sampling design can be utilized to obtain these estimates. At the second stage of computing, the vector of sample means and its covariance matrix were entered directly into the weighted least squares pro- gram, GENCAT (see Landis, Stanish, Freeman, and Koch38) to perform the various ANOVA hypothesis tests and final model-fitting computations. The specific command files used to generate the results for the ANOVA example in the subsequent section are listed in appendix II. ANOVA methodology Consider a linear model for the vector of g subclass or domain ratio means F' as ELF) = XB, (2) where X is a (g X u) matrix of known constants with u Yi ( 1 2) and n,= > (13) From these sample totals, the proportion of elements in subclass i and response category j can be estimated as P= n;/n;. (14) The denominator of (14) is not fixed by the sample design and is therefore a random variable. Hence, p;; is a ratio mean and subject to some theoretical diffi- culties. For one, the variances and covariances of the p;; generally are not known exactly. In practice, however, afirst order Taylor series expansion approximation can easily be computed for the variance of ratio means, provided by the &PSALMS routine in OSIRIS IV. For example, the estimated variance of ( 14) can be approx- imated by var (p;) = (p;)?[(n;)~2 var (n;) + (n;)~2 var (n,) = 2(n;n; )" cov (njn;)l, (15) where the variances and covariances var (n;), var (n;), and cov (n;;,n;) are estimated in a manner consistent with the sample design. If, for instance, there are exactly two primary selections in each stratum (i.e., a,=2forh=1,..., H),the variances and covariances of the cell and subclass totals can be computed as > wip = Rina)? (16) = > wim — Nina)? (17) var (n;) = var (n;) and = Rijn2)(Rin1 = Min2)- (18) An approximation similar to (15) can be given for the covariance of two proportions. In particular, cov (n;,n;) = 3 cov (ppiy) = pig) [(ynyp)~" cov (n,n; + (nin; )"" cov (n;,n;) — (yn, cov (nin) — (n;.n;)"" cov (n; ,n;)). Replication procedures also can be used to compute estimates of these variances and covariances. The estimation procedures for the proportions p; and the corresponding variances and covariances can be summarized compactly in vector notation. For each sample element, let the (sr x 1) vector of observed values y;.« be denoted as * Yamard: (19) Summing these vectors across elements in each primary selection, the weighted cluster totals are obtained as Yha = 24 Y hake (20) Let K denote an [s(r + 1) x sr] linear operator matrix that can be used to generate a vector of cell and subclass totals for each cluster from the Jha: Specifically, let I, K= ir %7} (21) Yhok = IY 1100200 Vi 2nakes + + where I, and I; denote sr and s dimension identity matrices, respectively, 1, denotes an r dimension vector of one’s, and ® denotes the Kronecker or direct product (see section 8.8 of Graybill).4! Then the [s(r +1) x 1] vector of cell and subclass totals denoted as ny, can be expressed as Phe = Kya. (22) Summing the cluster vectors n,, across clusters, the [s(r + 1) x 1] vector of weighted cell and subclass totals is n= 202 ltra (23) or, n = [141 .s oo Nsplly, . wis): As noted previously, the variances and covariances of the elements of n depend on the nature of the sample design. If, as before, a paired selection type of design is used, then the [s(r + 1) x s(r + 1)] variance-covari- ance matrix of # can be estimated as Va = > 7 Ryo) — po). (24) A more complete discussion of the conditions under which estimators such as (24) are appropriate is given in Kish and Hess.42 A series of transformations can now be applied ton and V, to obtain an (sr x 1) vector of the sample proportions Dj»and an (sr x sr) matrix of the variances and covariances of the p;. Let exp(-) and log(-) denote matrix operators which take the natural exponent and logarithm, respectively, of every element of the matrix argument (+). Also let 4 denote the [s(r +1) x sr] linear operator matrix A= [Z| wl, ® 5): (25) Thenthe (sr x 1) vector of sample proportions, denoted Pp, can be obtained as p = exp [A(log (n))]. (26) The Taylor series approximation to the variance- covariance matrix of 2 denoted V,, is the transforma- tion of V/, =D, AD;"V,D;'A'D,, (27) where D, and D,, denote (sr x sr) diagonal matrices with the tlements of the vectors p and n, respectively, along the diagonal. The matrix operations presented in expressions (19) through (27) are straightforward generalizations of the results in expressions (10) through (1 8). In addition, compounded sets of transformations, as in (26), can be developed for the vector p to obtain functions of the sample proportions, such as f3(p), . JP), with their corresponding Taylor series approximations to the variances and covariances as discussed in Forthofer and Koch43 and appendix I of the report by Koch, Landis, Freeman, and others.44 Thus, a vector of functions of sample proportions, denoted as F(p) = [fi®); . . ../4(p)], and its variance-covariance “matrix approximation V, can be computed directly from the “raw” survey data. The utility of this method is enhanced by the ability to obtain estimates of V,, in (24) for other survey sample designs besides paired selections. Multistage designs with multiple clusters in each stratum, stratified random samples, or systematic selections of clusters, are a few of the other designs that can be handled by this method. The integration of this computational procedure for categorical data from sample surveys with the flexible WLS procedure for analyzing categorical data has been implemented by Lepkowski37 and, with a slightly different computational procedure, by Freeman.45 Pro- vided a consistent estimate of the variance-covariance matrix of the vector of functions F(p) is available, the WLS procedure outlined in ANOVA methodology can be applied directly to the function vector and its covari- ance matrix associated with the contingency table. Contingency table example The data in table 21 were obtained from the cross- classification of current cigarette smoking status, race, and Periodontal Index. Since smoking history is avail- able only for those individuals included in the detailed 18 survey, the usable sample size is limited to a maximum of 3,854 adults ages 25-74. For this detailed sample, smoking data was available for 2,948 persons. After eliminating persons whose race was not white or black, there were 2,919 examined persons with complete data on these variables. The primary hypothesis addressed in this analysis is the relationships between the factors, cigarette smok- ing and race, and the response variable, periodontal index (PI). However, as discussed in considerable detail in Vital and Health Statistics, Series 11-No. 225,6 the relationship between PI and such factors as smoking and race are highly influenced by other co- variates such as frequency of tooth brushing. Conse- quently, the analyses of these frequency data will illustrate the contingency table methodology. They do not suggest substantive conclusions. The weighted frequency distribution of PI score and the proportion classified PI (Some) for each of these subclasses is shown in table 22. They are based on the weights for the detailed survey (tape location 170-175 on all public-use tapes) after standardization of the weights to sum to the total number examined in the detailed sample. In this context, it is critical that the weights be standardized to the number of examined persons in the detailed sample. Otherwise, the weighted frequencies will be population estimates and the analysis will not be conducted relative to the sample sizes actually utilized in the survey. In particular, note that the weighted frequencies for black people are all smaller than those actually examined. This reflects the oversampling in the design. Overall, the proportion of people estimated to have some periodontal disease is approximately 4 percent lower using the weights as compared to the unweighted estimates displayed in table 21 (51.8 percent vs. 55.7 percent). To incorporate the effects of the complex sample design in the analysis of these contingency table data, the proportions in table 22 were also computed under option 3 as a vector of ratio means using the method- ology discussed in the previous section. Consequently, the Taylor series-based variance estimates for these proportions were generated directly from the &PSALMS routine for this analysis. These results, together with the estimates obtained under options 1 and 2, are shown in table 23. As noted in the last column, for these proportions the square root of the design effects ranges from 1.22 to 1.88; the design effects range from 1.48 to 3.53. Note that the design effects for the proportions with PI (Some) are smaller for the current cigarette smokers than for those not smoking, regardless of race. For purposes of model fitting and hypothesis test- ing, the vector F of the subclass proportions with PI (Some) and its corresponding covariance matrix Veare shown in table 24. Initially, the variation among these proportions was investigated using the usual 22 factorial design matrix X, in the linear model formulation E,(F)=X,B;, where ir f°1 "1 1 -1 —1 X, = 1 -1 1-1 1 -1 —1 1 and Overall mean Differential effect for whites 5 Differential effect for smokers Interaction effect The estimated parameter vector for this model is obtained from the weighted least squares routine as b, = (0.606, —0.095,0.029, 0.033). Thus, hypotheses of the form H,:CB = 0 can be used to investigate the relative importance of these factors and their interaction in contributing to the variation among the proportion with PI (Some). Specifically, the three hypotheses for this model, together with their corresponding contrast matrices and resulting test statistics, are shown in table 25. For comparative purposes, these hypotheses were also tested under options 1 and 2, using the frequency data from tables 21 and 22, respectively. As shown in table 26, the importance of incorporating the design effects into the test criteria is quite pronounced. In particular, under option 1, each of the 3 hypotheses would have been rejected at the usual 5 percent level of significance. Although the race effect is highly significant, there is evidence among the estimates in F' and from the marginally significant test (p = 0.09) for H; in table 25 that smoking has a differential effect across race subclasses. To investigate this possibility formally, the variation among the estimates was characterized by the linear model E,(F) = X,B,, where 1 1 1 1 1 - X:=1y -1 0 1 =1. 0 Overall mean Differential effect for race Smoking effect for whites B,= As shown in table 27, the goodness of fit statistic (i.e., @ = 0.01) indicates that X, provides an adequate characterization for the variation among these propor- tions. Moreover, the smoking effect for the white population is highly significant (Q = 38.63) compared to the nonsignificant average smoking effects indicated in table 26 under X,. Moreover, the square root of the design effect is less than 1.05 for this smoking test statistic, whereas it was 1.45 under X,. Several remarks of caution about the model build- ing approach used here are appropriate. First, the reduced model could be criticized from the vantage point of “overfitting”” models to data. Clearly, the lack of fit statistic is extremely small. On the other hand, the model is consistent with the data in that the proportions are nearly identical for the black population regardless of smoking status. However, the objective of the model building is to find a linear model that adequately describes the variation in observed proportions and offers substantively appealing explanations for the rela- tionships among the variables of interest. Thus, the reduced model is in a certain sense ‘“‘overfitted.” However, it offers the substantive expert insight to complex relationships through a relatively straight- forward linear model framework. Second, the model was not obtained by successively fitting models until the best one was discovered. From a classical hypothe- sis testing point of view, one should always investigate H, the interaction hypothesis, prior to testing H; and H,, the main effects hypothesis. However, in this case, a more informative model was proposed by noting the “nested” effect of smoking. Thus, the issues of hier- archical testing should be considered carefully when proceeding with model reduction. Finally, the appro- priate significance level to be applied for the individual hypothesis tests in the model building is not clearly specified. Generally a “level” of 0.05 has been used as an acceptable criteria in choosing among models. But the particular significance level is arbitrary in a model building framework. The reduced model is the result of careful examination of the observed proportions in each subclass and the goodness of fit and hypothesis test statistics. The application of formal hypothesis testing is inappropriate in such an approach, but the methods and terminology are utilized nonetheless. Summary The analysis of data from large complex sample surveys is not a straightforward task. The analyst must consider many issues to develop an appropriate and efficient strategy for conducting the analysis. Such con- siderations should include not only the technical issues of the sample design, weights, or underlying assump- tions in the analytic procedures to be applied, but the fundamental inference issues concerning the nature of models to be developed from the data. Perhaps the first consideration should be inferential: are inferences to be made to the finite population from which the sample was selected or to some super- population or theoretical model of which the finite population may be a single realization? The two ap- proaches to inference for survey data—model-based and design-based—each offer the analyst difficult choices. The choice of a model-based inference leads to analytic strategies that ignore the complexities of the sample design and allow analysts to use routine statistical soft- ware for calculations. On the other hand, the model- based inference often requires stronger assumptions for the particular problem than does design-based inference. Further, the model is assumed to be perfectly specified. The approach presented here has been a design- based type of inference. Computationally, the design- based approach to inference is more difficult to develop and more costly to apply than is the model-based ap- proach, But many survey practitioners feel that it offers advantages when developing exploratory models from survey data. For example, consideration of weights, which account for unequal probabilities of selection adjustment for nonresponse, and adjustment for cover- age errors in the analysis, can protect the analyst from some types of misspecification error. Further, with use of variance and covariance estimates, which account for the complexity of the design, design-based inferences tend to be somewhat conservative compared with the model-based approach. ~ Some analysts argue that an hypothesis testing framework for survey data from a finite population con- cerning finite population parameters is inappropriate. 20 The perspective taken here has been that if the com- plexity of the survey design is taken into account, sub- stantively useful inferences are possible by applying existing model building methodologies to survey data. If a design-based inference approach is chosen, the technical details of the sample design must be con- sidered. The discussion in previous sections has indi- cated that the following design features ought to be carefully considered: eo What is the nature of the sample design? Was a stratified multistage sample design used? Were un- equal probabilities of selection applied? eo Were there adjustments for nonresponse or cover- age errors? Is there a weight variable or are there several weight variables that must be applied when different parts of the sample are analyzed? ® Are there important measurement issues that could affect survey analyses? Is item nonresponse an im- portant problem for some variables? Do inter- and intra-observer variability contribute to errors in the data? eo Given the sample design and various sources of error present in the data collection operations, how can estimates be formulated? How can such estima- tion procedures be incorporated into existing ana- lytic procedures? How can the results be interpreted, and what kind of inferences are appropriate in view of the complex survey design? These are not all of the issues that can or should be raised in this context. In many instances the analyst probably should consult with a survey practitioner or sampling specialist to resolve the technical issues. The effects of a complex sample design on inference can be quite dramatic, as illustrated in previous sec- tions for several types of analytic procedures (e.g., re- gression analysis, ANOVA). In most analyses, design effects are not negligible, even for means within sub- classes, regression coefficients, or chi-square criteria computed from contingency table analyses. On the other hand, the costs of computing can be large when the complexity of the sample design is used in the esti- mation process. Based on findings in this publication, the following recommendations about analytic strategy are suggested for users of NHANES I data from public use tape: ® A design-based inference, although difficult and costly to apply, is appropriate for such data. Esti- mation procedures, which account for the complex- ity of the survey design, should be used for the final analysis. e Investigate all preliminary hypotheses without re- gard to the design effects. Since estimated means and other statistics may change greatly when sam- pling weights are considered, sampling weights and weighted estimates should be used. Based either on significant results at the previous step or on relationships thought to be important from previous substantive considerations regardless of whether they were significant at the previous step, proceed with a more rigorous analysis using both the appropriate weights and sample design effects. Such a two stage design-based inference approach will be both less costly and more appropriate than other strategies that could be applied to the NHANES I data. 21 References INational Center for Health Statistics, H. W. Miller: Plan and operation of the Health and Nutrition Examination Survey, United States, 1971-1973. Vital and Health Statistics. Series 1-No. 10a. DHEW Pub. No. (PHS) 79-1310. Public Health Service, Wash- ington. U.S. Government Printing Office, Dec. 1978. 2Kish, L., Groves, R. M. and Krotki, K.: Sampling Errors for Fer- tility Surveys, WFS Occasional Paper No. 17. The Hague, Inter- national Statistical Institute. 1976. 3Verma, V., Scott, C. and O’Muircheartaigh, C.: Sample designs and sampling errors for the World Fertility Survey. J R Stat Soc A. 143(4):431-473, 1980. 4Survey Research Center Computer Support Group, OSIRIS IV User’s Manual, Institute for Social Research, 1979. SHidiroglou, M. A., Fuller, W. A., and Hickman, R. D.: SUPER- CARP. Survey Section of the Statistical Laboratory, Iowa State University, 6th ed., 1980. 6SAS Institute, Inc.: SAS User’s Guide, 1979 edition, J. T. Helvig and K. A. Council, eds., Raleigh, N.C. SAS Institute, Inc. TNational Center for Health Statistics, B. A. Burt, S. A. Eklund, J. R. Landis, and others: Diet and dental health, a study of relation- ships, United States, 1971-74. Vital and Health Statistics. Series 11-No. 225. DHHS Pub. No. (PHS) 82-1675. Public Health Service. Washington. U.S. Government Printing Office, Jan. 1981. 8National Center for Health Statistics, W. R. Harlan, A. L. Hull, R. P. Schmonder, and others: Dietary intake and cardiovascular risk factors, United States, 1971-75, Part I, blood pressure. Vital and Health Statistics. Series 11-No. 226. DHHS Pub. No. (PHS) 82-1676. Public Health Service. Washington. U.S. Government Printing Office. In preparation. 9National Center for Health Statistics, W. R. Harlan, S. A. Eklund, J. R. Landis, and others: Dietary intake and cardiovascular risk factors, United States, Part II, serum urate, serum cholesterol, and correlates. Vital and Health Statistics. Series 11-No. 227. DHHS Pub. No. (PHS) 82-1677. Public Health Service. Washington. U.S. Government Printing Office. In preparation. 10National Center for Health Statistics, A. Engel, R. Murphy, K. Maurer, and E. Collins: Plan and operation of the HANES I Augmentation Survey of adults 25-74 years, United States, 1974— 75. Vital and Health Statistics. Series 1-No. 14. DHEW Pub. No. (PHS) 78-1314. Public Health Service. Washington. U.S. Gov- ernment Printing Office, June 1978. 11Goodman, R. and Kish, L.: Controlled selection—a technique in probability sampling. J. Am. Stat. Assoc. 45, 350-373. 1950. 12National Center for Health Statistics, E. E. Bryant, M. G. Kovar, and H. Miller: A study of the effect of renumeration upon response 22 in the Health and Nutrition Examination Survey, United States. Vital and Health Statistics. Series 2-No. 67. DHEW Pub. No. (HRA) 76-1341, Health Resources Administration. Washington. U.S. Government Printing Office, Oct. 1975. 13National Center for Health Statistics: The HANES Study, Final Report. Prepared by the Institute for Survey Research. HSM—-110- 73-376. Health Services and Mental Health Administration. Philadelphia. Temple University. 14U.S. Department of Health, Education, and Welfare: A Com- parison and Analysis of Examined and Unexamined Persons on Medical History Characteristics for the First Round of the Health and Nutrition Examination Study HSM-110-73-371. Prepared by Westat Inc. Rockville, MD. Jan. 24, 1974. 15National Center for Health Statistics: Factors related to response in a Health Examination Survey: United States, 1960-1962. Vital and Health Statistics. Series 2-No. 36. DHEW Pub. No. 36 (HSM) 73-1263. Health Services and Mental Health Adminis- tration. Washington. U.S. Government Printing Office, Aug. 1969. 16Wesley L. Schaible, Ph.D., Acting Chief, Methodological Re- search Branch, National Center for Health Statistics: Memorandum to Arthur J. McDowell, Director. Division of Health Examination Statistics. June 21, 1974. 17National Center for Health Statistics: Blood Pressure Levels of Persons 6-74 Years, United States, 1971-1974. Vital and Health Statistics, Series 11-No. 203. U.S. Department of Health, Educa- tion, and Welfare. DHEW Publication No. (PHS) 78-1648. 1977. 18National Center for Health Statistics, H. W. Miller: Plan and operation of the Health and Nutrition Examination Survey, United States, 1971-1973. Vital and Health Statistics. Series 1-No. 10b. DHEW Pub. No. (PHS) 73-1310. Public Health Service. Wash- ington. U.S. Government Printing Office. 1978. 19Lepkowski, J. M.: Design effects for multivariate categorical interactions, doctoral thesis, University of Michigan, 1980. 20Vinter, S. T.: Survey sampling errors with OSIRIS IV. Paper presented at the COMPSTAT conference, Aug. 1980. 21Gurney, M. and Jewett, R. S.: Constructing Orthogonal Repli- cations for Variance Estimation. Am Stat Assoc. 70:819-821. 1975. 22Kmenta, J.: Elements of Econometrics. Macmillan Publishing Co., Inc. New York. 1971. 23Holt, D., Smith, T. M. F. and Winter P. D.: Regression Analysis of Data from Complex Surveys. J R Stat Soc A, 143(4):1980. 474-487. 24Snedecor, G. W. and Cochran, W. G.: Statistical Methods. The Iowa State University Press, Ames, Iowa. Sixth Ed., 1967. 25Beaton, G. H., Milner, J., Corey, P., and others: Sources of variance in 24-hour dietary recall data, implications for nutrition study design and interpretation. Am J Clin Nutr 32(6), Dec. 1979. 26Grizzle, J. E., Starmer, C. F. and Koch, G. G.: Analysis of categorical data by linear models. Biometrics 25:489-503, 1969. 27Koch, G. G. and Lemeshow, S.: An application of multivariate analysis to complex sample survey data. JJ Am Stat Assoc 67:780- 782, 1972. 28Koch, G. G., Freeman, D. H., Jr. and Freeman, J. L.: Strategies in the multivariate analysis of data from complex surveys. Int Stat Rev 43:59-78, 1975. 29Freeman, D. H., Jr. et al.: Strategies in the multivariate analysis of data from complex surveys II, an application to the United States National Health Interview Survey. Int Stat Rev. 44:317-330, 1976. 30Freeman, D. H., Jr. and Brock, D. B.: The role of covariance matrix estimation in the analysis of complex sample survey data, in N. K. Namboudri, ed. Survey Sampling and Measurement. New York. Academic Press, 121-140, 1978. 31National Center for Health Statistics, S. Abraham, M. Carroll, C. Johnson, and C. Dresser: Caloric and selected nutrient values for persons 1-74 years of age, first Health and Nutrition Exam- ination Survey, United States, 1971-74. Vital and Health Statis- tics. Series 11-No. 209. DHEW Pub. No. (PHS) 79-1657. Public Health Service. Washington. U.S. Government Printing Office, June 1979. 32Koch, G. G. and Stokes, M. E.: Annotated computer applica- tions of weighted least squares methods for illustrative analyses of examples involving health survey data. Technical report prepared for the National Center for Health Statistics. 1980. 33Koch, G. G., Stokes, M. E. and Brock, D.: Applications of weighted least squares methods for fitting variational models to health survey data. Proceedings of the American Statistical Asso- ciation Section on Survey Research Methods, 218-223. 1980. 34McCarthy, P.J.: Pseudoreplication, Half samples. Rev Int Stat Inst 37, 239-264, 1969. 35Kish, L. and Frankel, M. R.: Balanced repeated replications for standard errors. J Am Stat Assoc, 65:1071-1094. 36Lepkowski, J. M. and Landis, J. R.: Strategies in the analysis of the dental data from the HES and HANES. Contributed papers to the Third Data Use Conference, November 14-16, 1978, Part II, 99-115. 1979. 37Lepkowski, J. M., Design Effects for Multivariate Categorical Interactions. Unpublished doctoral dissertation, University of Michigan, 1980. 38Landis, J. R. et al.: A computer program for the generalized chi-square analysis of categorical data using weighted least squares (GENCAT). Comput Programs Biomed, 6:196-231, 1976. 39Shuster, J. J. and Downing, D. J.: Two-way contingency tables for complex sampling schemes. Biometrika, 63:271-276, 1976. 40Bhapkar, V. P. and Koch, G. G.: Hypotheses of ‘no interaction’ in multidimensional contingency tables. Technometrics, 10:107- 123, 1968. 41Graybill, F. A.: Introduction to Matrices with Applications in Statistics. Belmont, California: Wadsworth Publishing Company, Inc., 1969. 42K ish, L. and Hess, I. On variances of ratios and their differences in multistage samples. J Am Stat Assoc, 54:416-446. 1959. 43Forthofer, R. N. and Koch, G. G. An analysis for compounded functions of categorical data. Biometrics, 29:143-157. 1973. 44Koch, G. G., Landis, J. R., Freeman, J. L., Freeman, D. H., Jr. and Lehnen, R. G. A general methodology for the analysis of ex- periments with repeated measurement of categorical data. Bio- metrics, 33:133-158. 1977. 45Freeman, D. H., Jr., The regression analysis of data from com- plex sample surveys: an empirical investigation of covariance matrix estimation. Unpublished doctoral dissertation, University of North Carolina, 1975. 23 List of detailed tables 10. 12. 13. 24 NHANES | population estimates for examination locations 1-65, by sex, race, and age at examination: United States, 1971-74. . .. Sampling rates by age-sex groups for general sample of the NHANES I: United States, 1971-74 ............ Ws = 5 20% 3 Vs Subsampling rates by age-sex groups for detailed sample of the NHANES I: United States, 1971-74 ............ccovvvnnnnn. Percent distribution of adjustment factors for the NHANES I: United States, 1971-74 National Health and Nutrition Examina- tion Survey, survey locations 1-65, United States, 1971-74... Total number of examinees and those without dental examination records, by sex and age: National Health and Nutrition Examina- HON SUNBY, 1971748 ivuicivivns is 1 5 0s mmmalv amine oa ve sani Number of primary sampling units (PSU's) and number of examined persons for the general, detailed, and Augmentation Survey by stratum number for the NHANES | design: United States, 1971— Number of examined persons by race, sex, and stratum number in the NHANES | design: United States, 1971-74.............. Number of examined persons by race, sex, and stratum number in the NHANES | design for the detailed sample: United States, VOT Th wv mvinmes ys wn 6 4 ism wm aie wa © 3 5 aT AE Sh #2 Number of survey locations, type of examination, years of data collection, age of target population, number of examined persons, and location of appropriate weights on public use tape for NHANES | dative s «2 x 2 1 sx pinsiomminesy ws 55 + wuismmeiein vo 5s Comparative analyses of standard errors and design effects for multiple and paired sampling error computational units (SECU's) within certainty strata for the number of decayed, missing and filled (DMF) teeth, systolic blood pressure (SBP), and calories by age for NHANES | data: United States, 1971-74. ............ Number of examined persons, estimated means, standard devia- tions, standard errors of the mean, and design effects for the num- ber of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP), calories, and age under analysis options 1-3 for NHANES | data: United States, 1971-74 ................... Number of examined persons, estimated means, standard devia- tions, standard errors of the mean, and design effects for the num- ber of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SPB), and calories consumed daily within age groups under analysis options 1-3 for NHANES | data: United States, AOTITH.....3 0 crencrrents HE A ATL A ror Number of examined persons in subclasses determined by lowest 15 percentile and highest percentile of skinfold thickness, means, standard errors, test statistics, and design effects for serum cho- lesterol: United States, 1971-74 ......... $3 BRS ET 3 26 26 26 26 27 27 28 29 29 30 30 31 17. 18. 19. 20. 21. 22, 23. 24. 25. Summary of simple regression models of the number of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP), and calories consumed daily on age under analysis options 1-3 by race and sex for NHANES | data: United States, 1971-74... .. Summary of multiple regression models for the number of decayed, missing, and filled (DMF) teeth on age, race, sex, and sweets for 6,349 examined persons ages 11-30 under analysis options 1-3: United States, 1971-74 ........ iin Summary of multiple regression models for systolic blood pres- sure (SBP) on age, race, sex, and Quetelet’'s Index for 13,573 examined persons ages 18-74 under analysis options 1-3: United States, 197174... convnvrvinis iris tanmmmanes sven Summary of mean periodontal index (Pl) score and estimated standard errors and design effects by drinking and smoking classi- fication for NHANES | detailed sample: United States, 1971— Hypothesis tests for variation in mean periodontal index (Pl) score by cross-classification of drinking and smoking variables for NHANES | detailed sample: United States, 1971-74 ......... Hypothesis tests for variation in mean periodontal index (Pl) score by cross-classification of drinking and smoking variables (model with no interaction) for NHANES | detailed sample: United States, NBT V=Th.. . rcetnrsrmamansassntunta,s sid #5 8 853 oon WneTRELERS) & Distribution of sample elements according to the r levels of the response profile by the $:8UDCIas8es «cc. vs: vs vmavemenniey Number of examined persons ages 25-74 with periodontal index (P1) scores of zero (none) and greater than zero (some) by race and current smoking status for NHANES | detailed sample: United $1108, N0TV =F inre v4 5 4 14 5 wwwmamomn wes v3 4 3 SwEBUGLL £3 4 3 Weighted number of examined persons ages 25-74 with perio- dontal index (Pl) scores of zero (none) and greater than zero (some) by race and current smoking status for NHANES | detailed sample: United States, 1971-74 ............ciiiiinnnnnn. Distribution of proportion of some periodontal index (Pl > 0.0) and estimated standard errors and design effects by race and cur- rent cigarette smoking classifications for NHANES | detailed sample: United States, 1971-74 ..........civiinnnnnnnnnn Vector of subclass proportions of some periodontal index (Pl > 0.0) and estimated covariance matrix by race and current ciga- rette smoking classification for NHANES | detailed sample: United SALES, VOT N78... u 1.1 SETHE th 2 DE 8 SEER AE mi% SE 10 row Hypotheses, hypothesis matrices, and test statistics for the model x, relating the variation in the proportion of some periodontal index (Pl > 0.0) to race and current cigarette smoking classifica- tion using sample weights and design effects for NHANES | de- tailed sample: United States, 1971-=74..................... 33 33 34 34 34 35 35 35 35 36 36 26. Hypothesis tests for variation in the proportion of some perio- dontal index (Pl > 0.0) by cross-classification of race and smok- ing cross-classification for NHANES | detailed sample: -United . SALES, TOTV=Th. +. os nsimrsizson 1, 050 wb iwsh nk Mia SaRg 5.0 3 1 rm we 27. Hypothesis tests for variation in the proportion of some perio- dontal index (Pl > 0.0) by cross-classification of race and smok- ing (reduced model) for NHANES | detailed sample: United States, 197174... oi 25 Table 1. NHANES | population estimates for examination locations 1-65, by sex, race, and age at examination: United States, 1971-74 Estimated population Age at examination Male Female Total All races White Black All races White Black Total = conn onamemes 53 +30 e sab mime dos sense 193,976,381 94,239,866 82,740,899 10,413,986 99,736,515 86,867,546 11,999,935 VYBAT 11. tohnmisrncsen ors ivpie bie womdoneaming $303 3,313,458 1,693,074 1,401,508 280,212 1,620,384 1,327,657 257,289 Z-BYOOIS vont si he EE. 10 6,963,162 3,653,765 2,997,107 479,362 3,409,397 2,872,581 505,442 el BBY: s «cv mmwacsrip ws vo BIRRERIREEOD SEE 6,672,346 3,378,503 2,866,374 485,872 3,293,843 2,755,016 511,134 B-7vears ..............ciiinniirnnranns 7,193,663 3,652,322 3,060,888 573,867 3,541,341 2,951,927 576,578 B-0YBaIB sansiminun 08 8 5 hms emia ends 7,696,597 3,880,396 3,279,649 586,419 3,816,201 3,257,936 539,855 VO0=11 YoarS suiting «3s » snumwwmsicnowese ss 8,465,793 4,381,730 3,732,593 563,823 4,084,063 3,424,070 617,793 2 RT 12,335,321 6,312,519 5,397,061 879,377 6,022,802 5,122,189 836,252 VBA 7 YEAS vivnigns v7 5 4 4 bbw mama & bps 12,318,434 6,207,169 5,311,596 812,321 6,111,265 5,233,091 853,294 18B=1Q years «a vinny tus 5.5 45 63 wnkivinms we vv 7,352,200 3,673,321 3,206,467 404,045 3,678,879 3,158,930 504,417 20-24 Years... oii 17,325,038 8,109,775 7,094,036 866,201 9,215,263 7,972,486 1,073,358 25-3 Y@AIS . iii vv rss nna 26,936,001 13,002,514 11,594,115 1,231,793 13,933,487 12,160,578 1,646,337 BEA YOBIS + «uv vnsv suv 333 8a simran 22,268,477 10,675,731 9,515,530 1,004,953 11,692,746 10,111,458 1,318,050 ABBA YEAS. s wiwisiniviin v5 vv # 3 5 » a niwigieibie y 4 + 23,313,316 11,150,110 10,039,124 1,056,837 12,163,206 10,879,167 1,237,459 B5-64vyears..............ciiinniinnnn 19,049,001 9,072,586 8,274,948 702,647 9,976,415 9,037,157 871,098 BS =T74 YEAS . «ov vsvusive ins vss vmaimummess a 12,773,574 5,496,351 4,969,903 486,257 7.277.223 6,603,303 651,579 Table 2. Sampling rates by age-sex groups for general sample of the NHANES I: United States, 1971-74 Age and sex Sampling rate PB YOAIS . vovvssrvinih s vorsmamncnenes 8 2 4 ASTRA ECE 4 #8 STE C ECE #1 ARERR TATE EF & Hoh mm 4% BRERA 1/2 BVO YOATE (rie nits 4 3 BATA EERE SF bs IREETTRA es i0.0 0 ETRE THIRTA IRA Sr F8 AEA ZS RTA iA mc tin 8218 vA Erb nite ov 0, EERE RR RE 1/4 20-44 -Y8arS {MBN «+ «+ » vr taimemnsiene 1% 3a srs EBAREY © § SEISHRNST ESOL TE SEEIRR REET ¥ 8 SEE BARE $F ARR RR EE 1/4 20-44 VRAIS IWOMIBNY . . . coc unnvncnns ss rsnsnsnass snes es dah sn essen esssss esses sass ssn ns asses esses es sameness 1/2 BBB YORI rn 2.05.55 54 4 OR ER AA 0 ETRE AA 09: 59 Saison AA 0 0 Aen A A siatinr 8 let Rt me AW BREAN 4 3 1/4 BB=T4 YOBIS vr um. +v5000 0% S=4nas enna Hin S100 0 WTB A WEE Ar ESAS AE HE BE BU AR BREE RF SCAT SSE SE Ss S590 1h Re aaa ed 1 Table 3. Subsampling rates by age-sex groups for detailed sample of the NHANES I: United States, 1971-74 Age and sex Subsampling rate DSA YEAS KINO). . coirismme 0500 wren ieasimimalen sons 85 ba Tid die WTassnn & #704 i ERALTAED #9 5 mei ents wi 1m wri olesw ma sw wees: ms w) sis OR 2 2544 YEAIS AWOTIBNY: BO0-3.08: + svn vn i 59 58 3 SABER Er Ee 5 5 3 $00 SRmD RSE 4S IWDERRTAA § BE BARGER § Ee 1 0. 26 Table 5. Total number of examinees and those without dental examination records, by sex and age: National Health and Nutrition Examination Survey, 1971-74 Age Soin Male Female Bath Male Female sexes sexes Number without dental Total number examined fies examination records AlLAGES, TTR YBBIB. ...... vivo siminmswisims sy warm wiaanes vim smi aden vies vies 20,749 8,820 11,929 531 207 324 VB VBS vem virim vis v8 TIF FF ERAN TET 2 SADA G CERT HS FREE 2,895 1,469 1,425 78 36 42 B11 YAMS . otitis 2,057 1,026 1,031 63 30 33 F217 VRAIS. «cette teeta eee 2,126 1,064 1,062 48 18 30 VB=24 YBOI8:: v5 000000 5 8 5 v AIRE STE TERRA AE AGAR FE RRR 2,296 773 1,623 60 14 46 2B=Bh YBBIS + vvvv vrs vps rR ER EEE EE SEER EE EET 2,700 804 1,896 80 26 54 B5—44 YBAIS . Lo tiie 2,328 664 1,664 55 14 41 BBB YOOIS +i av vi vv is 3 EERE SERRE Sy EERE RATER YEE SERRA 1,601 765 836 43 22 21 BBB YAS. vvis wana vy MESES TEE SERRE TORE 1,267 598 669 33 10 23 BETA YBAIS . . cvs vrcrn snes vrais narrates earn er 3.479 1,657 1,822 71 37 34 Table 6. Number of primary sampling units (PSU's) and number of examined persons for the general, detailed, and augmentation survey by stratum number for the NHANES | design: United States, 1971-75 Number of PSU's in ; Number of examined persons sample survey design Stratum number General and Aiomanteiion General and Detailed Avcmenistion detailed g detailed only 9 BE si cor feinasninion Sunsmssmemmistinmesnissinie sess iBi mz ans ns sina ie iaaond 1,263 236 20,749 3,854 3,059 1 1.213 211: 4,511 853 701 rr rr rE ER ATA Sn 169 21 621 112 55 sn daira ennai AM ARATE int AE RA ROR MYERS Aan FE Rt 106 17 367 80 63 TT TTL 125 18 482 87 59 Br ee RT ERE ROTATE AER 156 21 737 129 60 erotics hi mon ron asmesama on Hori ano ama URS FR 197 24 741 143 97 I a rr aaa. 83 22 250 48 82 T rrr cr PEER SEER REEL PR STE SER 108 23 395 71 72 artnet A ar SAA AA 8 4 AAA RR ROR NR A 3 61 21 188 42 80 8 REE EL RIE AEE I ERAS TEE 89 21 304 57 64 TO sir von conn vv nivmpiin sys ennne serine Eee aE 119 23 429 84 69 i Tn 50 25 16,235 3,001 2,358 NOTE: In the certainty strata 1-10, PSU's are enumeration districts. In the noncertainty strata 11-35, PSU's are counties or groups of contiguous counties. 27 Table 7. Number of examined persons by race, sex, and stratum number in the NHANES | design: United States, 1971-74 Number of examined persons ages 1-74 years by race and sex Stratum number Total White Black Other White Black Other males males males females females females TOUa) svcininnivs 20 34 samimrarne ds wk » 5 #8 Balan 1.5 5 29 20,749 7,004 1,707 109 9,347 2,456 126 oz atmrritmrs £1 + 45 SE NERIRERIEES § B13 § # eRBE ERT 1 621 169 88 2 220 138 4 I EE Rg pr PP 367 146 24 0 157 38 4 B cisions 7% 1 ® nr SRIARE EE 1783 FREER AE YE 7 482 123 85 1 171 102 0 os ETEERTTE £2 LE 8 3 RETR RE § £8 8 Rane ££ 7 vv 8 737 198 102 1 255 162 9 Bir aiecinicinm 4 5¥ £0 5 8 0 iirbtiansin 4 #8 § + 5 3 hb Bardi 6 4 wi 741 232 65 13 328 88 15 Bas wiaiminiainin 4 % £00 8 ARASH 408 PRE MERITIRGE Ha TEE 250 67 35 2 85 57 4 Foams he % TEER FESR FRA TAER EEF RE 4] 395 85 90 0 93 127 0 Brest € 2% 2% 5 I RAR RA SR ARES REA 188 67 16 0 79 26 0 Osman #4 5 53 & BEM § 45 8 HEE BE 8 S12 304 109 13 1 149 32 0 VO owner sonra anne saves tts ArT RTRs red 429 138 32 13 190 37 19 VN orion s 2 55 3 5 33 SERIRAAIGUR £ ¥ © © 5 #5 SARS #5 7 3 481 205 4 0 267 3 2 WZ santos v + 1.6% FWRRRIIERIE #4 6 § #6 SEARRCRIAR € £5 50 517 198 14 0 286 17 2 1B nirnnnin ss ress arvana das RE ATONE EE RA ERY 531 232 2 2 290 4 1 VY cron fr 3 5 8 5 § F SIRERERERET E 11k 8% BOER £5 08 701 273 15 2 396 14 1 {© repre es RE I 486 185 20 4 226 43 8 BL, ec iwsia ss 1 4 #5 HARTER ¥ 8 8 18 1 SA PRER EE 4 RAR 563 178 68 5 211 98 3 VE mttclerird os 5.3 5 SERRE ha # BATE 3 SAUTE ER § 2 TE 4b 594 235 6 0 346 6 1 I er J gE Pp J 505 176 39 2 224 62 2 1D issinieniny 21 2 2S BAPGHAAG A HH 48 F AF IGAEIE CCE EY ETRE 585 237 12 4 317 14 1 20. us vinsin o 4 + 3 SEAETRETE 48 88 § ATREGRRE TE EY EY 446 171 13 1 246 14 1 BA Gum c rans b + tis SATE RE © R&A #4 790 344 0 0 446 0 0 D2 niin a SAREE € 4 8 BAECS EAE 551 114 107 3 141 185 1 i ae 619 167 85 0 249 116 2 Bh viv viva bf SEER RR & 8 SRST * £4 x 8 a 499 131 73 0 170 122 3 2B rnin 1.4% Sa 8 A RE EB § A 728 225 73 0 311 119 0 BB. oviinies tan Ta ETE tea E PERE SEE EE EES 887 232 156 0 305 194 0 istic hit 8 ATER ATEA 6 0 08 SARA sot 30 6 8 8 540 684 262 23 1 379 17 2 BT II Tran. pee, 1,001 259 174 0 327 241 0 BO. ivsnnins is sr sande ey tad van I VEY ERERGCE EES VE 20 634 222 51 1 292 68 0 BO rs 50 5.00 55 ARTE ms a4 4% SETA RAST AE 0 Amana 868 284 84 1 271 124 4 BY uinsnnees sa brn tnEE se 8 FORRES EERE ER 651 21 34 5 334 52 5 BR iri 2s mmm pee @ 0 F FINA VIS eS eRe 691 250 22 8 367 32 12 BB usr niin 03 Trem AR AREA BE 619 222 3 21 345 10 18 BA. iiinniicstrirnlussnney sve ses SHRPREEGEE EY SLRS 545 236 5 5 295 1 3 8B. inns SATERART CTR § 8A Shr oA Ee 1,059 411 74 1 479 93 1 28 Table 8. Number of examined persons by race, sex, and stratum number in the NHANES | design for the detailed sample: United States, 1971-74 Number of examined persons ages 25-74 years by race and sex Stratum number Total White Black Other White Black Other males males males females females females FOUR vv 005 3 Simmer 8 4 4 4 19% 2 HESSEN HELE 2 § 4 B00 3,854 1,541 277 21 1,667 335 13 mh rnniiit v7 78 3 SPEEA Aint £8 # Sir aah Sr FF 3 iA 112 37 13 1 34 27 0 Drown une in RGRAY # 4 $5.8 8 SRE SY £718 RAR 80 38 4 0 27 1 0 Btn ns 2 e 6 3 SERER CL FE § VFR EE FE xs an 87 23 18 0 29 17 0 Qos neni oo TAT sh es] B REO RA SAE vd tbat 1289 46 15 1 43 23 1 Bin Bis £5 5 4 ie IER a 4 4 8 FEARS AEE EE 58 28 PERE 143 60 11 4 55 12 1 B iintc vs id AEE A 4 SERRE EEA 48 17 7 1 12 1 0 282000554 1% FETTER RRITRAS ft £4,160 5 mpm mr AA 41 Ft ERE n 16 18 0 17 20 0 Biv ne er ARRE TEES 4 & TEER AREER TEES SREY 42 19 0 0 18 5 0 vera RES OB TRAE A SE eR 57 25 1 0 27 4 0 00.0: GIRS) ws Too 7 To WR ns Wom 45 Ai td SERSAER Em 84 34 8 4 30 Bb 3 VN snrmnincns oessompopie Biman 400 6 ARS R bande HR HR 0 EEE 100 45 0 0 53 1 1 12, conv sss ns sins distaste evs Ens ue reve ETE era 23 40 3 0 49 0 1 Bs 50.554 BETIS GFE 8 4 TR RAs Tara fra ra ma Ls Amps i 92 45 1 0 46 0 0 LE 129 54 1 0 70 4 0 ; NO 78 43 2 1 27 5 0 VB is + eran TH oir 44 RT TRA. rin a hs oom 101 29 13 0 41 18 0 er TERE pe, 107 52 1 0 54 0 0 Bu sini ummenrines 8 aan 00 SHENAE $A RA 4 81 41 4 1 28 7 0 Ei a ATS AHH 0 MIE 31 SA A 109 45 2 1 59 2 0 Ys ram GRR #13 6 8 RP Bb Esta 8 81 34 2 0 44 1 0 1 i Ca Sp 162 12 0 0 90 0 0 DD a RR 0 WR HRT A RATS TA 89 28 17 1 23 20 0 23 connie 8 p25 sna EERE 1 8 REAR EAE 3 112 33 16 0 48 15 0 a Sern 81 28 8 0 30 15 0 PB Wsmmiome: + i 44 35 HB HETERIE 4 3 £3 FP IEATRATT 40 48 156 67 8 0 67 14 0 BB immer n 55 5 Sn RTE © nh 2 4 @ ER FE 150 45 22 0 65 18 0 DF ccivimions 65 0 43 ERP EBGAS 5 1 E43 FDOT 4 08 5 141 65 6 0 68 1 1 2 ret esd 47S RIE ERTEEEY TRF UR REPT to na 182 57 26 0 64 35 0 2 i rnin: #5 4 nt iT bn 8 4 8 SEA © 5 7 nA 126 50 10 0 58 8 0 Boivin criss ramanuiend nis bes FE RCRERIGLE EE ETE RED 152 63 14 0 64 i 0 BE reeions ts 03 55 BERGER 1 HOY 4 18 38 VEE $0 Fe 113 49 3 1 51 8 1 BD imma & B55 4 4 SEAS Ra AE BARS © Sn 4% 123 51 2 2 61 6 1 BR innoiiinisisangamntn se ines FHSS EEE TEL $7 89 19 45 0 2 69 0 3 BY ivi 055 25.58 PRRSEREET ET E08 FRR £4 9 4 FF en 100 46 2 0 52 0 0 BB rns cing 158 FITAAERARS £88 P88 REET ay 8 Wl 224 99 19 1 94 1 0 Table 9. Number of survey locations, type of examination, years of data collection, age of target population, number of examined persons, and location of appropriate weights on public use tapes for NHANES | data FI . Age in years of Number of . Tape locations Survey locations and examination in sample design Year : target population examined persons of weights 1 LT SR PR pe perl Sia 1971-74 25-74 3,854 170-175 TBE NUIION. vc v50 5 4s sansgdnm sees s 28s ERFISH OLE HF 3 CPED 1971-74 1-74 20,749 176-181 BO=T00V Rtll..v vv cv wives mermininemecnin os = #erncemomasmsosinis Ane woreantiasn 1974-75 25-74 3,059 182-187 1=1002 detail. . o.oo eee eee eee 1971-75 25-74 6,913 188-193 1 Augmentation sample 2|ncludes augmentation sample 29 Table 10. Comparative analyses of standard errors and design effects for multiple and paired sampling error computational units (SECU’s) within certainty strata for the number of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP), and calories consumed daily by age for NHANES | data: United States, 1971-74 Multiple SECU'’s Paired SECU’s Ade Number of Maar 9g examined persons Standard error Square root of Standard error ~~ Square root of of mean design effect of mean design effect DMF teeth VPA BAPE viv 495 1 4 3 0 ncteriiaTond BHR 74% % 5 miss 20,749 14.723 0.166 2.094 0.161 2.034 V=17 YORI ovine + #4 4% yA RDFARHETE E06 8 5 5 3 3 BRTHEE 7.104 3.965 0.071 1.545 0.070 1.5638 V8=20. YRAIS siviaivivs 1.0 + 2 3 9 wiwevinwmses oe v.08 3» % www 2,297 11.924 0.237 1.766 0.237 1.768 25-34 VRAIS... . tits 2,694 16.918 0.261 1.823 0.262 1.826 Te 2,327 21.436 0.249 1.560 0.248 1.655 LL TE 1,599 22.826 0.216 1.085 0.232 1.164 B5—B4 Years... uit 1,262 25.744 0.291 1.278 0.279 1.224 BE=T0 YEAS. ini s vi 7 5 +3 4 GARGIRRAWE RB BEE #3 540m 3,466 27.727 0.154 1.283 0.154 1.278 SBP B78 YEAS: iia sv 5 3 5 3 3 5a Beniunie ton 4 8 4 06 ¥ b&b R2 17,658 123.95 0.424 2.292 0.409 2.21 OTT YOBIS .oivinivii vv 6.4 35 8 RRRERERNE & 4 8 F 8b RY BROES 4,085 108.24 0.492 2.207 0.498 2.234 BBA YOAB ov vv ss v5 43 yravwwmess ous see nms 2,290 118.89 0.466 1.573 0.441 1.489 25-34 years. .........c.onuunnnn TR ST 2,675 120.93 0.445 1.5634 0.440 1.515 BB=A4 YBAIS ci visss sai srRRanr es ey hea vE nS 2,317 125.64 0.580 1.479 0.603 1.536 AEB YOBIB ais vi v5v 4 3 wis mannnmess v5 4% 3s EaT VET 1,589 134.14 1.015 1.746 1.037 1.783 B5-B4 years.......... cc iiciiiiii iii nnn 1,255 142.11 0.826 1.214 0.804 1.181. BO =Th YOBIS vs ci v3 5 40a swmininis ns aay & 53 pnnias 3,447 150.01 0.793 1.820 0.784 1.799 Calories VT BAIR iramonss 2.3 + + 5.390 AFTRA #53 23 3 Fase 20,749 2000.0 17.80 2.923 17.88 2.937 V1 7 YOBIS cuales 15 05 4 BE GRBREE EE 0 F828 SRR 7.104 2011.0 20.75 2.106 20.03 2.033 YBm28 YOAS o's v4 + + 5 su mwmnimnirns sip ss § wiwim ummm 2,297 2294.8 37.02 1.660 35.32 1.584 26-34 VRAIS... outa 2,694 2177.5 27.66 1.479 29.44 1.573 B88 VOUS. vo v0 v0 + 7 EO BORSA LE A #85 Re 2,327 2042.9 28.33 1.545 28.94 1.578 ABBR YB... 5s. Ss + + wusmingesdiusnigpes o wa win 48 monies 1,599 1897.3 31.76 1.615 30.41 1.451 rE IE 1,262 17232 33.06 1.418 33.45 1.435 BB=T74 YOBIS, i 5+ 354s sno wmpesnme sv 1s 8s Ra mERHy 3,466 1518.9 20.68 1.870 19.99 1.808 Table 11. Number of examined persons, estimated means, standard deviations, standard errors of the mean, and design effects for the number of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP), calories consumed daily, and age under analysis options 1-3 for NHANES | data: United States, 1971-74 y Inclusion of sampling Sample Standard error Square root of Option number size Mean of mean design effect Weights Design 9 DMF teeth 0) ime errs ae me preemie ane lr Ewan ce eaten : No No 20,749 14.93 0.079 t. PR Yes No 20,749 14.72 0.075 1... SN NS Yes Yes 20,749 14.72 0.161 2.156 SBP 1 er TE EE TE No No 17,658 126.91 0.185 1 D1, oar wom wrens es A AR ER RRR Yes No 17,658 123.95 0.168 Xo Bev Tht Rh RE A TR RR AL SFA SE Yes Yes 17,658 123.95 0.409 2.442 Calories Be Bow ef EFT el Benn TE A RBA No No 20,749 1827.5 6.088 1. AON Yes No 20,749 2000.0 6.560 Yds JR Yes Yes 20,749 2000.0 17.883 2.726 Age Wosirerniovemerimsmiiosim Sie arr BE STE No No 20,749 32.23 0.159 1 cen ecrin Hamapasehg R wgse ie een znpea ns mt ve ps mi wan beni Yes No 20,749 30.61 0.140 T GT RTE USE SE EE RE ER Yes Yes 20,749 30.61 0.239 1.707 1category not applicable. 30 Table 12. Number of examined persons, estimated means, standard deviations, standard errors of the mean, and design effects for the number of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP), and calories consumed daily within age groups under analysis options 1-3 for NHANES | data: United States, 1971-74 Option 1 Option 2 Option 3 Number of Age examined Standard Standard Standard Standard Standard ~~ Square root persons Mean ia gz error of Mean reap error of Mean error of of design deviation deviation mean mean mean effect DMF teeth 1-74 years................. 20,749 14.935 11.4180 0.0793 14.723 10.7760 0.0748 14.723 ‘0.1613 2.156 V=37 YORB.iviovis v5 4 54 vinninin 7.104 3.338 3.8493 0.0457 3.965 4.0810 0.0484 3.965 0.0703 1.452 18-24 years... c.. cu 0uunn 2,297 12.050 6.4173 0.1339 11.924 6.2566 0.1305 11.924 0.2367 1.813 25-34 years. ............... 2,694 16.872 7.4408 0.1434 16.918 7.2497 0.1397 16.918 0.2618 1.874 36544 vyears................ 2,327 2127 7.6962 0.1595 21.436 7.3482 0.1523 21.436 0.2481 1.629 45-54 years. ............... 1,599 22.515 7.9700 0.1993 22.826 7.5709 0.1893 22.826 0.2320 1.226 BBB Years. ....co vii eunnnn 1,262 25.234 8.0990 0.2280 25.744 7.6022 0.2140 25.744 0.2790 1.304 65-74 years. ............... 3,466 27.608 7.0741 0.1202 27.727 6.7742 0.1151 27.727 0.1536 1.334 SBP 6-74 years. ................ 17,658 126.91 24.585 0.1850 123.95 22.262 0.1675 123.95 0.4090 2.442 O-V7 YBArS. viva vas naniive 4,085 108.67 14.245 0.2229 108.24 14.132 0.2211 108.24 0.4980 2.252 18-24 y0ar18..c00vs viv venus 2,290 117.96 14.166 0.2960 118.89 13.794 0.2883 118.89 0.4407 1.5629 25-34 years. ............... 2,675 119.90 15.006 0.2901 120.93 14.710 0.2844 120.93 0.4397 1.546 35-44 vyears................ 2,317 125.76 18.885 0.3923 125.64 17.665 0.3670 125.64 0.6026 1.642 . 45-54 years................ 1,589 135.10 23.176 0.5814 134.14 22.782 0.5715 134.14 1.0365 1.814 656-64 years. ............... 1,255 143.13 24.126 0.6810 142.11 23.453 0.6620 142.11 0.8040 1.215 65-74 years. ............... 3.447 161.02 25.580 0.4357 1560.01 25.056 0.4268 150.01 0.7840 1.836 Calories 1-74 years. ................ 20,749 1827.5 877.00 6.088 2000.0 944.91 6.560 2000.0 17.883 2.726 1-7 years. ..ovivnnsminamas 7.104 1880.4 830.42 9.853 2011.0 874.24 10.372 2011.0 20.033 1.931 18=28 YOAIB. «sivas as vas enue 2,297 2084.6 1068.70 22.298 2294.8 1136.6Q 23.715 2294.8 356.317 1.489 25-34 years. ............... 2,694 1954.5 971.00 18.708 2177.5 1050.10 20.232 2177.5 29.435 1.455 35-44 vyears................ 2,327 1829.0 884.65 18.339 2042.9 966.51: 20.036 2042.9 28.935 1.444 45-54 years. ............... 1,599 1840.4 838.33 20.965 1897.3 816.17 20.411 1897.3 30.410 1.490 65-64 years. ............... 1,262 1679.2 828.08 23.310 1723.2 814.02 22.914 1723.2 33.454 1.460 65-74 years................ 3,466 1497.2 651.06 11.059 1518.9 649.50 11.032 1518.9 19.991 1.812 31 Table 13. Number of examined persons in subclasses determined by lowest 15 percentile and highest percentile of skinfold thickness, means, standard errors, test statistics, and design effects for serum cholesterol: United States, 1971-74 Low skinfold High skinfold Option t Square root of p . Number of Standard Number of Standard statistic design effect ; Mean a Mean examinees error examinees error All males Yvan menmnireiens bey 1,030 198.7 1.41 1,015 223.1 1.55 11.6 x. 2 103} ERR a 1,030 191.6 1.34 1,015 221.8 1.60 14.3 1... Bios sinaanaeessn assay 1,030 191.6 1.59 1,015 221.8 2.39 9.5 1.5 Black males Vs ois smn hae 5 5 6 282 200.1 2.73 1565 222.2 4.03 4.7 oa De swine 1 54 282 193.4 2.81 155 226.6 4.89 6.4 re Bs snanen sins veka 282 193.4 4.00 155 226.6 8.80 34 2.1 White males a 748 198.2 1.65 860 223.3 1.68 10.6 Ys RD 5 STEER E A § 748 191.2 1.54 860° 221.2 1.70 12.8 x... Boi causamanoge teres ve 748 191.2 1.90 860 221.2 2.26 9.7 1.3 All females Vos ooumnmmwnsss 385s 1,852 197.4 1.19 1,637 224.7 1.24 15.9 x. es enna 8 Sree 1,652 196.1 1.19 1,637 225.9 1.25 17.3 1A Bana 5h 8 3 bk 1,652 196.1 2.00 1,637 225.9 83 134 1.3 Black females os sivnmimaiendie s 4% & are 288 191.8 2.40 488 221.0 2.25 8.5 1. 2D RTE RE # SA WE 288 193.0 2.39 488 224.2 2.11 9.6 1... Be eri alentas £3 3 Aaneih 288 193.0 "3.00 488 224.2 3.33 6.6 1.4 White females F's sntommenining 4 + v8 spain 1,364 198.5 1.34 1.149 226.2 1.49 13.8 1. Binsin enti 6p As §anee 1,364 196.5 2.93 1,149 226.3 1.62 14.8 Clr Bien ty Sag Sale 1,364 196.5 2.21 1,149 226.3 2.21 11.8 1.3 1category not applicable. 32 Table 14. Summary of simple regression models of the number of decayed, missing, and filled (DMF) teeth, systolic blood pressure (SBP), and calories consumed daily on age under analysis options 1-3 by race and sex for NHANES | data: United States, 1971-74 Weighted design NOTbErol Unweighted design (option 1) Race and sex nha (Option 2) (Option 3) Square root 2 . persons 2 Standard t- B Siepe Standard t- Standard t- o7 425i9n R Slope ea) ART i effect error statistic error statistic error statistic DMF on age Total ........... 20,749 0.67 0.408 0.0020 206.89 0.65 0.432 0.0022 196.52 0.0032 135.09 1.45 White males. .... 7,004 0.73 0.416 0.0030 138.91 0.67 0.440 0.0037 118.93 0.0042 105.49 1.13 Black males ..... 1,707 0.63 0.335 0.0062 54.44 0.47 0.308 0.0080 38.52 LI Ye Ys Other males ..... 109 0.53 0.317 0.0287 11.04 0.45 0.294 0.0316 9.28 Yen Vie Votes White females . .. 9,347 0.67 0.414 0.0030 136.49 0.68 0.439 0.0031 139.50 0.0053 82.76 1.69 Black females... . 2,456 0.59 0.391 0.0065 59.91 0.54 0.385 0.0072 53.29 1 on Voss 1 Other females. . . . 126 0.40 0.337 0.0372 9.07 0.25 0.244 0.0376 6.50 Yon Yay 1, SBP on age Totals viciie sve vs 17,658 0.40 0.730 0.0068 107.45 0.35 0.696 0.0071 98.11 0.0131 53.14 1.85 White males. . ... 5,854 0.36 0.605 0.0106 57.24 0.33 0.610 0.0115 53.14 0.0113 54.06 0.98 Black males ..... 1,326 0.46 0.815 0.0240 33.91 0.43 0.848 0.0269 31.53 Yo Ye Vid Other males ..... 89 0.35 0.762 0.1118 6.81 0.14 0.401 0.1064 3.77 1... Xr... Ye White females .. . 8,243 0.41 0.767 0.0102 75.57 0.38 0.734 0.0104 70.39 0.0188 39.03 1.80 Black females. . .. 2,037 0.47 0979 0.0230 42555 0.44 1.008 0.0252 40.05 Too 1... 1 Other females. . . . 109 0.40 0.920 0.1086 8.47 0.37 0818 0.1040 7.87 1 1... 1, Calories on age Total ........... 20,749 0.02 —4.90 0.2629 —-18.64 0.01 —5.50 0.3238 —16.99 317 —17.35 0.98 White males. .... 7,004 0.01 —3.39 0.4873 -6.95 0.00 —352 0.6102 —-5.78 .6314 —5.58 1.04 Black males ..... 1,707 0.01 —3.74 0.9217 —4.05 0.00 —1.08 1.292 —0.89 Y. 5 Tout Yo Other males ..... 109 0.00 1.00 3.598 0.28 0.05 12.50 5.101 2.45 1... he 1... White females . . . 9,347 0.04 —5.89 0.3034 —-19.41 0.04 —6.44 0.3315 -19.43 4339 —14.85 1.3 Black females... . 2,456 0.06 —8.39 0.6578 -12.75 0.06 =—9.45 0.7420 -12.74 Ye LS ob Other females. . . . 126 0.00 —1.23 3.474 -0.35 0.01 -3.35 3.899 —0.86 Yo a 1, 1category not applicable. Table 15. Summary of multiple regression models for the number of decayed, missing, and filled (DMF) teeth on age, race, sex, and sweets for 6,349 examined persons ages 11-30 under analysis options 1-3: United States, 1971-74 . Regression Standard error ; ls Square root of Veriable coefficient of coefficient Vsitisye design effect Unweighted SRS design (option 1) AOR retri aavrmsTTmn bie 4 § 1 4.8 SESRRESRALS § § 5 SGEENR aa t § 04 1 RRB w eet 2445 500 0.685 0.0130 52.42 1. EP CT eR ie Bp 0.875 0.0899 9.73 1 BOK 1 & + 5 arersiinmine o'% 5 & Biba RHE ES § 5 8 wR © & £58 3 AREER AE ¥ —0.491 0.0752 —6.52 1. BWEBIS + ¢ ov visiominnie so 5 2 5S ABEBUEBEE #3 3 3 SPFWHLCEED © 8 4 BARES HEEEE 155 5 25 0.057 0.0070 8.21 L. Weighted SRS design (option 2) BNIB + 0 0.0 vio wm 08 6 0 A 8 8 Rn AEE § 4 SAREE 8 8 6 a 0.705 0.0125 56.29 1 A Te es | i re 0.795 0.1072 7.42 ; SON avis ries R 200 BRET 5 3.8 8 5 ERIE 3 G0 § SANTA Kei —0.465 0.0698 -6.65 * 1. SWEETS + «+ 40 BT va PSB 00 ED SO ET 0% 3 08 DR Ra 049 Be £0 0.049 0.0068 7.17 1. Weighted complex sampling design (option 3) BOO... osetia as 8 4 pa 3 ARSE § Es eae Ee $i 0.705 0.0209 33.67 1.67 BEC. « : » (Smsnmme ood b 1 © IRRSTER GIES § 58 SRE Tv BF ERIE AE 0.795 0.2277 3.50 2.92 EE 7 vs & Ti TE © FEAR EE Row hd ARR 88 Fei abot lf oe —0.465 0.0928 -5.01 1.33 rE TI 0.049 0.0077 6.43 1.12 category not applicable. 33 Table 16. Summary of multiple regression models for systolic blood pressure (SBP) on age, race, sex, and Quetelet's Index for 13,573 examined persons ages 18-74 under analysis options 1-3: United States, 1971-74 . Regression Standard error or Square root of Verisbie coefficient of coefficient esatistic design effect Unweighted SRS design (option 1) AGE}, +575 2 Tr 3 7 5 rE Te is bapa eros ws th eine w + Biman 0.667 0.0096 69.44 1. RICE: 2c 10 whiners 7 24% 4 SRR TEES » 5a SURE + § 6 wR oil Fh 5 3.896 0.3938 9.89 Vo BOK ile « Tie Romie » 136 5 1 5 bnndpolhitin s 4 § 5.5 EHTWREES 2 T05 weal p BEANS RAT —1.885 0.3495 -5.39 1, CUBES NIB fries 2 5 75 3 5.9 5a adarss PRLS + 3 ETE 7 0758 sm sworn wad & worn 1.135 0.0335 33.88 L. Weighted SRS design (option 2) AGE, 210-5 25 2 SAPinS 5.3 5 § STRIATE 57 A 2 3 5% CORP 27% © 13 5 AAS 4 BE TES Rad 0.584 0.0102 57.49 Le BOGE 75 145» whinismniin ns 5 5 ohms © #5 Fa §o Surin ss 5 § 5 Abs 45 55 2553 5 2.908 0.4422 6.58 1. BRE 07» 3 vaio s lain 5 £3 13 5 asamoNIs # HE £3 A Cuma £2 5 3 SUSAR FRE 8A 55 —2.871 0.3162 —9.08 1. Quetelat’S INABX vce rv 1255s smaiamns E42 F 28% SREETRER IE £28 8 SATEEN £ LER AE BA 1177 0.0331 35.56 1. Weighted complex sampling design (option 3) BOR: i svansenrbisleattis v «vn a Manitidmiensd + § 0h 3 5 % CEERI © 5 £ © + SADE § ET 5 BER TE 0.584 0.0177 32.92 1.78 BECE : 2 vis stolons § #2 bo Banta § 5s § rar iabensn 54 4 5 o ie ssiaminsin hoo 3.5 on & pcan 2.908 0.8266 3.562 1.87 BOK suit 0 rR £ F 5 BERRIES 7 7 $ § § SRT BE 8 § Smee £5 —2.871 0.5206 —5.562 1.64 QUBLOIBES INABX =... «ov « + rminaitleiaiwainss 557 8 LFSEHE SSE 70 ® § SRIRBEVEE 3 9% 24% 1.177 0.0630 18.69 1.90 1 category not applicable. Table 17. Summary of mean periodontal index (Pl) score and estimated standard errors and design effects by drinking and smoking classification for NHANES | detailed sample: United States, 1971-74 . Unweighted design Weighted design Subclass SRS (option 1) Weighted SRS (option 2) Complex (option 3) number -_— ; Number Mean PI Standard i Mean PI Standard Standard Square root of Drinking Smoking examined examined score error score error error design effect None INBVBY oc 5a v5.0.0 atieaiaion winnie wei 417 1.618 0.1074 354.73 1.424 0.1018 0.1371 1.35 None Past tu ovr nr s snus nan 101 2.038 0.2366 74.74 1.836 0.2171 0.3044 1.40 None NOW. oo to dee Bormnine 4 Ee £4 mE 195 2.349 0.1733 162.63 1.904 0.1578 0.1339 0.85 Little INBVBE «inven wsisiesms vas ain Sormaie 479 0.961 0.0733 547.18 0.800 0.0663 0.0583 0.88 Little Past... urn vsines penirsse sn 214 1.280 0.1282 251.74 0.966 0.1137 0.1325 1.17 Little OW oi5 04 24s ivminiernntons v's vasa 9pm 483 1.738 0.0968 571.00 1.516 0.0896 0.1267 1.41 Moderate INGVBE ovis smnlinins vn op nnieinivminte 178 1.003 0.1303 196.80 0.853 0.1271 0.1887 1.48 Moderate PASE oil nivienimn site vis sre laine 166 1.148 0.1341 198.95 0.930 0.1195 0.1120 0.94 Moderate INOW 5 45.5.5 4 ieiatabelaa ies 4 £150. 4 Farmed 483 1.731 0.0984 540.25 1.463 0.0902 0.1256 1.39 Heavy NOVY ..vivnansawnres vis pavmivnes 30 1.774 0.3815 29.42 1.420 0.3592 0.4219 117 Heavy PASE « . ov unsmsinenin dus ss memenrng 32 1.769 0.3391 36.06 1.754 0.3290 0.3646 1.31 Heavy NOW. ; 5 550 a0marins + 24 fh awmsEs 165 2.029 0.1801 198.79 1.690 0.1676 0.2211 1.32 Table 18. Hypothesis tests for variation in mean periodontal index (Pl) score by cross-classification of drinking and smoking variables for NHANES | detailed sample: United States, 1971-74 Chi-square test criteria and significance levels SoUrEe OE Valls tion Degree of Unweighted Weighted Weighted freedom SRS design SRS design complex design Q P-value Q P-value Q P-value DADKING AD): x vivivivniis sb + 444 AAAS HS § ¥ ¥ RE 3 29.54 0.00 28.83 0.00 26.91 0.00 SMOKING (8): 1 vwmvivne tv 42 dav mmmitinwes 4 + ¥ 3 »wmmsEney 2 13.568 0.00 7.89 0.02 9.14 0.01 DYES, vu vv is seis fon S's 3 samme iit? § «99 STR 6 2.52 0.87 5.54 0.48 3.78 0.71 34 Table 19. Hypothesis tests for variation in mean periodontal index (Pl) score by cross-classification of drinking and smoking variables (model with no interaction) for NHANES | detailed sample: United States, 1971-74 Chi-square test criteria and significance levels Source of variation Degree of Unweighted Weighted Weighted freedom SRS design SRS design complex design Q p-value Q p-value Q p-value DOnKING (De sss nner s suas sos mmisain y vee smhaais 3 51.58 0.00 48.89 0.00 40.92 0.00 SMOKING (S) +1 «vm sme rasmran ws Fra asi 2 76.85 0.00 65.45 0.00 42.70 0.00 RACKGERIL. fin5. voemuarmmminmmarbonmmn ss wrasse mteiin ssn Soinsmesn 6 2.52 0.87 5.564 0.48 3.78 0.71 Table 20. Distribution of sample elements according to the r levels of the response profile by the s subclasses Response profile Subclass . Total 7 2 r D aiaioinion = a 4% Wome wR ARGE % 8 ® 1 GREATER E68 PAREN b 5 § CRATER 4 £8 § 9 Ree ni n2 La nr ni, Ds iin vad PARERE EEE DE EP VRRRAEE § 8 eA ER & § RRA § A SRR 4 4 @ @ Gln end n21 na2 wa nar na. Biss ie eR RAE AE TE TER AMSAT E TEE 3 DRRTSEET EER 8 See ig wR 2 88 SIS w 836k AR ns1 ns2 iis nsr ns. Table 21. Number of examined persons ages 25-74 with periodontal index (Pl) scores of zero (none) and greater than zero (some) by race and current smoking status for NHANES | detailed sample: United States, 1971-74 Surent Number £1 score Proportion Race eigasetle examined rT Pl (some) smoker None Some AN SUBIBCIS, + «vvnunnmns 0 15 8 vee wEREREL SE £250 ERTL © 52GB IA, § 2 FE 2,919 1,294 1.625 0.557 WIILE. vse 500 20 SRImETA0 2 0 RRR mR eetctoh & & Sui lmeoes riot: #0 8 A irae Baa 00 Yes 851 351 500 0.588 WILE avs oo va 500 wma e 05 8 BE ATS 5 8 Bw Aa 4 8 Moai edn hes. a rah No 1,574 821 753 0.478 BEI Faasmezsiin a onsniprosaisssamsmens tone a om coinsarmensm pease ay vas SUEATEINIED 3 Aen TETRIS 8 0 3 Yes 230 58 172 0.748 BLACK ivi 0.45.8 3 BE 2050 0.0 0H HA BAR A AR rs ARERR armed No 264 64 200 0.758 Table 22. Weighted number of examined persons ages 25-74 with periodontal index (Pl) scores of zero (none) and greater than zero (some) by race and current smoking status for NHANES | detailed sample: United States, 1971-74 Current Weighted Pl score Pp rtion Race cigarette number oe 0 ) smoker examined None Some Sore Al SUBJBEIR.. ; + 43 FRET Era 5 59 dimen vow etme fa a rT 3,137.5 1,611.3 1,626.2 0.518 White Bene lt Sle re ersmensemstsnecsials o wep SE 3 4 SERRE EE i Eee Yes 1,076.3 459.0 617.3 0.574 WIE. cvvaion 205 3 Haaren a s.00 « 5 5 nari minin fin vows oamibimnac spo Liwmmm mime ares No 1,727.2 952.5 774.7 0.449 BOCK: bivvrviiivinn s + £5 3 swans eH £8 55 0.5 RR Wa § 4S SRE KE Rea ee Yes 171.2 51.8 119.4 0.697 Bl BCR crensmcor ess wih Horses atirrs § PHRASE £06 ARTIST Be § 8 RRR § No 162.8 48.0 114.8 0.705 35 Table 23. Distribution of proportion of some periodontal index (PI > 0.0) and estimated standard errors and design effects by race and current cigarette smoking classification for NHANES | detailed sample: United States, 1971-74 Unweighted design Weighted design Subclass SRS (option 1) option ; ; Weighted SRS (option 2) Complex (option 3) r number h 2 ; Number Proportion Standard 3 Proportion Standard Standard Square root of Race Current cigarette smoking examined examined PI (some) error PI (some) error error design effect WHE: - VBS iv ive ivom in ihimsaiuapncns sv obbidldpie ois 851 0.588 0.0169 1,076.3 0.574 0.0151 0.0250 1.66 White . INOS. 1: «sans vie sents 105 SEE ITEwS 1,574 0.478 0.0126 1,727.2 0.449 0.0120 0.0225 1.88 Black Yes... coavevursmnnna LT, 230 0.748 0.0286 171.2 0.697 0.0351 0.0427 1.22 Black NO =. 5 + 55mm mmtsremmicn a» son minis 264 0.758 0.0264 162.8 0.705 0.0357 0.0547 + 1.83 Table 24. Vector of subclass proportions of some periodontal index (PI > 0.0) and estimated covariance matrix by race and current cigarette smoking classification for NHANES | detailed sample: United States, 1971-74 Subclass #roponijon Estimated covariance matrix X 103 PI (some) Race Current cigarette smoker White YES. os cinin um smens £355 Ean rnnEe se 0 peach iene vo Se Se 0.574 0.626 0.364 0.059 0.000 White ING. vrnmmnon mim ow ow ama SERRE 0.449 0.506 0.070 0.020 Black NEE. i iisstnnninnsins iis dit Pt iGal An A A SEERA ed A Aan hb Aon ee 0.697 1.825 —0.411 Black NO « covsnssnsssnes HEA) WH EE A RAE ea es Fw 0.705 2.995 Table 25. Hypotheses, hypothesis matrices, and test statistics for the model X1 relating the variation in the proportion of some periodontal index (Pl > 0.0) to race and current cigarette smoking classification using sample weights and design effects for NHANES | detailed sample: United States, 1971-74 Chi-square Degree of Hypothesis Hypothesis matrix S2aLtIC Tre COI P-value Hy: There is no variation due to the effect Of FACE wu uw is vss snumnimas ves sss wnwesas isssisy [0100] 26.02 1 <0.01 Ha: There is no variation due to the effect of smoking. ..................coiiiiiiii... [0010] 2.27 1 0.13 H3: There is no variation due to the interaction between race and smoking .................. [0001] 2.92 1 0.09 Table 26. Hypothesis tests for variation in the proportion of some periodontal index (PI > 0.0) by cross-classification of race and smoking cross-classification for NHANES | detailed sample: United States, 1971-74 Chi-square test criteria and significance levels to Degree of Unweighted Weighted Weighted Square root of S F HrcEolalalion freedom SRS design SRS design complex design design effect Q P-value Q P-value Q P-value Race (RY: « , sueniwiniss i 13 sans vad 1 98.59 0.00 50.20 0.00 26.02 <0.01 1.39 Smoking (S). -....ciiviviieinann 1 5.04 0.02 4.78 0.03 2.27 0.13 1.45 BRS vs «3 52a miosis sons mains 1 7.22 0.01 6.11 0.01 2.92 0.09 1.45 36 Table 27. Hypothesis tests for variation in the proportion of some periodontal index (Pl > 0.0) by cross-classification of race and smoking (reduced model) for NHANES | detailed sample: United States, 1971-74 Chi-square test criteria and significance levels Source of variation Degree of Unweighted Weighted Weighted Square root of freedom SRS design SRS design complex design design effect Q P-value Qa P-value Q P-value White versus black mean. ........ 1 99.47 0.00 50.17 0.00 27.20 <0.01 1.36 Smoking with whites ............ 1 26.87 - 0.00 42.19 0.00 38.63 <0.01 1.05 Lack of fit. ..............couunn. 1 0.06 0.80 0.02 0.88 0.01 0.92 1.41 37 Appendixes Contents I. Definitions of terms and Variables . ..............uiititttit titties 39 DICLARY VAHIBDNOS.. +. 2 5.5 vam 5:9 55 500000 5 10100 0 0 iB 0 me mm, Ho 810 0 0) 0) 0 RL i 0 39 Denial 3nd MBOICE) VANBDIES -..... +. +5555 5 5res 2.3 5d lo F £10 $15 51 3.90 3 0 oo 0 wiim wie 9 101 ma Bd oem Hn 58.4100 800 0 00 Wh mm im 39 BoRaVIOTAl VATIAIIEE . «vv vv w vow sien are wow 0 wow wow mbites 0700 Wilp 548 6 370 010 ls 0 Bln 9 at min wie ww wc 1 Er 00 ee te 4 39 1:2 'COMPUtING CONION Cart THIS. sv ve wom wis vm mie i his 5 29 053 16 518 #08 578 49 i+ W000 008 5 B08 600 04 W018 970 0 6.0 000 3. wid 8 Wah mowers 40 Means and variances. ............. 300 ip 3 55 023 2 3.5 04 5 10 fs x 12 Ta 1 rT me wan si ou LE 1 5 003 0 0 3 0 37 8 40 Regression MOIS .. ott it ttt ttt tt ete e et tte ee tee ee 40 BINOVA coh 5 500 Tt 3 wii Fe 5.30 E05 0 008 B10 0B 00 00 0 5 mw #50 0 800 00 0 4 0 3 0% 00 0 98 0 3 1 00 92 ce 4 00 BW 9 8 40 COMINGBIICY TAIIBS: . .. - c vn 50 6 5: 50m 515 Twa 518 0: 8 © 808 8 5 Weim 01% m0 80 Bok oi 40 3 0 Bm 6 300 B00 HL 5 8 0 00 6 8 8 0 Tw 90 GLE 9 we 40 38 Appendix |. Definitions of terms and variables Dietary variables Calories—The total energy intake determined from the 24-hour dietary recall measured in kilocalories. Sweets—The sum of the reported frequencies for the ingestion of food from the three categories of desserts and sweets, candy, and beverages (sweetened, carbonated, and non-carbonated). Dental and medical variables Periodontal index—Periodontal index score for entire mouth as given in the data provided. DMF—Sum of decayed (D), missing (M), and filled (F) permanent teeth. Quetelet’s index—Body mass index which stand- ardizes weight for height and permits indirect prediction of adiposity. Defined as weight+height? using weight in kilograms and height in centimeters. Behavioral variables Drinking—Categorical variable concerning alcohol consumption derived from three other variables. The four categories are 1) Those who claimed not to have had a drink in the past (called “none” in the tables), 2) Those who claimed to drink no more than once a week and when they did drink had three or fewer drinks (called “little” in the tables), 3) Those who stated they drank more often than once a week but have three or fewer drinks at a time, or those who drink no more often than once a week but have four or more drinks when they do drink (called “moderate” in the tables), and 4) Those who claimed to drink more often than once a week and have four or more drinks at a time (called “heavy” in tables). Smoking—Categorical variable derived from sev- eral other variables. The three categories are 1) Never have used tobacco in quantities up to or equal to the amounts stated in the medical history questionnaire, that is, at least 100 cigarettes, 50 cigars, or three packages of pipe tobacco during the subjects’ lifetime. 2) Have used tobacco at least up to the amounts stated for at least one of the categories stated in the questionnaire, but do not use tobacco now, and 3) Used tobacco at the time of the interview, in amounts at least as large as those stated in the questionnaire. Current cigarette smoker—Categorical variable for current cigarette smoking status. The categories are 1) Have smoked more than 100 cigarettes and smoke cigarettes now, and 2) Have never smoked more than 100 cigarettes or do not smoke cigarettes now. 39 Appendix Il. Computing control card files The following subsections contain representative control card files that were used for the run illustrated in the corresponding previous section. All the program statements are intended for the OSIRIS IV system available at the University of Michigan, except for the ANOVA and contingency table analyses which were processed sequentially through OSIRIS IV and then through the GENCAT weighted least squares program. Means and variances Example 1 shows the OSIRIS IV commands were used to generate the multiple SECU results for DMF teeth and calories displayed in table 10. &USTATS computes means, standard deviations, and standard errors using the weighted data, whereas &PSALMS computes estimates and sampling errors for ratio means from stratified clustered sample designs. Regression models Example 2 shows the OSIRIS IV commands were used to generate the simple regression model results for DMF teeth on age and calories on age shown in table 14. ®RESSN computes standard regressions using the weighted data ignoring the sample design, whereas &REPERR computes regression statistics and their sampling errors for data from clustered sample designs using balanced half sample replications. 40 ANOVA Example 3 shows the five command files were used to generate the results under options 1-3 displayed in tables 17-19. Step 1 uses the OSIRIS IV &USTATS command to generate means and their standard errors (unweighted and weighted) for the 12 drinking and smoking classifications. The vector of means and the corresponding covariance matrix then were read under the direct input option of GENCAT to generate the analyses for options 1 and 2, according to whether the sampling weights were included in the analysis in Steps 2 and 3, respectively. Finally, Steps 4 and 5 were used to generate the ratio means and the covariance matrix under the cluster sample design and to produce the chi- square statistics under option 3. Contingency tables Example 4 shows the five command files were used to generate the results in tables 21-27. Step 1 uses the OSIRIS IV &TABLES command to obtain the 4 X 2 table with race-current cigarette smoking categories forming the rows and periodontal index forming the columns (none, some). These unweighted and weighted tables then were used as input to GENCAT in Steps 2 and 3. Step 4 utilizes the &PSALMS routine to obtain the variance-covariance matrix for the weighted pro- portions taking into account the complex sample design. Finally, step 5 contains the control cards needed to run GENCAT on the weighted data. Example 1 $RUN ISR:OSIRIS.IV SPRINT=xPRINT#* &RECODE R1=1 NAME R1'COUNTER' &END &USTATS DICTIN=-DICT DATAIN=-DATA WEIGHTED STATS FOR DMF AND CALORIES BY AGE GROUP VARS=V6042,V203 WTVAR=VS0- REP=(V49=1-17/18-24/25-34/35-44/45-54/55-64/65-74/1-74) &END &PSALMS DICTIN=-DICT DATAIN=-DATA SAMPLING ERROR ANALYSIS OF DMF AND CALORIES BY AGE GROUP R=1 WTVAR=V90 SECU=V96 ST=V91- REP=(V49=1-17/18-24/25-34/35-44/45-54/55-64/65-74/1-74) MOD=MULT ST=1-10 SECU=169,106,125,156,197,83,108,61,89,119 MOD=PAIR ST= PAR=V6042/R1 PAR=V6042/R1 PAR=V6042/R1 PAR=V6042/R1 PAR=V6042/R1 PAR=V6042/R1 PAR=V6042/R1 PAR=V6042/R1 PAR=V203/R1 PAR=V203/R1 PAR=V203/R1 PAR=V203/R1 PAR=V203/R1 PAR=V203/R1 PAR=V203/R1 PAR=V203/R1 &END 11-35 SUB=1, 1 SUB=2,2 SUB=3, 3 SUB=4,4 SUB=5,5 SUB=6,6 SUB=7,7 SUB=8,8 SUB=1, 1 SUB=2, 2 SUB=3, 3 SUB=4,4 SUB=5,5 SUB=6, 6 SUB=7,7 SUB=8, 8 41 42 Example 2 $RUN ISR:O0SIRIS.IV SPRINT=%PRINT#* ®RESSN DICTIN=-DICT DATAIN=-DATA REGRESSION OF DMF AND CALORIES ON AGE WTVAR=V90 V=V49 DEPV=V6042 V=V49 DEPV=V203 &END ®RESSN INCLUDE V50=1 AND REGRESSION OF DMF WTVAR=V90 V=V49 DEPV=V6042 V=V49 DEPV=V203 &END ®RESSN INCLUDE V50=2 AND REGRESSION OF DMF WTVAR=V90 V=V49 DEPV=V6042 V=V49 DEPV=V203 &END ®RESSN INCLUDE V50=3 AND REGRESSION OF DMF WTVAR=V90 V=V49 DEPV=V6042 V=V4S9 DEPV=V203 &END ®RESSN INCLUDE V50=1 AND REGRESSION OF DMF WTVAR=VI0 V=V49 DEPV=V6042 V=V49 DEPV=V203 &END ®RESSN INCLUDE V50=2 AND REGRESSION OF DMF WTVAR=V90 V=V49 DEPV=V6042 V=V4S DEPV=V203 &END ®RESSN INCLUDE V50=3 AND REGRESSION OF DMF WTVAR=V90 V=V49 DEPV=V6042 V=V49 DEPV=V203 &END V51=1 AND CALORIES V51=1 AND CALORIES V51=1 AND CALORIES V51=2 AND CALORIES V51=2 AND CALORIES V51=2 AND CALORIES ON ON ON ON ON ON &REPERR DICTIN=-DICT DATAIN=-DATA BHS MODEL FOR DMF AND CALORIES AGE AGE AGE AGE AGE AGE BY BY BY BY BY BY RACE-SEX RACE-SEX RACE-SEX RACE-SEX RACE-SEX RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS ,RCOEFF ,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END CATEGORIES CATEGORIES CATEGORIES CATEGORIES CATEGORIES CATEGORIES &REPERR INCLUDE V50=1 AND V51=1 BHS MODEL FOR DMF AND CALORIES BY RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS, RCOEFF ,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END &REPERR INCLUDE V50=2 AND V51=1 BHS MODEL FOR DMF AND CALORIES BY RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS ,RCOEFF,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END &REPERR INCLUDE V50=3 AND V51=1 BHS MODEL FOR DMF AND CALORIES BY RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS, RCOEFF,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END &REPERR INCLUDE V50=1 AND V51=2 BHS MODEL FOR DMF AND CALORIES BY RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS, RCOEFF ,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END &REPERR INCLUDE V50=2 AND V51=2 BHS MODEL FOR DMF AND CALORIES BY RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS, RCOEFF,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END &REPERR INCLUDE V50=3 AND V51=2 BHS MODEL FOR DMF AND CALORIES BY RACE-SEX SECU=V97 ST=V91 WTVAR=V94 VAR=49,6042,203- STATS= (MEANS ,RCOEFF,MULTR) REGR=TOT ST=1-35 MOD=BHS V=49 DEPV=6042,203 &END CATEGORIES CATEGORIES CATEGORIES CATEGORIES CATEGORIES CATEGORIES 43 44 STEP STEP . 00000000 Example 3 1 %% MEANS AND VARIANCES (WEIGHTED AND UNWEIGHTED) *# $RUN ISR:0SIRIS.IV SPRINT=%PRINTx &RECODE=1 R1=BRAC(V4501, 1 R2=BRAC(V7501, 1 R4=COMBINE R2(3 R6=V1089 MDATA R4(99) &END &USTATS DICTIN=DICTREP4 DATAIN=DATAREP4 UNIVARIATE STATISTICS RECODE=1 VARS=6005,6008 REP=(R4=0/1/2/3/4/5/6/7/8/9/10/11) &END &USTATS WEIGHTED UNIVARIATE STATISTICS RECODE=1 VARS=6005,6008 REP=(R4=0/1/2/3/4/5/6/7/8/9/10/11) WT=R6 &END =0,2=1 =0,2=1 ) ,R1(4 ~—w = 2 ** GENCAT ANALYSIS USING UNWEIGHTED DATA ** $RUN SJS6:GENCAT 1=*%*SOURCE* 3=*PRINT#* 8=-TEMP . B 3 1 1 UNWEIGHTED ANALYSIS OF P.I. 12 1 (6F9.4) 1.617914 2.037624 2.348564 .961378 1.280140 1.738261 1.003090 1.148133 1.731346 1.773667 1.768750 2.028848 3 (6F9.3) 0.0115326 0.0560125 0.0300307 0.0053.778 0.0164335 0.0093619 0.0169859 0.0179936 0.0096814 1.455348 1.149650 0.0324495 7 1 12 1 (12F1.0) FULL MODEL 0000001 000011011 000000011011 000000000011 000001001001 000000001001 000000000001 8 1 2 (12F1.0) SMOKE EFFECT / FULL MODEL 01 001 8 1 3 (12F1.0) DRINK EFFECT / FULL MODEL 0001 00001 000001 7 1 6 1 (12F1.0) MODEL WITH NO INTERACTION "M1111 STEP 011011011011 001001001001 000111111111 000000111111 000000000111 : 8 1 2 (6F1.0) SMOKE EFFECT / NO INTERACTION 01 001 , B 1 3 (6F1.0) DRINK EFFECT / NO INTERACTION 0001 ‘ 00001 000001 7 1 4 1 (12F1.0) 111000000111 000111111000 012000000012 000001001000 8 1 1 (42.0) i-1.0'0 8 1 1 (4F1.0) 0010 8 1 1 (4F1.0) 0001 3 ** GENCAT ANALYSIS USING WEIGHTED DATA *x* 5 3 1 1 WEIGHTED ANALYSIS OF P.I. 12 1 (6F9.4) 1.423589 1.835681 1.903959 0.800040 0.965758 1.516229 0.852681 0.929804 1.462482 al Ja2n308s 1.689494 3 6F9.3 0.0103659 0.0471113 0.0249002 0.0044020 0.0129284 0.0080.248 0.0161660 0.0142776 0.0081406 1.290554 1.082204 0.0280853 7 1 12 1 (12F1.0) FULL MODEL 000000011011 000000000011 000001001001 000000001001 000000000001 8 1 2 (12F1.0) SMOKE EFFECT / FULL MODEL 01 001 8 1 3 (12F1.0) DRINK EFFECT / FULL MODEL 0001 00001 45 46 STEP 000001 7 1 6 1 (12F1.0) MODEL WITH NO INTERACTION 111311311111 011011011011 001001001001 000111111111 000000111111 000000000111 8 1 2 (6F1.0) SMOKE EFFECT / NO INTERACTION 01 001 8 1 3 (6F1.0) DRINK EFFECT / NO INTERACTION 0001 00001 000001 7 1 4 1 (12F1.0) 111000000111 000111111000 012000000012 000001001000 8 1 1 (4F2.0) 1-100 8 1 1 (4F1.0) 0010 8 1 1 (4F1.0) 0001 4 x* &PSALMS RUN TO GENERATE RATIO MEANS & THEIR COVARIANCE STRUCTURE UNDER THE CLUSTERED DESIGN *#% $RUN ISR:OSIRIS.IV SPRINT=-PR &RECODE R1=V4501 R2=V7501 TABLE A,COLS 1-3,ROWS 1(1-3),2(4-6),3(7-9),4(10-12) ENDTAB R3=TABLE(R1,R2,TAB=A) R100=1 R101=V1089 MDATA R1(99),R2(99) &END &PSALMS DICTIN=DICTREP4 DATAIN=DATAREP4 OUTPUT=-SE4 HANES MEAN P.I. BY DRINK-SMOKE CATEGORIES (DETAILED WEIGHTS) R=1 SORT=4000 PSU=V97 ST=VS1 W=R101 REP=(R3=1/2/3/4/5/6/7/8/9/10/11/12) OUT ST=1-35 MOD=PAIR PAR=V6008/R100-V6008/R100 SUB=1,1,2,2 P=FULL SUB=1,1,3,3 P=FULL SUB=1,1,4,4 P=FULL SUB=1,1,5,5 P=FULL SUB=1,1,6,6 P=FULL SUB=1,1,7,7 P=FULL SUB=1,1,8,8 P=FULL SUB=1,1,9,9 P=FULL SUB=1,1,10,10 P=FULL SUB=1,1,11,11 P=FULL SUB=1,1,12,12 P=FULL SUB=2,2,3,3 P=FULL SUB=2,2,4,4 P=FULL SUB=2,2,5,5 P=FULL SUB=2,2,6,6 P=FULL SUB=2,2,7,7 P=FULL SUB=2,2,8,8 P=FULL SUB=2,2,9,9 P=FULL SUB=2,2,10,10 P=FULL SUB=2,2,11,11 P=FULL SUB=2,2,12,12 P=FULL SUB=3,3,4,4 P=FULL SUB=3,3,5,5 P=FULL SUB=3,3,6,6 P=FULL SUBB=3,3,7,7 P=FULL SuB=3,3,8,8 P=FULL SUB=3,3,9,9 P=FULL SUB=3,3,10,10 P=FULL SuB=3,3,11,11 P=FULL SUB=3,3,12,12 P=FULL SUB=4,4,5,5 P=FULL SUB=4,4,6,6 P=FULL SUB=4,4,7,7 P=FULL SUB=4,4,8,8 P=FULL SUB=4,4,9,9 P=FULL SUB=4,4,10,10 P=FULL SUB=4,4,11,11 P=FULL SUB=4,4,12,12 P=FULL SUB=5,5,6,6 P=FULL SUB=5,5,7,7 P=FULL SUB=5,5,8,8 P=FULL SUB=5,5,9,9 P=FULL SUB=5,5,10,10 P=FULL SUB=5,5,11,11 P=FULL SUB=5,5,12,12 P=FULL SUB=6,6,7,7 P=FULL SUB=6,6,8,8 P=FULL SUB=6,6,9,9 P=FULL SUB=6,6,10,10 P=FULL SUB=6,6,11,11 P=FULL SUB=6,6,12,12 P=FULL SuB=7,7,8,8 P=FULL SUB=7,7,9,9 P=FULL SuB=7,7,10,10 P=FULL SUB=7,7,11,11 P=FULL SuB=7,7,12,12 P=FULL SUB=8,8,9,9 P=FULL SUB=8,8,10,10 P=FULL SUB=8,8,11,11 P=FULL SUB=8,8,12,12 P=FULL SUB=9,9,10,10 P=FULL SUB=9,9,11,11 P=FULL SUB=9,9,12,12 P=FULL SUB=10,10, 11, 11 P=FULL SUB=10,10,12,12 P=FULL SUuB=11,11,12,12 P=FULL &END &SM15:MATGEN INPUT=-SE4 3=-T 47 48 STEP 12 1 70 $ENDFILE 5 *x GENCAT ANALYSIS USING CLUSTER DESIGN COVARIANCE MATRIX $RUN SJS6:GENCAT 1=*SOURCE* 3=*PRINT* 4=-T 8=-U 5 3 4 1 DETAILED SAMPLE WEIGHT ANALYSIS OF P.I. 12 12 1 12 1 (12F1.0) FULL MODEL 00000000 000011011 000000011011 000000000011 000001001001 000000001001 000000000001 8 1 2 (6F1.0) SMOKE EFFECT / FULL MODEL 01 001 8 1 3 (6F1.0) DRINK EFFECT / FULL MODEL 0001 00001 000001 7 1 6 1 (12F1.0) MODEL WITH NO INTERACTION 1171111171111 011011011011 001001001001 000111111111 000000111111 000000000111 8 1 2 (6F1.0) SMOKE EFFECT / NO INTERACTION 01 001 8 1 3 (6F1.0) DRINK EFFECT / NO INTERACTION 0001 00001 000001 7 1 4 1 (12F1.0) 4 PARAMETER MODEL 111000000111 000111111000 012000000012 000001001000 8 1 1 (4F2.0) 1-100 B 1 1 (4F1.0) 001 RB 1 1 (4F1.0) 0001 49 50 Example 4 STEP 1 ** UNWEIGHTED AND WEIGHTED FREQUENCIES #*x* STEP DIS INT FI=NEW.REP3 R @NEW INT FI=NEW.REP3 V=ALL DES Vv=89,90,91 ONEWAY V=50,100,7339,6008 OP=x TWOWAY V=50,7339 OP=% TWOWAY V=% OP=%x C=V92:1 TWOWAY V=% OP=%x C=V92:1%xV100:NONE TWOWAY V=x OP=% C=V92:1%xV100:SOME $R ISR:OSIRIS.IV SPRINT=%PRINTx* &MIDASFILE INPUT=NEW.REP3 &RECODE RECODE= 1 IF MDATA(V91) THEN REJECT R1=BRAC(V50,1=0,2=1) R2=BRAC(V7339,1=0, R3=BRAC(V100,1=0,2 R4=COMBINE R2(2),R R5=BRAC(R4,0=0, 1=1 R6=3854.%V91 MDATA R3(99),R5(99) &END &TABLES BIVARIATE FREQUENCIES: UNWEIGHTED RECODE= 1 VAR=R3 ST=R5 &END &TABLES BIVARIATE FREQUENCIES: WEIGHTED RECODE=1 WTVAR=R6 VAR=R3 ST=R5 &END 2 *% GENCAT ANALYSIS OF UNWEIGHTED FREQUENCIES #*x* $RUN SJS6:GENCAT 1=%SOURCE* 3=-PRINT 8=-V 5 1 1 UNWEIGHTED ANALYSIS ° 4 2 (2F4.0) 351 500 821 753 58 172 64 200 1 2 4 1 1 1 (2F1.0) 01 7 1 4 1 (4F2.0) FULL MODEL {iva 1 1-1-1 1-1 1-1 1-1-1 1 8 1 1 (4F1.0) TEST FOR RI EFFECT (RACE) 0100 8 1 1 (4F1.0) TEST FOR R2 EFFECT (SMOKE) 0010 STEP 8 1 1 BY R2 INTERACTION 0001 7 1 3 NO INTERACTION 1111 I 1=1=1 1~% I=} 8 1 1 NO INTERACTION 010 8 1 1 NO INTERACTION 001 3 *% GENCAT ANALYSIS OF WEIGHTED FREQUENCIES ** 5 1 1 4 2 458.98 617.31 952.49 774.69 51.80 119,36 48.04 114.80 1 2 4 1 01 7 1 4 1111 1 1-1-1 1-1 1-1 1-1-1 1 8 1 1 EFFECT (RACE) 0100 8 1 1 EFFECT (SMOKE) 0010 8 1 1 BY R2 INTERACTION 0001 % 1 3 NO INTERACTION ¥.9-9 1 1 1-1-1 1-1 1-1 8 1 1, NO INTERACTION 010 8 1 1 NO INTERACTION 001 1 (4F1.0) (4F2.0) (3F1.0) {3F1.0) (2F7.2) (2F1.0) (4F2.0) (4F1.0) (4F1.0) (4F1.0) (4F2.0) (3F1.0) (3F1.0) TEST FOR RI MODEL WITH R1 EFFECT / R2 EFFECT / WEIGHTED ANALYSIS FULL MODEL TEST FOR RI TEST FOR R2 TEST FOR RI MODEL WITH R1 EFFECT / R2 EFFECT / STEP 4 ** &PSALMS RUN TO GENERATE COVARIANCE MATRIX OF WEIGHTED FREQUENCIES UNDER CLUSTER SAMPLE DESIGN *x* $COPY -PRINT *MSINK#* $RUN ISR:0SIRIS.IV SPRINT=%xPRINT* &MIDASFILE INPUT=NEW.REP3 &RECODE 51 52 STEP * % R1=V50 R2=V7339 TABLE A,COLS 1-2,ROWS 1(1-2),2(3-4) ENDTAB R3=TABLE (R1,R2,TAB=A) R4=BRAC(V100,1=0,2=1) R100=1 R101=3854,%V91 MDATA R1(99),R2(99),R4(99) &END &PSALMS OUTPUT=-SE HANES 4 X 2 TABLES (SMOKING VS RACE) R=1 PSU=V97 ST=V191 W=R101 REP=(R3=1/2/3/4) OUT SORT=4000 ST=1-35 MOD=PAIR NUM=12 PAR=R4/R100-R4/R100 SUB=1,1,2,2 P=FULL NUM=13 SUB=1,1,3,3 P=FULL NUM=14 SUB=1,1,4,4 P=FULL NUM=23 SUB=2,2,3,3 P=FULL NUM=24 SUB=2,2,4,4 P=FULL NUM=34 SUB=3,3,4,4 P=FULL &END &SM15:MATGEN INPUT=-SE 3=-T 4 1 70 $ENDFILE 5 xx GENCAT ANALYSIS OF WEIGHTED FREQUENCIES UNDER OPTION 3 $RUN SJS6:GENCAT 1=*SOURCE* 3=-0UT 4=-T 8=-0 5 3 4 1 WEIGHTED ANALYSIS (VIA PSALMS) 4 4 1 2 7 1 4 1 (4F2.0) FULL MODEL i 1°11 1 1-1-1 1-1 1-1 1-1-1 1 8 1 1 (4F1.0) TEST FOR RI EFFECT (RACE) 0100 8 1 1 (4F1.0) TEST FOR R2 EFFECT (SMOKE) 0010 8 1 1 (4F1.0) TEST FOR RI BY R2 INTERACTION 0001 ’ 7 1 3 1 (4F2.0) MODEL WITH NO INTERACTION "17% 1 ) Yt] 1=1 1=1 8 1 1 (3F1.0) R1 EFFECT / NO INTERACTION 010 8 1 1 (3F1.0) R2 EFFECT / NO INTERACTION 001 $COPY -OUT *PRINTx ¥U.S. GOVERNMENT PRINTING OFFICE: 1982-361-161:514 ! ST ERAS RETURN PUBLIC HEALTH LIBRARY TOm==p 42 Warren Hall _ fl Vital and Health Statistics series descriptions SERIES 1. SERIES 2. SERIES 3. SERIES 4. SERIES 10. SERIES 11. SERIES 12. SERIES 13. Programs and Collection Procedures.—Reports describing the general programs of the National Center for Health Statis- tics and its offices and divisions and the data collection methods used. They also include definitions and other material necessary for understanding the data. Data Evaluation and Methods Research.—Studies of new statistical methodology including experimental tests of new survey methods, studies of vital statistics collection methods, new analytical techniques, objective evaluations of reliability of collected data, and contributions to statistical theory. Analytical and Epidemiological Studies.— Reports presenting analytical or interpretive studies based on vital and health sta- tistics, carrying the analysis further than the expository types of reports in the other series. Documents and Committee Reports.—Final reports of major committees concerned with vital and health statistics and documents such as recommended model vital registration laws and revised birth and death certificates. Data From the National Health Interview Survey.— Statistics on illness, accidental injuries, disability, use of hospital, medical, dental, and other services, and other health-related topics, all based on data collected in the continuing national household interview survey. Data From the National Health Examination Survey and the National Health and Nutrition Examination Survey.—Data from direct examination, testing, and measurement of national samples of the civilian noninstitutionalized population provide the basis for (1) estimates of the medically defined prevalence of specific diseases in the United States and the distributions of the population with respect to physical, physiological, and psychological characteristics and (2) analysis of relationships among the various measurements without reference to an explicit finite universe of persons. Data From the Institutionalized Population Surveys.—Dis- continued in 1975. Reports from these surveys are included in Series 13. Data on Health Resources Utilization.—Statistics on the utili- zation of health manpower and facilities providing long-term care, ambulatory care, hospital care, and family planning services. SERIES 14. SERIES 15. SERIES 20. SERIES 21. SERIES 22. SERIES 23. Data on Health Resources: Manpower and Facilities.— Statistics on the numbers, geographic distribution, and char- acteristics of health resources including physicians, dentists, nurses, other health occupations, hospitals, nursing homes, and outpatient facilities. Data From Special Surveys.— Statistics on health and health- related topics collected in special surveys that are not a part of the continuing data systems of the National Center for Health Statistics. Data on Mortality.—Various statistics on mortality other than as included in regular annual or monthly reports. Special analyses by cause of death, age, and other demographic variables; geographic and time series analyses; and statistics on characteristics of deaths not available from the vital records based on sample surveys of those records. Data on Natality, Marriage, and Divorce.— Various statistics on natality, marriage, and divorce other than as included in regular annual or monthly reports. Special analyses by demo- graphic variables; geographic and time series analyses; studies of fertility; and statistics on characteristics of births not available from the vital records based on sample surveys of those records. Data From the National Monthly and Natality Surveys.— Discontinued in 1975. Reports from these sample surveys based on vital records are included in Series 20 and 21, respectively. Data From the National Survey of Family Growth.—Statis- tics on fertility, family formation and dissolution, family planning, and related maternal and infant health topics derived from a periodic survey of a nationwide probability sample of ever-married women 15-44 years of age. For a list of titles of reports published in these series, write to: Scientific and Technical Information Branch National Center for Health Statistics Public Health Service Hyattsville, Md. 20782 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES POSTAGE AND FEES PAID Public Health Service U.S. DEPARTMENT OF HHS Office of Health Research, Statistics, and Technology HHS 396 National Center for Health Statistics 3700 East-West Highway Third Class Hyattsville, Maryland 20782 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE, $300 4 41451 From the Office of Health Research, Statistics, and Technology , DHHS Publication No. (PHS) 82-1366, Series 2, No. 92 For a listing of publications in the VITAL AND HEALTH STATISTICS series call 301-436—-NCHS U.C. BERKELEY LIBRARIES C0e120bL59 rs 4 ee Gnd d rE