Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
APPENDIX C Federal Health Data Sources The United States possesses one of the world's leading statistical sys- tems, its current status having evolved over the course of the nation's histo- ry. As the role of the federal government has increased over the years, so has the demand for more and better statistics at the national and subnational levels. The federal statistical system in the United States is decentralized, with authority and responsibility for statistical activities spread among sev- eral agencies of the government. There exists an abundance of statistical information related to the health status and the health care industry of this nation. A number of federal and state agencies gather and use health-related data. The private sector is also a major collector and user of such data. Yet, as shown in the body of the report, the current national data systems do not provide the information needed to assess the effect of changes in the financing, organization, and delivery of health care or the impact of socioeconomic trends on the quali- ty, cost, and outcomes of care. An understanding of the strengths and weaknesses of the current status of federal statistics and the prospects for continued development and progress in future years requires knowledge of the genesis of these systems and an understanding of the social, economic, public health and political factors that have led to a gradually increasing involvement of both the executive and legislative branches of the federal government in the collection and maintenance of health statistics. 125
126 APPENDIX C EVOLUTION OF FEDERAL lIEALTlI STATISTICS For many years after the formation of the federal union, responsibility for the collection of statistics in the health field was located primarily at the level of the state and local governments. Federal involvement was limited to a large extent to meeting the information needs for the most pressing health and related problems affecting the country and to the collection of vital records from the states. Notification of diseases in this country began in the colonial times at the local level, particularly in port cities. However, the collection of the data was limited to times when epidemics of pestilential diseases threatened or were in progress. Statewide notification was first instituted in 1883, when Michigan passed a law requiring reporting of certain diseases (Bureau of the Census, 1960~. In response to the need for nationwide information on epidemic diseas- es, Congress passed a law in 1878 authorizing the collection of morbidity reports for use in connection with quarantine measures against pestilential diseases such as cholera, smallpox, plague, and yellow fever. This law was the start of the responsibilities of the Public Health Service (PHS) in the collection of data on communicable diseases and the surveillance of the incidence and distribution of diseases. A year later, a specific appropriation was made for such collection. In 1893, Congress passed another law that provided for the collection of information each week from state and municipal authorities throughout the United States. By the 1950s national health data sources were primarily comprised of the reporting by state and territorial authorities of morbidity statistics for communicable diseases and their surveillance and the death registration system. Completeness of reporting continued to vary greatly, and many illnesses were not reported to the authorities. Various health programs pro- duced statistics specific to their programs, but they were neither national in scope nor comparable for different states. CulTent data that were represen- tative of the general health status of the U.S. population were not available regularly. A national health survey was conducted by the Works Project Adminis- tration in 1935-1936 as part of a comprehensive National Health Inventory, covering 700,000 urban families in the depressed economic conditions of the time. The survey had two purposesto promote knowledge of the population's health status and to create jobs for the unemployed. Twenty years later, statisticians inside and outside the government were still relying primarily on the 1935-1936 National Health Survey and a series of supplemental studies carried out in the 1940s and early l950s as the basis for estimating the current level of illness and disability in the United
FEDERAL HEALTH DATA SOURCES 127 States.i Those data had become outdated. During the interim much had happened to change the country, the people, and their health status. The Great Depression was over; the country had gone through a major war and had recovered; the population of the United States was more urbanized; medical care had advanced and the people's health had improved. During that period major advances had been made in statistical survey theory and methodology. The late 1940s and the early l950s saw a series of local surveys conducted by various state arid private organizations to study morbidity, disability, and handicapping conditions. Many of these studies broke new ground but none was national in scope. The Bureau of the Census and the PHS had provided advice to those carrying out these local surveys. It was clear by the early l950s that the federal government's role in health statistics could best be carried out by providing leadership in devel- opment of a national health data resource based on the probability sampling of households. The National Health Survey Act In 1953 the National Committee on Vital and Health Statistics (NCVHS)2 recommended that a national health survey be established on a permanent basis. (NCHS, Ser.4, No.5, 1966) In 1956 Congress enacted PL 84-652, the National Health Survey Act (reproduced in Appendix D). The act au- thorized a three-part program: a continuing survey using probability sam- pling methods, special studies, and methodological studies. It authorized the surgeon general to make available technical advice and assistance on the application of statistical methods in health and medicine. The content of the program of the National Health Survey was influ- enced by the varied nature of the data needed and the requirement in the act that corresponding information be obtained for both the well and the ill population. The National Health Survey was conceived as a program of surveys, with its objectives and approaches being flexible to adapt to chang- ing needs for data and advances in statistical method. The activities of the National Health Survey as it is currently comprised are divided into three 1 these studies were supplements to the Census Bureau's Current Population Survey, re- quested by the Social Security Administration and the Public Health Service. The supplement in 1943 (repeated in 1949 and 1950) dealt with the prevalence of disability, and another in 1951 dealt with the prevalence of arthritis. These were the first applications of national probability sampling of the population to the measurement of health. 2The U.S. National Committee on Vital and Health Statistics was established in 1949 by the surgeon general of the U.S. Public Health Service to put into effect a recommendation of the First World Health Assembly. On July 23, 1974, this committee was legislatively established under Section 306(k) of the Public Health Service Act (PL 93-353).
128 APPENDIX C major parts, with data collected from three major types of sources. These are (1) the Health Interview Survey, (2) the Health Examination Survey, and (3) the Health Records Surveys. A year after the passage of the National Health Survey Act, the House- hold Health Interview Survey (later renamed the National Health Interview Survey) began with the objective of producing general data on the nation's health. For several years this Household Health Interview Survey was the only major activity of the National Health Survey while the special studies also authorized by the law were being developed. In 1959, a second major component of the National Health Survey, the Health Examination Survey, was started, aimed at gathering health data by means of a physical examina- tion of a sample of the population, and reached full-scale data collection by spring 1961. Special studies also authorized by the law were being devel- oped. By 1961 several of the survey activities had begun that later made up the Health Records Survey program. CURRENT DATA SoURCES3 The federal health statistics system, like the federal statistics system as a whole, is decentralized across the federal government. The majority of health statistical activities are undertaken by the Department of Health and Human Services, although the Department of Defense, the Veterans Admin- istration, the Environmental Protection Agency, the Bureau of Labor Statis- tics, and other agencies collect and use health statistics in carrying out their respective functions. Within the Department of Health and Human Services, the organization and decision-making structure for health statistical activities is further de- centralized among the various organizational components of the department that have distinct responsibilities for gathering data directly related to their . . . . program responsl '1 ltles. The section below briefly describes the data systems that these agencies operate that either contribute, or could contribute, to the understanding of the dynamics of our health delivery system and its impact on the health status of the American people. No attempt is made to provide an exhaustive listing or summary of all the health statistical activities carried out by feder- al agencies and programs. Only selected statistical activities of the Nation- al Center for Health Statistics (NCHS) and the department relevant to this present study are described. ~ . The information in this section is based on' and excerpted frond publications and other documents of the agencies of the federal government. These sources are provided at the end of the report.
FEDERAL HEALTH DATA SOURCES National Center for Health Statistics 129 NCHS is the focal federal agency for the production, analysis, and dissemination of general-purpose health statistics. It is recognized as one of the principal statistical agencies of the federal government. The center was established in 1960 when the National Office of Vital Statistics and the National Health Survey Program in the Division of Public Health Methods were placed in one agency within the U S. Public Health Service. Since its creation, the center has been moved to various locations within the organizational structure of the Public Health Service. Most recently, in 1987, it was placed as one of the Centers in the Centers for Disease Control (CDC). Section 306 of the Public Health Service Act establishes the center in the Department of Health and Human Services and specifies that it shall collect statistics on: . . . The extent and nature of illness and disability of the population of the United States (or of any groupings of the people included in the population), including life expectancy, the incidence of various acute and chronic illnesses, and infant and maternal morbidity and mor- tality; The impact of illness and disability of the population on the econo- my of the United States and on other aspects of the well-being of its population (or such groupings); · Environmental, social, and other health hazards; Determinants of health; · Health resources, including physicians, dentists, nurses, and other health professionals by specialty and type of practice, and the sup- ply of services by hospitals, extended care facilities, home health agencies; . . Utilization of health care, including (1) ambulatory health services by specialties and types of practice of the health professionals pro- viding such services and (2) services of hospitals, extended care facilities, home health agencies, and other institutions; Health care costs and financing, including the trends in health care prices and cost, the sources of payments for health care services, and federal, state, and local governmental expenditures for health care services; and Family formation, growth, and dissolution.
130 National Health Interview Survey APPENDIX C The National Health Interview Survey (NHIS) is the principal source of information on the general health status of the civilian, noninstitutionalized population of the United States. The NHIS is a cross-sectional household interview sample survey conducted annually since 1957. Interviews are conducted each week throughout the year. The survey follows a multistage probability sampling of households. It is designed in such a way that the sample scheduled for each week is representative of the target population and the weekly samples are additive over time. The NHIS redesign in 1985 included a reduction of the primary sam- pling units (PSUs) from 376 to 198 for sampling efficiency and an oversam- pling of the black population. Data are collected from approximately 50,000 households including about 130,000 persons in a calendar year. The survey provides national estimates on the incidence of acute illness and accidental injuries, the prevalence of chronic conditions and impair- ments, the extent of disability, the use of medical services, and other health- related topics by demographic and socioeconomic characteristics. The questions in the core part of the questionnaire are asked each year. They solicit information to measure illness and injuries, days of disability including bed-days, work-loss days, school-loss days, and other days lost due to acute or chronic conditions. Included in the core are questions on limitation of activity caused by chronic conditions or impairments, hospital- izations, doctor contacts, perceived health status, as well as the social, eco- nomic, and demographic characteristics of the interviewed person. Ques- tions on dental visits were also part of the core from 1975 through 1981. The core questionnaire includes six lists of chronic conditions. Each condition list concentrates on a group of chronic conditions involving a specific system of the body, such as digestive, skin and musculoskeletal, circulatory, and respiratory. Prior to 1978, questions on only one condition list were asked each year. Beginning in 1978, each of six representative subsamples has been asked questions based on one of the six lists of chronic conditions. In this manner national estimates on each of the six body systems are obtained during the same interview year. Questions on current health topics are included in supplements to the core questionnaire and change from year to year. They have been included in the NHIS since 1959, permitting the survey to collect information on a wide variety of health topics and to be responsive to the country's changing information needs. Because a supplement is added to the ongoing NHIS, which is an established vehicle resulting in statistically reliable data for the U.S. population, it is an efficient way to measure on a large scale and in a timely manner a topic of current concern. Some supplements, such as health insurance and out-of-pocket health expenditures, are repeated period-
FEDERAL HEALTH DATA SOURCES 131 ically to allow measurement of change over time. For instance, supple- ments in 1979 were included on immunization, smoking, home health care, eye care, residential mobility, and retirement income. In 1980, the supple- ments on home health care, residential mobility, and retirement income were continued and a supplement on health insurance was added. In 1981, a child health supplement was included throughout the year. The 1982 NHIS included supplements on health insurance and preventive care. The 1983 NHIS contained questions on alcohol, dental care, physician services, and health insurance. In 1984 the supplement was devoted to health issues of aging; in 1985, health promotion and disease prevention; in 1986, health insurance, dental health, vitamin and' mineral intake, longest job worked, and functional limitations; in 1987, cancer risk factors and child adoption; in 1989, mental health. Because of budgetary constraints, beginning in 1985 the supplements have been funded by the agencies requesting the . special topics. With acquired immunodeficiency syndrome (AIDS) reaching epidemic proportions in the United States, and with no cure or vaccine to prevent infection by the virus at this time, a supplement on AIDS knowledge and attitudes was initiated as part of the 1987 NHIS in order to determine the level of knowledge of the U.S. population and to help assess the effective- ness of educational campaigns. This supplement with necessary modifica- tions has been continued throughout 1988, 1989, 1990, and 1991 and will provide continued measurement of the public's knowledge and attitudes about AIDS and changes in these characteristics over time. ~ . ~ ~ ~ ~ _ ~ TO AT TV National Health and Nutrition Examination Survey The National Health Survey Act authorized special studies in addition to the continuing household interview survey. Furthermore, the recommen- dations of the National Committee on Vital and Health Statistics on the subject had also envisioned special studies to include one "to obtain data on undiagnosed and non-manifest diseases" (NCHS, Ser.1, No. 4, 1973~. The first Health Examination Survey (HES) was launched in 1959 and was com- pleted in 1962. A traveling staff and mobile examination units were em- nloYed for the survey. Data were collected through direct standardized 1~ _1 ___ 1~~ ~~_~~ BAND LEA ~^qel~r~m~ntc An physical examinations, clinical and laboratory LUStS, ~~U Ill~a~uiclll~l~L~ vie the total prevalence of certain chronic diseases as well as the distributions of various physical and physiological measures, including blood pressure and serum cholesterol levels. year cycles through the 1 960s. A decade later the survey was restructured and expanded to include measurements of nutritional status of the population and subsequent moni- toring of changes in that status over time. This change reflected the coun- The survey was conducted at two- to four-
32 APPENDIX C try's and the federal government's shift of emphasis from illness to health maintenance. The survey name was changed to the National Health and Nutrition Examination Survey (NHANES). The first cycle of this expanded survey, NHANES, was initiated in 1970 with data collection beginning in 1971. To date, five nationwide surveys using health examination procedures have been completed since the inception of this survey program in 1960. In 1982, the survey was directed toward persons of Hispanic origin or descent and was conducted in areas with high concentration of Hispanic population. At the present time, NHANES III is in the field. As with the earlier sur- veys, NfIANES III involves interviews, physical examinations, and bio- chemical and hematological tests of a representative sample of the U.S. population. The objectives of both the early HES and the NHANES have been collection of data that can be obtained best or only by direct physical exam- ination, clinical and laboratory tests, and related measurement procedures. Prevalence data are collected for specifically defined diseases or conditions of ill health and normative health-related measurement data are collected that show distributions of the total population with respect to particular parameters, such as blood pressure, visual acuity, and serum cholesterol level. The sample design for all the HES and NHANES cycles has been multi- stage, highly clustered probability samples to provide representative nation- al estimates. All of the samples are stratified by broad geographical regions and by population density grouping. At the present time, NHANES III is in the field. As with the earlier surveys, NHANES III will involve interviews, physical examinations, and biochemical and hematological tests of a representative sample of the U.S. population. Health Records Survey Program Soon after the establishment of the National Health Interview Survey and the Health Examination Survey, it became obvious that information on hospitalized illness and injury, out-of-hospital medically attended illness, long-term institutionalized conditions, and data on the use of providers of services and facilities were either not available on a national basis or were inadequate. As a consequence, by 1960, when the National Health Survey and the National Office of Vital Statistics were combined to form the National Center for Health Statistics, developmental work already had begun for surveys based on records of health facilities and providers of health services. The Health Records Survey (HRS) program, was conceived originally
FEDERAL HEALTH DATA SOURCES 133 as one of the three major data collection program activities of the National Health Survey. The surveys under this program were designed to augment the NHIS and NHANES by providing comprehensive statistics on the health of the institutionalized population as well as on the use of all types of health facilities and services. As stated by NCHS at the time of development of the HRS program, an important distinction between the Health Records Survey program and the other programs of the National Health Survey was that the data are obtained from the provider or facility providing service instead of from the recipient of the service. The rationale for the reliance on the provider, as stated by NCHS, was that providers can have the most accurate and detailed data on diagnosis and treatment, and that providers are a cost-effective source for identifying events such as hospitalization, surgery, and long-term institu- tionalization, which are relatively rare events in the total population. The original concept of the HRS encompassed records from all types of medical facilities and institutions, including those from such facilities as hospitals, clinics, dispensaries, physicians' and dentists' offices, and nurs- ing homes. In practice, however, the data are limited to that obtained in the four current surveys: National Hospital Discharge Survey, National Ambu- latory Medical Care Survey, National Nursing Home Survey, and the Na- tional Master Facility Inventory. The major part of the data collected is based on existing records. Mechanisms used to obtain the data include abstracts of medical records of institutions, completion of patient encounter forms by physicians, compilation of data from states and professional associations, and purchase of data from private abstract services. The four component surveys of the HRS, as they have evolved over the years, have provided data for monitoring the supply, organization, and use of health care in specified service setting categories. It is important to note here that these surveys provide data on events or on the characteristics of the provider in a given category. In the 1970s the term Health Records Survey Program was dropped; these surveys began to be referred to as provider surveys and form the basis of the proposed National Health Care Survey, the subject of this study. The four current records, or provider surveys, are described briefly below. National Master Facility Inventory. As a first order of business to implement the records-based survey program, NCHS developed a current unduplicated list of all facilities for inpatient care in the United States. The National Master Facility Inventory (NMFI) is this list. The NMFI has served as the universe or sampling frame for the health records surveys. Conceptually it includes all types of facilities within scope of the HRS program, with the flexibility to include other types of establishments or
34 APPENDIX C health facilities as the program expands. It also serves as a source of statis- tics on the number, type, and distribution of inpatient facilities in the Unit- ed States. The program of the NMFI includes the development and maintenance of a list of names and addresses of all facilities or establishments within its scope and the collection of information that describes them with respect to their size, type, and current status of business. The information not only provides a basis for stratifying the NMFI into homogenous groups for the purpose of sampling designs, but also provides important national statistics about the availability of such facilities nationwide. The scope of the inventory in its initial stage of development included all hospitals in the United States with 6 or more beds, as well as all resident institutions, except for nursing or personal care homes with less than 3 beds. Institutions included in the NMFI were establishments that were in business to provide medical, nursing, personal, or custodial care to groups of unrelated individuals who have no usual place of residence elsewhere. Such as: (1) Establishments that provided nursing care to the aged, infirm, or chronically ill. These included places commonly referred to as nursing homes, homes for the aged, rest homes, boarding homes for the aged, and homes for the needy such as almshouses, county homes, and poor farms. The primary factor that determined if an establish- ment was in scope was its function of providing some kind of care besides room and board. Residential schools or homes for the deaf. (3) Residential schools or homes for the blind. (4) Homes for unwed mothers. (5) Orphan asylums and homes for dependent children. (6) Homes for crippled children. (7) Homes for incurables. (8) Residential schools or detention homes for juvenile delinquents. (9) Prisons, reformatories, and penitentiaries that are operated by the federal or state governments. The NMFI did not include special dwelling places or group; darters such as hotels, private residential clubs, fraternity or sorority houses, mon- asteries, nurse's homes, flophouses, or labor camps.
FEDERAL HEALTH DATA SOURCES Data items collected in the NMFI included: . Ownership of the institution, · Type of services provided, Number of beds, Number and kinds of staff, Patient census, Annual number of discharges, and Annual number of admissions. 135 The NMFI has been conducted periodically since 1963. The inventory has been kept current by the periodic addition of names and addresses obtained from state licensing and other agencies for all newly established inpatient facilities. In addition, annual surveys of hospitals and periodic surveys of nursing homes and other facilities are used to update name and location, type of business, number of beds, and number of residents or patients in the facilities and to identify those facilities that have gone out of business. In the 1970s, all data were obtained through the state statistical agen- cies participating in the Cooperative Health Statistics System (CHSS). Most of these states, however, had data only on nursing and related care homes. Some of the CHHS states collected data only for nursing homes; NCHS surveyed the other related care homes directly for these states. Hence the scope was limited to data on hospitals and nursing and related care homes. After the elimination of CHHS, NCHS did not go back to collecting data directly from hospitals. Hospital data are purchased from the American Hospital Association and SMG Marketing Group, Inc. In the 1980s the NMFI was once again centralized when the Inventory of Long Term Care Providers was conducted in 1986 (information based on telephone conver- sations with NCHS staff). The scope of the 1991 NMFI, renamed the National Health Provider Inventory (NHPI), has been expanded to include information about facilities providing health care, including hospitals, skilled nursing and long-term care units of hospitals, nursing and related care homes, facilities that pro- vide some kind of personal care, board and care or domiciliary care (such as residential care homes, group homes, homes for the aged, family care homes, adult foster care homes, personal care homes, adult congregate living facil- ities, residential community care facilities, domiciliary care homes), homes for mentally retarded or developmentally disabled people, homes for the mentally ill (such as halfway houses, residential centers for emotionally
36 APPENDIX C disturbed children, community residences, and supervised apartments, shel- ters, and'hostels), halfway or quarter-way houses for alcohol or drug abus- ers, and hospices and home health agencies. Information will be obtained on the certification status under Medicaid or Medicare of the nursing facili- ty, if it is certified under Medicaid as an intermediate care facility for the mentally retarded, and if it is a nursing home unit of a retirement center, as well as the race, sex, and age of residents. The 1991 NHPI will be conduct- ed centrally by the Bureau of the Census as the collecting agent for the NCHS. National Hospital Discharge Survey. The National Hospital Discharge Survey (NHDS) is a continuing nationwide sample survey of short-stay hospitals in the United States. It has been the principal source of data since 1965 on inpatient use of hospitals by characteristics of patients, expected payment source, length of stay, diagnoses, surgical and nonsurgical proce- dures, and the patterns of use of care in hospitals of different bed size, ownership type, and geographic region. The scope of NHDS is limited to discharges from nonfederal hospitals located in the 50 states and the District of Columbia. Only hospitals having 6 or more beds for patient use and an average length of stay for all patients of less than 30 days or those whose specialty is general (medical or surgi- cal), maternity, or children's general are included in the sample. The NHDS was initiated in fall 1964, the short-stay hospitals listed in the NMFI as of 1964 serving as the sampling frame. The sample was updated periodically with samples of hospitals that opened later. Within each sample hospital, a systematic sample of discharges was selected. Staning in 1985, in addition to manual selection of the sample and transcription of data, NCHS purchased from commercial abstracting services discharge medical abstracts 'and selected samples from these tapes. Until 1987 a two-stage stratified sample design was used and the sam- ple was selected from a frame of short-stay hospitals listed in the National Master Facility Inventory. i': 1988 when the phased implementation of the National Health Care Survey began, the NHDS was redesigned to incorpo- rate an integrated cluster sample design. A subsample of the NHIS PSUs serves as the sampling frame for NHDS. The redesigned survey includes a three-stage clustered stratified design the primary sampling units (PSUs), the hospitals within the PSUs, and discharges within hospitals. The rede- signed sample contains 542 hospitals, approximately one-third of the sam- ple hospitals providing data through commercial abstract services In order to ensure continuity of the survey, should a problem arise with tic.- purchase of these data and to provide narrative information for special studies, the design also includes a subsample of 128 hospitals that continue to submit data through the traditional hard copy method of data collection with the
FEDERAL HEALTH DATA SOURCES 137 Bureau of the Census or hospital personnel completing the medical ab- stracts. The data obtained from the medical abstract form and the abstract ser- vice data tapes include: Date of birth Sex Race Ethnicity Marital status Final diagnoses Surgical and nonsurgical procedures and dates of procedures Expected sources of payment · Medical record number Admission and discharge dates Discharge status and disposition Residence zip code The medical record number and patient zip code are confidential infor- mation and are not available to the public. Up to seven diagnoses and four procedures are coded for each discharge. National Ambulatory Medical Care Survey. In the early 1970s the recognition that there was little known on a national basis about medical care provided in physicians' offices that was comparable to information available for hospital care and nursing home care, led to the establishment of the National Ambulatory Medical Care Survey (NAMCS). NAMCS, first fielded in 1973, was designed as a continuing national sample survey to meet the needs and demands for statistical information about the provi- sion of ambulatory medical care services in the United States. Today, however, ambulatory health services are provided in a variety of settings including physicians' offices, health maintenance organizations, and hospi- tal outpatient clinics. Yet NAMCS still is limited to visits to "office-based, patient care, non-federal physicians." As currently designed, it does not include visits to hospital-based physicians; visits to anesthesiologists, pa- thologists, and radiologists; and visits to physicians who are principally engaged in teaching, research, or administration. Telephone contacts with physicians are also excluded.
138 APPENDIX C The NAMCS was conducted continuously from 1973 through 1981. From 1982 through 1984 the survey was suspended due to budget con- straints. The survey was conducted again in 1985. In 1989 survey activi- ties were resumed again on a continuous basis as part of the National Health Care Survey. During 1973-1985 a multistage probability sample design was used in- volving samples of PSUs, physician practices within PSUs, and patient vis- its within physician practices. A modified probability-proportional-to-size sampling procedure using separate sampling frames for metropolitan statis- tical areas (MSAs) and for nonmetropolitan counties was used to select the first stage sample PSUs independently within each of the four census re- gions. The second stage consisted of a probability sample of nonfederal, office-based physicians selected from master files maintained by the Amer- ican Medical Association and the American Osteopathic Association. The third stage is the selection of patient visits within the annual practices of the sample physicians. The 1985 sample included 5,032 physicians and 71,594 patient visits. In 1989, when the survey was included as a component of the National Health Care Survey, a subsample of the NHIS PSUs was used as the first stage of sample selection. The sample size for the 1989 and 1990 cycles has been reduced by one-half of the 1985 sample size to include 2,500 physicians in office-based practice, with an expected sample of 45,000 pa- tient visits annually. Also, beginning with 1989, data for the survey are being collected annually under an interagency agreement with the Bureau of the Census. Data items for the 1989 and 1990 surveys have remained constant to allow for aggregation over a two-year period. Data are collected on the characteristics of patients, their diagnoses and symptoms, and on the diagnostic procedures, therapy, drugs Prescribed, and payment source associated with each visit to the physician. characteristics of the physician, such as specialty and type of practice. These data are obtained by means of the induction interview form and patient log and record form. Information is obtained on the physician's specialty, type of practice, and staff. In 1989 questions were included about HMOs and other prepaid programs, on the availability of in-office laborato- ry testing to provide data on the number of physician offices having in- office laboratory equipment, the nature of in-office tests performed, the degree of quality control exercised in the performance of the tests, and the physicians' knowledge of state testing regulations. For the 1989-1990 NAMCS, a patient record is completed for each sampled patient visit and contains the following items: . Patient zip code (new in 1989) · Date of visit
FEDERAL HEALTH DATA SOURCES Date of birth Sex Race Ethnicity Expected sourcets) of payment New patient · Referral . Patient complaint, symptom or reason for visit · Physiciants diagnoses . 139 Has patient been seen before Screening/diagnostic services includes changes from the 1985 cat- egories, nearly all of which are tests for screening and early detec- tion of disease. Medication therapy Nonmedication therapyincludes changes from the 1985 catego- ries. New categories include consultation concerning weight re- duction, cholesterol reduction, smoking cessation, family planning, and breast self-examination instruction. Disposition Duration of visit NCHS plans call for expanding the scope of this survey to cover hospi- tal outpatient aIld emergency departments beginning in late 1991. National Nursing Home Survey. The National Nursing Home Survey (NNHS) has been conducted periodically since 1963 and most recently in 1985. Data are collected on a wide range of information about facilities, services provided, staff, financial characteristics, and on residents' personal and health characteristics. In the 1960s, two surveys of nursing homes were conducted on an ad hoc basis. The need for comprehensive data on a continuing basis became obvious, with the establishment of the Medicaid and Medicare program, increasing use of nursing homes by the elderly population whose numbers were growing, and the overall increasing involvement of the federal govern- ment in these and other long-term care issues. In 1973 the National Nurs- ing Home Survey was established in NCHS as a continuing program. The Master Facility Inventory has served as the sampling frame for the surveys.
140 APPENDIX C Subsequent to the two ad hoc surveys in the 1960s, three national nurs- ing home surveys have been conducted. The first national survey was conducted in 1973-1974, the second was conducted in 1977, and the third was conducted in 1985. The next survey is scheduled for 1995. The 1985 NNfIS was a sample survey of nursing and related care homes, their residents, discharges, and staff in the conterminous United States that regularly provide nursing or personal care services to residents, regardless of the level of service provision or the certification status of the facility. Facilities were either free-standing establishments or nursing care units of hospitals, retirement centers, or similar institutions maintaining financial and employee records separate from those of the larger institutions. Resi- dential care facilities were excluded from the scope of the survey, as were those serving only persons with specific health problems, such as mental retardation or alcoholism. A stratified two-stage probability sample design was used, the first stage being selection of facilities and the second stage being selection of residents, discharges, and registered nurses from the sample facilities. Data were collected on four major areas: (1) Facility. Inflation was obtained by personal interview with the administrator and by mail-back questionnaires. Topics covered in- cluded: (1) size, ownership, license and Medicare and Medicaid certification status, number of beds, services offered; (2) staffing patterns including characteristics of registered nurses, work sched- ule, experience, activities in facility, demographic characteristics, and salary; (3) costs, including total expenses and for major com- ponents of operation. (2) Current residents. Topics covered included: demographic charac- teristics, activities of daily living, status and living arrangements prior to admission, conditions at admission and at interview, re- ceipt of services, cognitive and emotional status, charges, sources of payment, history of nursing home utilization, and hospitalization during stay. (3) Discharged residents. Topics covered included: demographic char- acteristics, history of nursing home utilization, hospitalization dur- ing stay, status at discharge, conditions at admission and discharge, sources of payment. Resident information was obtained by personal interview with the nurse who referred to the medical record. (4) Follow-up. Follow-up information on the two patient samples was collected by means of a computer-assisted telephone interview with a next of kin of the current or discharged resident. The next of kin
FEDERAL HEALTH DATA SOURCES 141 interviewed was identified in the current resident and discharged resident questionnaires and included relatives, guardians, and any- one familiar with the sampled resident. Topics covered included: (1) subject's living arrangements, health, and functional status during the period preceding the admission, (2) history of previous nursing home use activities of daily living, (3) current status and living arrangements, and (4) Medicaid spend- down. In an attempt to meet the increasing demand for information related to the dynamics of long-term care use, follow-up data were collected in the 1985 NNfIS covering a period of one year subsequent to the initial nursing home admission. Since many episodes of alternating institutional, acute hospital, and community care are of considerably longer duration, it became clear that the period of observation needed to be extended. Moreover, data compiled on an event basis do not reveal the interconnections within epi- sodes of illness or utilization or provide information on transitions between hospital, nursing home, and community care. In response to this need, the National Nursing Home Survey Follow-up (NNHS follow-up) was estab- lished as a collaborative project between the NCHS and the National Insti- tute on Aging. The NNHS follow-up is a longitudinal that follows a cohort of surviv- ina current residents and discharged residents sampled from the 1985 NNHS. The primary purpose of the follow-up study is to provide data on the flow of persons in and out of long-term care facilities and hospitals. The study obtains lon;,itudinal information not readily available in the medical record, such as the resident's health and functional status prior to admission, the reason for admission, and a history of previous nursing home admissions, the resident's current functional status, current living arrangements, and nursing home and hospital use subsequent to the last contact and sources of payment. The NNHS follow-up has initiated three additional contacts with the cohort since the 1985 NNHS. Interviews for all three waves were conduct- ed using a computer-assisted telephone interview (CATI) system. Data collection and data entry are concurrent and computer-controlled. The first wave of the follow-up was conducted between August and December 1987. The study population consisted of all sampled residents who were still alive, approximately 6,600 subjects. If the subject could not participate, interviews with proxy respondents (primarily next of kin) were conducted. For subjects residing in facilities, administrators or other facili- ty personnel were interviewed. Wave II of the NNHS follow-up was conducted from July to November 1988, approximately 12 months after the completion of the first wave, and
142 APPENDIX C involved approximately 4,000 subjects. The same questionnaire and the CATI system were used for both waves. Respondents were asked about the subject's vital status, living arrangements, and nursing home stays, hospital stays, and sources of payment for stays occurring between the wave I and wave II interviews. For deceased residents the information on cause of death was obtained through the National Death Index. Interviewing for wave III of the follow-up began the first week of February 1990 and was completed in April 1990. Wave III extends the period of observation for surviving residents in the 1985 NNHS for an additional 18 months and involves approximately 3,000 subjects. The same questionnaire, CATI system, and procedures used in the previous two waves were again applied to wave III. In addition, seven new questions are asked at the end of the interview concerning the disposition of the subject's own home. Other Related Health Surveys of the Department of Health and Human Services As stated earlier, many other major data systems are maintained in the other DHHS agencies to meet their special purpose and programmatic needs for statistical information. Some of the major potentially related data sys- tems are briefly described below. Centers for Disease Control The Centers for Disease Control (CDC) is the federal agency within the Public Health Service charged with protecting the public health of the na- tion by providing leadership and direction in the prevention and control of diseases and other preventable conditions and responding to public health emergencies. It is composed of nine major components, one of which is the National Center for Health Statistics. Behavioral Risk Factor Surveillance System. This surveillance system provides official state health agencies with the funding, training, and con- sultation necessary to permit them to routinely collect behalf Oral risk factor information. The data items include weight control, hypertension, physical activity, obesity, mammography, alcohol consumption, seatbelt use, tobacco use, HIV/AIDS, preventive health practices, and routine demographic infor- mation such as age, sex, race, and education. In all, 46 states and jurisdictions participate in this system, which was created in the early 1980s. Each of the states conduct surveys of their noninstitutionalized adult population during the course of the year with 600-3,000 respondents interviewed in each state yearly. Telephone inter-
FEDERAL HEALTH DATA SOURCES 143 viewing is employed, and states can add modules of questions to meet special needs. States use the data from this system to develop and evaluate programs and to track progress in reducing behavioral risk factors over time. Agency for Health Care Policy and Research The Agency for Health Care Policy and Research was established in 1989 by the Omnibus Budget Reconciliation Act (42 U.S.C. 299) as the successor to the National Center for Health Services Research and Health Care Technology Assessment. The agency is the federal government's fo- cal point for health services research. National Medical Expenditure Surrey. The 1987 National Medical Expenditure Survey (NMES), conducted by the Agency for Health Care Poli- cy and Research, continues a series of national health care expenditures sur- veys. The last two were conducted in 1980 and 1977. The primary purpose of NMES is to provide the executive branch, the Department of Health and Human Services, and the Congress with national estimates of the use of and expenditures for health care services and of health insurance coverage. The 1987 NMES is comprised of three components: (1) the Household Survey, (2) an institutional population component that includes the popula- tion resident in or admitted to nursing homes and facilities for the mentally retarded over the course of the survey year, and (3) the Survey of American Olin and Alaska Natives (SATAN), which includes all persons eligible for care through the Indian Health Service and living on or near reserva- tions. Because of the importance of long-term care for the nation's health policy agenda, NMES also provides national estimates on long-term care services and expenditures. Taken together, the three mayor components of NMES provide com- prehensive, population-based measures of health and functional status, es- timates of insurance coverage and ot the use of services, expenditures, employment, income, and sources of payment for the calendar year 1987. The long-term care supplement provides estimates of persons with func- tional disabilities and impairments and their use of formal home and com- munity services, as well as the burden felt by family and friends of pro- . . Yang care. Although these periodic expenditure surveys provide the most compre- hensive data on health expenditures and insurance coverage, they are con- ~ , ~ ducted at infrequent intervals at great cost and are slow to produce accessi- ble data. Medicare Beneficiary Health Status Registry. A Medicare Beneficiary Health Status Registry (MBHSR) is being developed as a collaborative ef-
144 APPENDIX C fort between the Agency for Health Care Policy and Research and the Health Care Financing Administration. The registry is funded by AHCPR and will be implemented by HCFA. It will be a longitudinal database containing information on elderly Medicare beneficiaries obtained by linking health status information from survey data with HCFA's administrative enrollment and utilization files. The survey will gather data on approximately 2 per- cent of the elderly at the time they enter the Medicare program and at intervals of two to five years thereafter. Data will be gathered of each new Medicare cohort on risk factors, functional status, sociodemographic vari- ables, medical history, and quality of life, thus permitting a projection of their needs. By enrolling successive cohorts over many years, changes in the health and utilization of new entering groups can be monitored over time. Through repeated contacts with the same individuals while they are Medicare beneficiaries, the progression of health and disease can be mea- sured, the relationship between the Medicare program and the health status of its beneficiaries can be examined, and the effectiveness of specific health interventions in influencing the health status of individuals of the group can be evaluated. The primary goals of the registry are to provide information on measur- ing the relationship of Medicare-reimbursed services to the health status of Medicare beneficiaries; to describe, analyze, and understand health and dis- ease longitudinally in successive cohorts of Medicare beneficiaries; to de- scribe, analyze, and understand the use and costs of services, lor~g-term utilization patterns, and lifetime Medicare costs of individual coh`~;~s: to monitor access to care in special populations; to assess the effectix Bless and impact of specific medical and surgical interventions on the health, quality of life, well- -a. and functional status of Medicare beneficiaries; and to monitor existing, DlIHS "Healthy People 2000" objectives and devel- op additional ones. Design and field testing began in September 1990. Implementation is expected in fiscal 1993. Agency for Alcohol, Drug Abuse, and Mental Health Administration The Agency for Ale- - ~l, Drug Abuse, and Mental Health Administra- tion (ADAMHA) province a national focus for the federal effort to increase knowledge and to promote effective strategies in handling health problems and issues associated with the use and abuse of alcohol and drugs and with mental illness and mental health. Its responsibilities include gathering and analyzing data about the extent of alcohol, drug abuse, and mental health problems and the national response to these needs in terms of planning, estab- lishing, and evaluating alcoholism, drug abuse, and mental health programs.
FEDERAL HEALTH DATA SOURCES . 145 Mental Health Statistics. The National Institute of Mental Health (NIMH) in the Alcohol, Drug Abuse, and Mental Health Administration has been collecting data on inpatient care of the mentally ill in public and private mental hospitals and in psychiatric wards of general hospitals since 1947, when it took over that function from the Bureau of the Census, which had been collecting information on hospitalization of the mentally ill since 1840. Recognizing that the annual census alone could not provide the range of basic data required on the facilities providing, inpatient mental health care, over the years NIMB has taken several steps to correct these shortcomings and, as needs arose, added other types of care settings to the data collection effort and instituted a sample survey program to meet specialized needs. Today NIMH conducts several inventories of mental health organiza- tions. These data systems are based on questionnaires mailed every other year to mental health organizations in the United States, including psychiat- ric hospitals, psychiatric services in nonfederal general hospitals, Veterans Administration psychiatric services, residential treatment centers for emo- tionally disturbed children, freestanding outpatient psychiatric clinics, par- tial care organizations, and freestanding, and multiple care mental health organizations not elsewhere classified. Information is collected on types of services provided (inpatient, outpa- tient, and partial care); number of inpatient beds; number of inpatient, out- patient, and partial care additions; average daily and end-of-year inpatient census; patient characteristics; staffing by types of discipline; expenditures; and revenue by source (1983 and 1986 only). Staffing information is col- lected as of a sample week at the time the inventory is mailed, and informa- tion on types of services and beds are collected as of the beginning, of the next year. The sample survey program was initiated in 1969 and today includes ~A~;C`c,;~e tm ~tOtP rn''ntv and private mental hospitals, outpatient psychi- atric services, and Veterans Administration psychiatric services. The pur- pose of these surveys is to obtain data on the sociodemographic, clinical, and treatment characteristics of patients served by these facilities. In gener- aulllla;~o He ~-~~ ~ .~ ~ ~ rat . al these surveys have been conducted approximately every five years for any specific type of facility. Data are collected on a sample of admissions from a national sample of facilities for one month in the year on age, sex, race, marital status, education, prior psychiatric care, type of services re- ceived, and referral on discharge. The NCHS historically has not gathered data on mental health facilities for their National Master Facility Inventory (NMFI). In the late 1960s, when the expanded and consolidated inventory of mental health organiza- tions was initiated by NIMH, the two agencies reached agreement that NCHS would not collect data on mental health facilities for the NMFI, and NIMH would provide the data collected in their inventory to NCHS for incorpora-
146 . APPENDIX C lion in the NMFI. These data were then added to the NMFI and reported in the category "other health facilities." NCHS also has not collected data on the mentally ill on a regular basis in the general health care sector in the same manner that it collects informa- tion on general health care. Some information, however, is routinely col- lected, and other information has been collected on an ad hoc basis if re- sources are made available by an interested agency. For example: The National Hospital Discharge Survey collects information on the mentally ill admitted to short-stay, nonfederal hospitals. The National Ambulatory Medical Care Survey collects informa- tion on the mentally ill seen in physicians' offices, including psy- chiatrists' offices. In the 1991 NAMCS a separate item on depres- sion is included in the list of diagnoses instead of relying on physicians to report it. The National Nursing Home Survey collects information on mental illness in nursing home residents. The National Health and Nutrition Examination Surveys have col- lected depression data. The NHANES-III, currently in the field, is administering a depression and a manic-depressive instrument. The 1989 National Health Interview Survey included a supplement on mental health in collaboration with, and funded by, NIMH. The overall purpose of this supplement was to estimate the number and characteristics of chronically mentally ill persons in the civilian, noninstitutionalized population of the United States. Chronic men- tal illness is defined by NIMH as "severe and persistent disability resulting primarily from mental illness." It was not the intent of this supplement to estimate the prevalence of any particular disor- der. Alcohol and Drug Abuse Statistics. ADAMHA has instituted several studies to obtain data on the prevalence and trends in alcohol and drug abuse and the treatment facilities. An illustrative few are described below: The National Household Survey of Drug Abuse is a biennial, national sample survey of household population age 12 and older. The survey pro- vides estimates of the lifetime, last year, and current use of marijuana, cigarettes, and alcohol and attitudes about drug abuse. Youths ages 12-17 and young adults ages 18-25 are oversampled, as are blacks and Hispanics. The 1988 survey is the ninth in a series that began in 1971 and is based on personal interviews of a sample of 8,814 persons age 12 and over.
FEDERAL HEALTH DATA SOURCES 147 The National Drug And Alcoholism Treatment Unit Survey (NDA- TUS) is a national survey of the scope and use of treatment and prevention services and resources, both privately and publicly funded. This survey is conducted periodically by the National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism. Information is ob- tained from over 12,000 treatment, prevention, and service units on types of services provided, client capacity and census, and funding amounts, as well as client demographic information. Through this survey a National Directo- ry of Drug Abuse and Alcoholism Treatment and Prevention Programs is produced. The Sample Survey of NDATUS Clients will be a sample survey of clients drawn from a representative sample of the drug abuse and alcohol treatment units. The study will provide for the first time data on informa- tion on the characteristics of alcohol and drug abuse treatment programs and on types and duration of treatment and collateral client-focused activi- ties. Client characteristics will include previous treatment history, diagnos- tic data about drug and alcohol use, amount, patterns, and related impair- ments. Clients will be reinterviewed at 3 and 15 months to determine subsequent treatment and changes In use and impairments. Prevalence and Impact of Alcohol and Drue Abuse Dependence in Short-term General Hospitals wait be a one-time study to determine the national prevalence rates of alcohol abuse and alcoholism and polysub- stance abuse in short-term, general, nonfederal hospitals; it will also deter- mine the impact of alcohol and other drug abuse and dependencies on hos- pital resource utilization, charges, and costs. Drug Abuse Warning Network (DAWN) is a large-scale ongoing drug abuse data collection system based on information from a nonrandom sam- cle of emergency room and medical examiner facilities. DAWN collects informntion only about those drug abuse occurrences that have resulted in a _ .. ^, _, ~ _ , . . .. ~ _ ., medical crisis or death. Major objectives include the mon~tonng OT patterns and trends of drug abuse, identification of substances associated with drug abuse episodes, and the assessment of drug-related consequences and other health hazards. Health Care Financing Administration The Health Care Financing Administration (HCFA) was established in 1977, placing under one administration the oversight of the Medicare and Medicaid programs and related federal medical care quality control staffs. HCFA maintains a comprehensive system of administrative data on Medicare
148 APPENDIX C enrollment, use, diagnosis, provider characteristics, costs, charges, and pay- ments based on the claims payment files. Administrative data systems, of course, are a by-product of program administration. Although they have many potential uses, they cannot tell us, for example, about services cur- rently not covered by Medicare but that may be likely targets of new legis- lation, or about sources of payment other than Medicare, people who do not use Medicare services, or characteristics of beneficiaries that determine their use of services and costs. HCFA has taken several steps to enhance the usefulness of its data by developing systems to obtain information about the health status and socio- demographic variables of the beneficiaries that can be linked to the admin- istrative files for analysis. Some of these databases are described below. Medicare Current Beneficiary Survey. The Medicare Current Benefi- ciary Survey (MCBS), developed by the Health Care Financing Administra- tion, is designed to be an ongoing multipurpose survey for use by all com- ponents of HCFA, the department, and others concerned with Medicare policy. The MCBS is focused on health care use and expenditure in the Medicare population and determinants thereof, issues that are of prime im- portance to HCFA. The core of the MCBS is a series of interviews of a representative sample of the Medicare population to obtain information on their patterns of health service use and cost over time, sources of coverage and payment, assets and income, demographic characteristics, health and functional sta- tus, health and work history, and family supports. The same people will be interviewed three times a year over several years to observe changes in use with changes in coverage, and to observe processes that occur over time, such as institutionalization or spending down of assets. The survey will produce data sets suitable for both longitudinal and cross-sectional analysis. In addition to the core items, periodic or one-time supplements will be included in each of the three annual rounds. These will collect information on relatively stable characteristics of the respondents such as work history, or on special topics of current policy interest to BCFA, such as access to care or perceptions about health maintenance organizations. They may also contain questions that need not be asked each time but may be asked annu- ally, such as income and assets. The MCBS sample will consist of 12,000 individuals sampled from the Medicare Enrollment File to be representative of the Medicare population as a whole and by age group, enrollment type (aged or disabled), urban or rural residence, ant among the aced, whether or not institutionalized. The sample will be adjusted annually for attrition, as well as for newly eligible persons. Survey data files will be matched to HCFA claims payment and other administrative records such as the National Death Index, Social Security
FEDERAL HEALTH DATA SOURCES 149 records, and the Area Resource File. Pilot testing of all forms and proce- dures has been completed and data collection is under way. Nursing Home Quality Assessment. The Omnibus Reconciliation Act of 1987 requires that each resident of a Medicare or Medicaid certified nursing facility have a periodic standardized, accurate, and reproducible comprehensive assessment of his or her functional capacity using a Resi- debt Assessment Instrument (RAI) specified by the state. In addition, the instrument must include a uniform minimum data set of core elements, common definitions, and utilization guidelines specified by HCFA. According to HCFA plans, implementation will be phased in so that initially the RAI will be administered only for new admissions. Reassess- ments will be conducted when significant changes in the patient's status have occurred. The database generated from the RAI over time has the potential for being used for research purposes in addition to the administra- tive uses for which the instrument was originally developed, which raises concerns about the quality and reliability of the data. Medicare Provider Analysis and ReviewMedicare Stay Records. An Inpatient Hospital Stay Record containing information about a person's stay in a hospital from admission through discharge is created for 100 per- cent of Medicare enrollees. Stay records are used to compile many health insurance utilization tables and the Medicare Provider Analysis and Re- view (MEDPAR) files. This record contains use information from bills, demographic information about the enrollee, and Provider characteristics data. Stay records are also prepared for stays in skilled nursing facilities (SNF). The SNF stay record is essentially the same as the hospital record. Part B Medicare Annual Data System. The Part B Medicare Annual Data System (BMAD) system collects information in each carrier's history of claims processed. The data are produced and submitted from the Medi- care claims history for the preceding year and are submitted annually. The Part B data are provided in four separate files: (1) The Procedures File- provides an array of every procedure processed showing the relative fre- quency, submitted charge, and paid amounts. This file dives HCFA detail on all the services processed by Medicare carriers. (2) The Prevailing Charge File supplies HCFA with the prevailing charge limits for every procedure shown on a carrier's procedure file submittal. It allows HCFA to study and accurately project payment levels. (3) The Provider Filecon- tains line-by-line detail from claims history of procedures rendered by a 1- percent sample of physician suppliers. These data allow HCFA to study the impact of actual and physicians' projected program changes on the suppli- ers. (4) The Beneficiary Filethis is a 5-percent beneficiary history sam- ple file. It contains the detailed claims history of procedures received by
150 APPENDIX C the sample beneficiaries. This file also collects the claims detail for all End Stage Renal Disease (ESRD) beneficiaries. This file allows HCFA to link, based on the health insurance claims number, a beneficiary's Part A and Part B service utilization data. CONCLUSION The brief descriptions presented above of only some of the national health data systems maintained by the federal government provide a clear illustration of a large number of statistical activities in place and new ones being developed. Despite these various activities, we still lack important statistical information needed for monitoring and evaluating the important changes occurring in the availability, financing, and delivery of health care. The need for these data will become even more critical in the years to come, as the country debates and attempts to reach informed decisions on major health care reform, an issue that will be on the forefront for the next several years in the federal government, the professional community, and the public.
