Click for next page ( 112


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 111
APPENDIX B Survey of Users of National Health Care Statistics INTRODUCTION Evaluation of plans for the National Health Care Survey requires de- tailed knowledge of the needs for the data that such a survey could produce. There is no systematic body of information about the users of NCHS data on health care or about the data needs of the users. In order to address its charge to evaluate plans for a National Health Care Survey, early in the study the panel decided to obtain the views of a wide group of users, policy makers, and other interested parties. Although the panel is comprised of experienced users of health care data and persons highly knowledgeable of the health care delivery system, it wanted to learn from users about the kind of health care data they use; how they use the data; their experience in using data from the current NCHS provider surveys (the National Hospital Discharge Survey, the National Ambulatory Medical Care Survey, the Na- tional Nursing Home Survey, and the National Master Facility Inventory); the problems they have encountered in doing so, their assessment of their quality, cost, and accessibility; their current needs for data on health care and their anticipated future needs; and their comments on the NCHS plans for the integrated National Health Care Survey. As a first step in this direction, the panel decided to conduct an infor- mal survey of users. The information obtained from users had significant influence on the panel in formulating its recommendations. The proposed 111

OCR for page 111
2 APPENDIX B National Health Care Survey, however, cannot meet all the needs for data of all persons. Choices had to be made in terms of costs and benefits associat- ed with policy making in the public and private sectors, as well as the burden on the public. The panel takes full responsibility for the ways in which the users' views have been considered and incorporated in the recom- mendations. METHOD OF DATA COLLECTION The panel had neither the time nor the resources to conduct a large- scale structured user survey designed to meet the standards of survey re- search. The panel therefore obtained the views of users through (1) presen- tations at panel meetings by key users of health care data and policy officials, mostly from the Department of Health and Human Services and some from the private sector and (2) through focused group interviews of selected federal health officials, legislative staff, researchers, and state vital and health statisticians. Approximately 75 users and some nonusers representing the executive and legislative branches of the federal government, health statisticians and vital registrars from several states, and individual researchers in the private sector participated in the survey. The panel recognizes the limited nature of its efforts to obtain the views of users. For instance, the group interviewed is not necessarily representative of all current and potential users of NCHS health care data. Moreover, the expressed needs for data often reflect the mission of the agency interviewed. The interviews were conducted in group sessions with the aid of an open-ended interview guide rather than a structured questionnaire; it ap- pears at the end of this appendix. The guide was adapted to the needs and functions of the specific group interviewed and to the responses received during the specific interview. For several of the interviews, once the pur- pose of the interview was explained, it rapidly became an open-ended dis- cussion. Nevertheless, attempts were made to ensure that all of the areas in the interview guide were covered even in this type of discussion. The areas in which questions were asked included the following: Uses of data from the NCHS Problems with uses of the data Current data needs Health care and health status issues of most concern for the next few years Knowledge of NCHS plans for an integrated National Health Care Survey Views on various features of the survey, such as: Coverage of the survey Content of the survey

OCR for page 111
SURVEY OF USERS 113 Longitudinal data Geographic detail Data on subpopulations Other. Initial contact was made with the group to be interviewed by telephone, followed by a confirmation letter and information about the panel study and the nature of the inquiry. For the few persons who could not attend the group sessions, information was obtained by telephone interview. The decision was made to concentrate mostly on federal agencies in the Washington, D.C., area. The following is a list of agencies that were repre- sented in interviews: Agency for Health Care Policy and Research, PHS Alcohol, Drug Abuse, and Mental Health Administration, PHS Office of the Assistant Secretary for Planning and Evaluation, DHHS Health Resources and Services Administration, PHS Social Security Administration, DHHS Health Care Financing Administration, DHHS National Institute for Occupational Safety and Health, CDC, PHS National Institute of Diabetes and Digestive and Kidney Diseases, NIH, PHS National Institute on Aging, NIH, PHS Staff, House Subcommittee on Health and Environment State health agencies in Connecticut, New York, Pennsylvania, North Carolina, South Carolina, Iowa, Michigan, Utah, Wisconsin, Maine In addition, several individual data users from the private sector were interviewed. FINDINGS The survey of users of health care data was not a scientific sample and the results cannot be expressed statistically. Therefore, the findings are summarized in narrative form on the basis of responses to questions as well as unsolicited comments from the respondents. Uses of Existing NCHS Provider Surveys Most of the respondents were familiar with the four provider surveys, but not equally familiar with all four. Almost all of the groups interviewed had used data from at least one of the surveys. A majority of those inter- viewed responded that they were more familiar with the National Health Interview Survey (NHIS) and the National Nursing Home Survey (NNfIS)

OCR for page 111
114 APPENDIX B than with the National Hospital Discharge Survey (NHDS) and had used, or are currently using, data from these surveys more than the other provider surveys. The respondents were asked in what format or media do they usually obtain the data from NClIS. Published data and special tabulations ap- peared to be the usual format, followed by public use tapes, reflecting the heavy users of the data. During the course of the interviews, the respondents were asked to describe the uses of the data and their experiences in using the data from these surveys. In responding, they tended to focus primarily on the prob- lems they encountered in using the survey data, often for purposes for which the data were not specifically intended. Many of the uses of the data were ad hoc applications to obtain answers to specific questions. The following are provided as illustrative examples of uses of the NCHS health care data and are by no means intended to be exhaustive. Analysis of hospital utilization rates by diagnosis and the distribu- tion of procedures performed. Descriptive analyses of data on specific diseasesfor example, hos- pitalization for diabetes using NHDS, the extent of ambulatory care for diabetes using NAMCS, and some aspects of cost of care for diabetes using the NNHS. Use of the NHDS for a study of hip fractures. Problems Encountered With the Data The respondents mentioned a number of problems that they encoun- tered with the data from the current provider surveys. In general, these problems reflected the needs of the agencies for data that were not being met at the present time. To some extent, therefore, the list of problems reads like a list of data needs. The following list is illustrative of the problems surfaced for each of the provider surveys. National Hospital Discharge Survey . The survey is not person-based. It samples discharges and there- fore does not measure what happens to individuals. There is no inflation on multiple admissions, the patient's contact with oth- er providers, and the episode of illness. There is no information on disposition following discharge from the hospital.

OCR for page 111
SURVEY OF USERS . . 115 There is no information on payers or on cost of care. Infollllation on race/ethnicity is not very useful because of a high proportion of "unknowns." There is no information on occupation of the person discharged. There is no hospital identification that would permit linkage to other data sets. The sample size is inadequate to produce estimates for small geo- graphical levels, subgroups of the population, and relatively rare events. There is insufficient detail in the item content and lack of depth in the topics covered. National Ambulatory Medical Care Survey . . Visits' not persons, are sampled, so there is no information on other contacts with this or other providers. There is no information on outcomes of care. There are no data on cost of care only on source of payment. More diagnostic information is needed five diagnoses rather than three. There is no information on occupation. Diagnostic data are outdated. There are no data on practice of providers and on their characteristics. There is no information on the other providers who assisted the . . physician. The survey does not include data on physician contacts with pa- tients in the hospital. The sample size is too small. The survey is conducted at too infrequent intervals. National Nursing Home Survey The survey does not sample persons. There are no longitudinal data on the person's condition or events and circumstances before and after the specific nursing home stay.

OCR for page 111
116 . . The payer i nformation is inadequate. The sample size is too small for many purposes. Data on subpopulations are inadequate. The rates of refusals and nonresponse are too high. The 8-year gap between surveys is too long. National Master Facility Inventory APPENDIX B The National Master Facility Inventory serves as the sampling, frame for the provider surveys. The overriding comment on the part of the users was that the coverage of the inventory was incomplete. The list of providers that the users proposed for inclusion are discussed below in the section on coverage of the proposed survey. Looking Ahead to the National Health Care Survey Looking ahead to the next few years, respondents were asked (1) what aspects of health care and related health status issues they will be most co-ed with and (2) to comment on the features of the proposed integrat- ed N=rional Health Care Survey that would best meet their needs for data. Key issues of concern in the years ahead identified by respondents are summarized below: . Uninsured and underserved people. The aging population, long-te~ care for the elderly and nonelder- ly, care for the aged and cost of care to the aged, incidence and prevalence of chronic illness, disability, access to home health care. The impact of changes in financing policies on patients and on quality of services, access to care, and outcomes of care. Costs of prescription drugs. Income and expenditure data for health care, including long-term care. The extent of shifts from inpatient hospital care to ambulatory care. Costs and outcomes of care in inpatient and ambulatory settings. Access to care for HIV-infected persons, including adolescents, children, and those in rural areas. Effects on health habits and lifestyles of prevention programs.

OCR for page 111
SURVEY OF USERS 117 The respondents were asked what features would make the proposed integrated National Health Care Survey best meet their needs for data. The sections that follow summarize under several topic headings the comments of the respondents on the features of the proposed survey that would best meet their needs for data. Content. Almost all the groups interviewed expressed a need for data on the cost of health care. The subject was raised in the context of inforrna- tion on health insurance, out-of-pocket expenditures, managed care, types of providers, etc. Although some of these questions can be answered by data from the large national surveys such as the National Medical Expendi- ture Survey, these surveys are conducted too infrequently and the results of these surveys become outdated very rapidly because of changing costs, chan;,es in financing of health care, changes in settings for specific kinds of care, etc. Some of the groups interviewed were interested in the number and characteristics of the medically uninsured, the extent to which they move in and out of insurance coverage, what medical care, if any, they receive, and in what settings. Persor`-based versus provider-based data. As indicated earlier, almost all of the users interviewed indicated the need for data on the health care received by individuals. This included health care received from a range of providers for specific conditions, costs of such care, etc. Coverage, The respondents to the survey of users suggested 37 differ- ent types of providers not now being surveyed by NCHS that should be considered for inclusion in the National Health Care Survey: Adult day care centers Alcohol and drug abuse centers Board and care homes Chiropractors Chronic disease hospitals Emergency clinics Endoscopy units Foster care Group and residential care facilities HMOs Home care facilities Homes for the aged Hospices Hospital outpatient services Managed care facilities

OCR for page 111
118 APPENDIX B Maternal and child health clinics Mental hospitals Mentally retarded facilities Neighborhood health centers Orphanages Podiatrists PPOs Private insurance claims data Psychiatric social workers Psychiatric nurses Psychologists Psychosocial rehabilitation centers Public health clinics Radiology units Rehabilitation medicine units Retirement facilities Rural health centers Shelters Surgical centers Veterans Administration facilities Visiting nurses Walk-in clinics Some of the providers on the list are already under consideration by the NCHS for inclusion and others are already included in surveys being con- ducted by other agencies. Frequency. Very few of the respondents to the survey of users men- tioned the frequency of data collection for the proposed survey. The com- ment most often heard in this area was that 8 years was much too long a gap between National Nursing Home Surveys. Unit of analysis. The basic issue here is whether data collected should be based on a sample of events, such as doctor visits or hospital discharges, on characteristics of providers, or on persons and the medical care they receive. The primary interest of most of the users surveyed was in data on individuals in order to be able to answer questions about episodes of illness, costs of care over specific periods of time, the use of multiple providers, transition from one type of facility to another, access to health care, etc. Several users were interested in data on providers in order to get basic information on the supply of health resources, including characteristics of providers, staffing patterns of providers, number of new nursing home beds, etc. Most also wanted some information on medical care events in addition to the data on individuals in order to measure the volume of care. They

OCR for page 111
SURVEY OF USERS 119 wanted more detailed information than presently available on the cost of service, sources of payment, specifics of services provided, etc. Geographic detail. The estimates made from the current NCHS provid- er surveys are for the national level and for four broad regions of the United - , ~_ _ ~ _^ ~ ^ ~ ~;~ l~r~1 th^~7 ~D^~ ~ ^. ~ ,[~ ~ States. The respondents were asKea al wnaL XC;u',lavl~ l~V~1 L11~' 11~- data. Only a few indicated that they were interested only In na~c~na~ TV; data, and several felt that the four regions were too broad. Many wanted . , . ~ ~ . _ 1 1 ~ . ~ 1 data at the state level, and several have need for data at the county or community level and for rural areas. Subpopulations. Most of the respondents were interested in data for specific subpopulations: blacks, Hispanics, Asians (and specific subgroups of Hispanics and Asians), American Indians, disabled children, agricultural Donulations. specific industrial groups, older workers, disabled workers, children In general, groups defined by socioeconomic level, elderly, and nonelderly in institutions. r -r - - , , Longitudinal data. In most of the user interviews the question of their need for longitudinal data was raised. The term longitudinal was used in at v least two ways: (1) following a cohort of the general population to assess their use of the health care system over a period of time or (2) Following a cohort of persons seen by a particular type of provider to assess outcome or to determine what additional care they received from the entire range of providers over a period of time. Each of these approaches implies obtaining information from respondents about the health care they received over a period of time and then obtaining more specific information from their providers. Several of the respondents expressed the need for longitudinal data on health care to serve a variety of purposes. The most common of these was to be able to assess episodes of illness, which would involve following individuals through the health care system as they had contacts with provid- ers for specific conditions. Respondents in three group sessions indicated the need for data on outcomes of health care. However, they were not specific about what measures they would use and for what purpose, sug- gesting the need for specification of meaningful measures of outcome and the design of the longitudinal aspects of the health care survey to obtain such measures. BAAS r_^V~-A~ ~~ One ot the examples given was to be able to follow a Smart of n~rcnn~ rli~rhn~raed from hospitals to assess the extent of subse- quent hospitalization, contacts with other providers, or death. Several of the respondents recognized two critical problems in obtaining longitudinal data on health care: (1) the need for identifiers on the individuals to be followed and their informed consent to allow their information to be linked with other data sets and (2) the cost of conducting follow-up surveys. Some

OCR for page 111
120 APPENDIX B concern also was expressed about the possible risk of jeopardizing other parts of the survey because of the cost of carrying out longitudinal studies of health care. Linkage. On the subject of linkage with other data sets, most of the respondents wanted to see linkages not just with other NCHS databases, but with social security records, HCFA databases, and several databases in the private sector. It was recognized that, in order to link the databases, identi- f~ers will be needed.

OCR for page 111
SURVEY OF USERS Pane! on the National Health Care Survey Guide for Interviewing Users of Health Care Statistics 121 THE FOLLOWING SHOULD BE USED ONLY AS AN OPEN-ENDED GUIDE, TO BE ADAPTED TO THE NEEDS AND INTERESTS OF THE SPECIFIC GROUPS AND TO THE RESPONSES RECEIVED. [The meeting will begin with an introduction and a five minute background of the study, the objectives of the session and the sur- veys covered in the questions.] Current Data Needs 1) What aspects of health care and related health status issues are you currently most concerned with? 2) What kinds of programmatic action, policy decisions or research products result or could result from your work in this area? 3) What types of statistical information do you require to carry out analyses and or make decisions in this area? Use of Data from NCHS 4) Have you used, or are you currently using, any data from NCHS on these subjects? User Nonuser IF NONUSER GO TO QUESTION 14 5) Are you a user of any of the following NCHS surveys of health care? yes National Master Facility Inventory or Inventory of Long-Term Care Places National Hospital Discharge Survey National Ambulatory Medical Care Survey National Nursing Home Survey No IF YES, ASK WHICH SURVEYS. IF NO, GO TO QUESTION 14

OCR for page 111
122 Questions for Users APPENDIX B 6) In what format or media do you usually obtain these data publi- cations, special tabulations, public use tapes, etc. ? IF PUBLIC USE TAPES ARE USED ASK: How many tapes have you used in the last 2 years? Are you on the mailing list for NCHS publications? yes No FOR EACH OF THE HEALTH CARE SURVEYS THAT YOU SAID YOU USE: 7) What uses have you made, or do you make, of data from this sur- vey? 8) What end products derive from your use? (testimony, policy pa- pers, review of legislative proposals, budget estimates, reports, re- search papers, etc.) Obtain some citations, if possible? 9) What do you consider to be the strong points of these surveys? Problem Areas 10) What do you consider to be the weak points of these surveys? IF NONE, GO TO QUESTION 14 11) What kinds of problems have you encountered in attempting to use data from these surveys? WHICH OF THESE SURVEYS? 12) Have you had projects for which these surveys should be uniquely suited but cannot be used because of problems of design or lack of relevant data elements or sample size or other problems? IF YES, ASK TO ELABORATE Do you have any suggestions for solving these problems? 13) Do you use other sources of data in these subject areas? IF YES, WHAT ARE THEY?

OCR for page 111
SURVEY OF USERS Looking Forward 123 14) Looking ahead to the next few years, what aspects of health care and related health status issues do you think that you will be most concerned about? 15) For the past few years NCHS has been developing an integrated national health care survey including a follow-up component. Have you been aware of NCHS plans for an integrated National Health Care Survey ? Yes IF NO, BRIEFLY DESCRIBE no The new survey might include information covering a number of topic areas: Health care facilities Health care services provided Health manpower Recipients of health care Outcomes of health care The context within which the care is provided 16) Do you think data on any of these topics will be of particular interest to youwhich ones and for what purposesshould any other topics be considered? 17) From your perspective what features would make the proposed in- tegrated National Health Care Survey best meet your needs for data? ~ It's Aicr''cc rf~v~rn~Pr isles first. a. Let s Discuss coverage 15bU~b Ills. Do you have any thoughts on which facilities and places should be covered in this new National Health Care Survey ? b. Turning to the content of the survey, we are especially inter- ested in those areas that are emerging or are of increased im- portance. What content areas should be covered in the new survey? c. How frequently do you need the data you mentioned earlier? d. Some data uses require longitudinal estimates, tracking chang- es over time for the same respondents. To what extent are longitudinal data needed in the area of health care statistics?

OCR for page 111
24 APPENDIX B PROBE FOR SPECIFIC EXAMPLES Some data needs suggest different approaches to survey design. One in particular that we seek your views on is unit of analysis. In the area of health care surveys, the unit of analysis may be facility (hospital, nursing home, doctor's office), event (admission, dis- charge, visit, particular services), or person (patient, client): e. From your viewpoint, how important is it to obtain data on the facility? event? individual? For whet kinds of analyses would data on each of these be important? Another feature of interest may be geographic detail. Are national data sufficient for your purposes, or do you need data for geographic subdivisions? If the latter, in what geo- graphic detail? IF RESPONSE IS "SMALL-AREA" OR "LOCAL," PROBE FOR DEFINITION g. Do you need information on any specific subpopulatior~s? WHICH GROUPS? h. Are there any other features you feel should be addressed in the review of the National Health Care Survey? 18) Are there other sets of data to which you would wish to link data from the National Health Care Survey? IF YES, WHAT ARE THEY? Can you give examples of how these data would be used - that is, what kinds of programmatic actions, policy decisions or research products might result from an analysis of the data? What identifiers would be needed for the linkageSocial Security number, Medicare number, Medicaid number, or something else? Do you see any problems in obtaining them? 19) In what format should data be made available in the future?- published reports, data tapes, special tabulations, or any other me- dia? WRAP UP