Skip Navigation U.S. Department of Health and Human Services
Agency for Healthcare Research Quality
Archive print banner
Future Directions for Community-Based Long-Term Care Health Services Research

This information is for reference purposes only. It was current when produced and may now be outdated. Archive material is no longer maintained, and some links may not work. Persons with disabilities having difficulty accessing this information should contact us at: Let us know the nature of the problem, the Web address of what you want, and your contact information.

Please go to for current information.

Day Two

Data Presentations

The goal of day 2 was to present a summary of the most important data sources that contain community-based long-term care information and query attendees about their use of these data, inviting any suggestions they had about possible improvements. The morning session included five informal presentations of six data sets, followed by an afternoon discussion of data gaps and a concluding discussion on how to fill these gaps.

The presentations included the following data sets:

  • Medical Expenditure Panel Survey.
  • Medicare Current Beneficiary Survey.
  • Health and Retirement Study.
  • Health Dynamics Among the Oldest Old.
  • National Long-Term Care Survey.
  • National Home and Hospice Care Survey.

Medical Expenditure Panel Survey (MEPS)

The MEPS is an ongoing, nationally representative survey designed to capture various aspects of medical care for the noninstitutionalized U.S. population. The MEPS consists of four survey components: Household, Medical Provider, Insurance, and Nursing Home (1996 only). The Household Component collects data on families and individuals across the Nation and is drawn from a nationally representative subsample of households that participated in the prior year's National Center for Health Statistics (NCHS) National Health Interview Survey. AHRQ began fielding MEPS in March 1996. AHRQ conducts MEPS in conjunction with NCHS and through contracts with Westat, a survey research firm located in Rockville, MD.

MEPS is the most recent in a series of medical expenditure surveys that began in 1977 as the National Medical Care Expenditure Survey and later became the National Medical Expenditure Survey. The purpose of these surveys is to provide nationally representative estimates of health care use, expenditures, sources of payments, and insurance coverage for the U.S. noninstitutionalized population.

Community-Based Variables. Utilization of community-based long-term care can be identified from the Household Component of the MEPS with variables specific to adults and children.

Variables available to identify chronic illness and disability for adults include the following:

  • Conditions and impairments associated with use of health care services.
  • Work, school, or housework limitations.
  • Functional limitations (e.g., walking, bending).
  • Receipt of help with ADLs or IADLs (one question for each category).
  • Social role limitations.
  • Assistive devices.
  • Hearing, vision, and cognitive limitations.

Variables available to measure disability for children include the following:

  • For children ages 4 and under:
    —Limitations in any activities.
    —Limitations in play activities.
    —Participation in special programs or early intervention services.
  • For children over age 4:
    —Interpersonal problems.
    —Limitations in school attendance.
    —Enrolled in special education or related services.
    —Limited in activities other than school.

Home Health Care. Home health data have been collected for every year between 1996 to present; data for 1996-99 have been released. Sample sizes may be small for particular subpopulations but are expected to increase in the future.

Home health care data are collected each round and identify all sources of help including family, neighbors, friends, and paid care. Agency-based providers are identified and caregivers from agencies are tracked on a monthly basis. The survey respondent provides estimates of the number of visits per month and average length of time per visit.

Details captured about visits include type of provider, nature of help, expenditures for care, and sources of payment. Additionally, the type of health problem associated with the receipt of care is captured.

Supplements. A long-term care supplement was fielded in 1996, 1997, and 1998. Persons were screened into the supplement if they were flagged for any of the following in any round: ADL or IADL deficits, use of special equipment, functional limitations, work, housework, school limitations, social limitations, cognitive limitations, vision or hearing deficits, or child with limited activities or school attendance limitations. The supplement includes details on ADLs and IADLs, communication limitations, developmental delays, special diets, employment accommodations, transportation, and the condition that is associated with the limitations.

Concurrent with the long-term care supplement, a caregiver supplement was administered in 1996, 1997, and 1998. Persons were screened in using the same criteria as the long-term care supplement. It captured the type of help received at home, time spent by caregiver, and changes in living arrangements. Potential caregivers outside of the household were also identified.

More information on MEPS can be found at

Medicare Current Beneficiary Survey (MCBS)

The Medicare Current Beneficiary Survey, sponsored by CMS, has been in the field since 1991. It is an ongoing, nationally representative sample of approximately 12,000 Medicare beneficiaries. The purpose of this survey is to develop reliable information on Medicare-covered and non-Medicare-covered services and costs, to study transitions in beneficiaries' lives, and to monitor the effect of change on the Medicare beneficiary population. Data are collected for beneficiaries living in the community and in health care institutions and are supplemented with Medicare claims data.

There are three rounds of data collection per year. Persons are followed for 4 years, from setting to setting, to capture 3 calendar years' worth of health care use and expenditures and transitions in care. The design is an overlapping panel with an oversample of the disabled and oldest old. Persons are sampled from the Medicare rolls; and, after the first round, questions about cost and use of services are asked for the period since the last interview.

Community-Based Variables. Questions about general health, chronic conditions, and functional limitations (ADLs and IADLs) are asked. Use of health care includes Medicare-covered services, noncovered services, and preventive services. Total payments by source of payment (up to eight sources) are captured. Health status measures include the following:

  • General health.
  • Vision and hearing.
  • Height and weight.
  • Presence of chronic conditions.
  • Continence.
  • IADLs and ADLs.
  • Physical functioning and locomotion.
  • Preventive services: mammogram and Pap smear, flu and pneumonia shots, prostate specific antigen and digital exam.

Additional questions about the living arrangements of survey respondents are being planned.

Home Health Care. Home health care questions are limited to a simple acknowledgment of receipt of care at home, type of organization providing care, reasons for not using a home health provider, time per visit, receipt of medical care (e.g., changing dressings, giving shots), help with IADLs or ADLs, receipt of personal care from other nonresident sources, and the type of relationship to respondent. Affiliations with Department of Veterans Affairs facilities or health maintenance organizations are also captured.

Health and Retirement Study (HRS)

The Health and Retirement Study is a nationally representative, longitudinal survey of more than 12,000 persons born during the years 1931 to 1941. The focus of the survey is on health, retirement decisions, and the physical and financial well-being of Americans in later life. It is intended to provide data for researchers, policy analysts, and program planners who are making major policy decisions that affect retirement, health insurance, saving, and economic well-being. The University of Michigan Institute for Social Research conducts the HRS under a cooperative agreement with the National Institute on Aging.

HRS Questionnaire Topics.

  • Health and cognitive conditions and status.
  • Retirement plans and perspectives.
  • Attitudes, preferences, expectations, and subjective probabilities.
  • Family structure and transfers.
  • Employment status and job history.
  • Job demands and requirements.
  • Disability.
  • Demographic background.
  • Housing.
  • Income and net worth.
  • Health insurance and pension plans.

Wave 1 data collection was completed in February 1994. Wave 2 data collection ended in May 1996. Wave 3 data collection ended in 1998; future waves will largely replicate HRS 1998 in design, format, coverage, structure, and measurement. Wave 4 data collection was fielded as a joint data collection effort with Health Dynamics Among the Oldest Old and ended in March 1999.

Health Dynamics Among the Oldest Old (AHEAD)

The AHEAD study is a nationally representative, longitudinal survey of more than 7,000 persons over age 70 and provides data to address a broad range of scientific questions focused on the interplay of resources and late life health transitions. Among these issues are the costs of illness borne by the family; differences in how resources are used to offset cognitive, physical, and functional losses; the effectiveness of various care arrangements in preserving function and delaying institutionalization; the extent to which transfers from kin buffer the assets of older persons and slow transitions to late life impoverishment; and the extent of and mechanisms for spending down private assets as well as Medicaid spend-down.

AHEAD Questionnaire Topics.

  • Cognitive performance.
  • Physical and functional health.
  • Economic status (assets and income).
  • Claims on transfer programs and contingent claims.
  • Spend-down and Medicaid eligibility.
  • Family structure, caregiving, and financial transfers.
  • Demographic characteristics.
  • Housing (including access to services).
  • Service use (community and nursing home).
  • Out-of-pocket costs for all services.

Other Variables. Both HRS and AHEAD contain data on utilization of care, health, living arrangements, and economic resources/behavior. Respondents who reported one or more of 12 physical limitations were asked questions about difficulties in ADLs. All respondents were asked IADL questions. Questions about time spent by helpers, cost, and availability of helpers were asked.

Both surveys include home care use and utilization data in the past 2 years and use of other special services (e.g., adult care center, social worker, outpatient rehabilitation, transportation, and meals). Use and expenditure information was also captured for hospitals, nursing homes, outpatient services, dental care, and prescription drugs.

Health measures include self-ratings and change, disease inventory with some severity measures, pain assessment, cognitive competencies, subjective expectation of life span, and need for nursing care. Living arrangement data include housing, services, ownership, and value. Data about children and parents (if applicable) are also gathered.

More information can be found at Exit Disclaimer

National Long-Term Care Survey (NLTCS)

The National Long-Term Care Survey is a nationally representative survey of Medicare beneficiaries age 65 and over with chronic functional disabilities. Data have been collected for 1982, 1984, and every 5 years since. Persons (in the community and health care institutions) are sampled from the Medicare beneficiary rolls, and those with functional limitations are oversampled.

The NLTCS for 1982, 1984, 1989, and 1994 are designed to measure the point prevalence of chronic (90 days or more) disability in the U.S. elderly Medicare enrolled population and changes (both improvement and incidence) in chronic disability (and institutionalization) over time. Slightly different sampling procedures were utilized for each of the survey years.2

For the elderly population with disabilities, the NLTCS provides cross-sectional and longitudinal data on: prevalence and patterns of functional limitations, medical conditions and recent medical problems, use of health services, kinds and amounts of formal and informal long-term care services used, demographic characteristics, public and private expenditures for health services, and housing and neighborhood characteristics.

2 At the time of the writing of this report, a final version of the 1999 data was not available.

Community Questionnaire. The community portion of the survey includes information on the following:

  • Condition list.
  • ADL status (detailed questions on six ADLs).
  • IADL status (detailed questions on seven IADLs).
  • Source, type, and amount of informal help.
  • Source, type, amount, and payer for formal help.
  • Questions on range of motion and impairment (Nagi items).
  • Activity list.
  • Nutrition.
  • Social activities.
  • Alcohol consumption and smoking.
  • Other functioning (mental, emotional, behavioral).
  • Housing and neighborhood characteristics.
  • Health insurance.
  • Medical providers and prescription medicines.
  • Cognitive functioning.
  • Military service, ethnicity, income, assets.

Informal Caregiver Survey. A separate (but linked) informal caregiver supplement was administered in the surveys for 1982, 1989, and 1999. A caregiver was defined as someone who provided help to a sample member with one or more IADLs or ADLs in the week prior to the sample interview. The caregiver was either a paid or unpaid relative or unpaid nonrelative.

The supplement includes the following information:

  • Caregiver name, address, and basic demographics.
  • Relationship to sample member.
  • Amount and kinds of help provided.
  • Information on care provided by others.
  • Caregiver's living and work situation.
  • Caregiver's health and functional status.
  • Caregiver's income and asset.

The instruments used for the informal caregiver surveys have not remained static across the life of the NLTCS. The 1982 and 1989 caregiver surveys are similar in content although the 1989 survey was shortened.

The 1999 survey contains detailed caregiver information and includes asking caregivers whether they helped with IADLs, ADLs, toileting, medications, and other medical care, as well as the number of times this help was given. Other questions about the amount of time the caregiver can leave the sample person alone and questions on problems and inconveniences experienced by the caregiver also are asked. Sources of and satisfaction with outside help, caregiver experience, living situation, and work situation are other areas covered.

More information can be found at

National Home and Hospice Care Survey (NHHCS)

The National Home and Hospice Care Survey is a set of surveys of a random sample of home and hospice care agencies in the United States. The survey was been fielded in 1992, 1993, 1994, 1996, and 1998. The survey includes all types of agencies that provide home health and hospice care without regard to whether they are Medicare or Medicaid certified or whether they are licensed. Information is collected about the agency, current patients, and discharges.

All data collected in the NHHCS consist of three files: agency, current patients, and discharged patients. Each record includes a common agency number to allow linkage between them.

Agency Variables.

  • Type of ownership.
  • Agency affiliation.
  • Certification status.
  • Number of current patients.
  • Services available: 1993, 1994, 1996, 1998.
  • 1992 only:
    —Number of full- and part-time staff.
    —Part-time staff hours.
    —Nonpayroll staff hours.

Patient Variables (Current and Discharged).

  • Demographics (age, sex, race, marital status).
  • Current living arrangements.
  • Referral source.
  • Primary caregiver and relationship.
  • Diagnoses at admission, time of survey, and discharge (if applicable).
  • Reason for discharge.
  • Type of care received.
  • Living arrangements at discharge.
  • Aids used.
  • Vision and hearing status.
  • Activities of daily living.
  • Instrumental activities of daily living.
  • Services provided.
  • Service providers.
  • Amount billed for care.
  • Primary and secondary sources of payment.
  • Number of visits: 1992, 1993, 1994.
  • Surgical and diagnostic procedures related to admission: 1996 only.
  • Dates covered by bill: 1992, 1993, 1994, 1998.

More information can be found at

Improving Data

Prior to the meeting, participants were asked to comment on ways they would like to see the data sets improved. Highlights from their responses follow:

  • National long-term care data on persons living in the community have focused mainly on the elderly; more data are needed about the nonelderly.
  • Researchers expressed concern that some rich data sets such as the disability supplement to the National Health Interview Survey are only one-time efforts; thus, trends in care and health status cannot be monitored. As a consequence, we cannot get parallel information about disability trends for the nonelderly as we get for the elderly from the NLTCS, for example.
  • Information on quality of long-term care services provided in the community is lacking.
  • Analyses are often limited by small sample sizes, especially when trying to compare subpopulations of the long-term care population, such as children with special needs, adults with disabilities, or persons with specific chronic conditions.
  • Combining data to link all care received to an individual is often impossible when care is provided across settings and financed by different payers.
  • Data on informal caregivers need to be expanded to better capture influence on caregivers' lives, including influence on work and social, psychological, and physical burden.
  • Access to data has become very difficult in some cases because of confidentiality concerns.
  • Administrative data such as OASIS should be linkable with national data to reduce the burden of survey data on providers.
  • Outcome measures in national data need to be improved to increase the sensitivity of measures to change.
  • Information about satisfaction with community-based long-term care services currently is not available in national survey data.

Identifying and Prioritizing Data Gaps

Following the presentation of data sets, a discussion was held to summarize and prioritize the most important data gaps in community-based long-term care. A number of concerns were emphasized:

  • Participants agreed that both facility-based and population-based surveys were important. Many stressed that in either type of survey, person-level longitudinal or repeated panel, information is needed to enable the monitoring of changes in quality, outcomes, access, and cost—an important goal for national data. These data can be supplemented periodically with special studies, but a coherent strategy should be put into place.
  • Priority should be given to increasing the sample sizes of the long-term care populations so that separate analyses of subpopulations can be made.
  • Administrative data should be made more useful by improving data linkages (e.g., improved linkages across payers, common person-based identifiers, and centralized data cleaning). Access should be improved to enable researchers to take advantage of these important data. Participants were sensitive to confidentiality issues, but expressed concerns that confidentiality issues should be balanced by concerns about facilitating research. They expressed concerns about limitations in access and constraints on linking survey data to Social Security Administration, Medicaid, and Medicare administrative data.
  • Caregiver data, especially for children and adults with disabilities, need to be collected. More information is needed on how caregiving decisions are made and how caregiving responsibilities are allocated among potential caregivers; estimates are needed about caregiver burden—economic, physical, and psychological.
  • Information about service-enhanced housing needs to be improved to better capture long-term care services provided in home-like environments. This would also improve our ability to capture transitions from home and across different levels of service-enhanced settings.
  • Participants viewed the development of a frame of residential long-term care providers as very important and believed it should include all types of service-enhanced settings, if possible. However, they acknowledged that it is difficult to build a frame when there are many small facilities. This is precisely the nature of the board- and-care market, especially those places serving the nonelderly. These facilities have grown greatly in number but are usually small places serving only a few people.
  • Descriptive data on care received should be improved to capture both informal and formal care.
  • Disability measures for children should be improved.

Other Data Issues

Measuring Unmet Need in ADLs and IADLs. Some participants advocated for improving the measurement of unmet need. The discussion centered around the NLTCS and whether the low estimates of unmet need in that survey were real or artifacts of measurement. There seemed to be a consensus that the measurement of unmet need in the NLTCS could be improved. To get valid responses, a debate exists about whether one can ask if someone has an unmet need, or if more specific questions should be asked about the consequences of the lack of available help. More specific questions were asked in the Longitudinal Survey on Aging. Estimates of unmet need collected periodically would provide an indication of how the long-term care system is meeting needs over time.

State-Level Data. Participants emphasized that most of community-based long-term care is financed at the State level, and consequently, policy decisions are often made by States, not the Federal Government. There are tremendous differences in the array of government programs that exist across States as well as how States use Medicaid waiver programs. There are no consistent data collection systems for State programs.

Some States are able to monitor their home and community-based services well while others are not. Consequently, it is difficult to compare States on almost all dimensions of service delivery. Participants emphasized that it wasn't enough to mandate the availability of data unless a uniform data system was developed. In addition, most national data sets are not designed to provide State estimates. Some participants believed that an agency of the Federal Government might help stimulate a consistent data collection system. Others discussed the importance of studying a few States in detail or even a few major cities. As an alternative approach, participants discussed developing a State data system that could be linked to national data so that the impact of the variation in programs across States could be studied.

Capturing Transitions. Most participants wanted improved data on transitions in care and health, including capturing the important changes and being able to study transition decisions. Another concern is that the change period captured should be small enough to include important clinical changes. For the Medicare population, the MCBS follows persons across all settings and captures all use and expenditures either from Medicare and Medicare bills or from self-report over a 3-year period. For persons of all ages and all payers, MEPS follows persons for 2 years and captures all changes in home health on a monthly basis. It only captures changes in functioning three times over the 2 years, allowing only for approximate annual changes in status to be measured. For home health care, the potential for linking OASIS data, which include health changes over a 60-day period, was discussed. National surveys do not currently link to the OASIS data.

Although the MCBS and MEPS follow persons across settings, the sample size is a concern. To understand transition decisions, information needs to be collected at admission or at discharge. Both service and health data need to be collected and linked.

In addition to capturing clinical information inadequately, participants believed that data do not sufficiently capture social transitions. For children with special health care needs, transitions to school are important but are not generally captured in health surveys. For adults with disabilities, transitions in work, family, and other social activities should be captured.

Special Issues Related to CSHCN

For children with special health care needs, participants emphasized concerns about sample size and screening criteria. Because educational services also contribute to health outcomes, the importance of linking educational activities with health care and long-term care arrangements was also stressed. Children with special health care needs comprise the smallest of the three main disability groups, but this population has its own service delivery issues. Therefore, CSHCN must be analyzed separately, which, in turn, makes oversampling this population of paramount importance.

There are plans to oversample CSHCN in the MEPS, but this will depend on future budgets. Also, a common definition of CSHCN is needed. An example of this is the development of a CSHCN screener that is being incorporated into the MEPS by the Child and Adolescent Health Measurement Initiative's Living with Illness Task Force.

Measuring Home Care Quality and Staffing

Participants believed that it was important to include quality measures for long-term care even in national surveys and that home care outcome measures could be broadened. The importance of better capturing consumer satisfaction and the staffing profile in home care agencies was stressed. Information should be collected on staff ratios, training, and wages, whether the staff are independent providers or affiliated with an agency, and whether they are directly hired by the patients.

Related to staffing for home care is the need for better information about long-term care management. Care management is difficult to measure because there are no common definitions, financial incentives vary, and care intensity varies. Participants recommended collecting a measure of case load, services provided, and professional training. Clients may receive care from more than one care manager, which should also be captured in some way.


This meeting provided AHRQ with a broad view of the data needs and research priorities for community-based long-term care research. Participants communicated a number of important messages. The following recommendations summarize the next steps that are needed to enhance information about community-based long-term care.

  • Improve basic information about the caregiving patterns and service use of persons of all ages, but especially for those under age 65.
  • Evaluate new models and financing mechanisms for improving outcomes from community-based care.
  • Evaluate models to improve methods to encourage the hiring and retention of new staff.
  • Expand our ability to measure quality beyond home health including less clinical forms of home care.
  • Develop outcome measures that are sensitive to the differing goals of the three long-term care populations: CSHCN, adults with disabilities, and the elderly.
  • Develop better measures of satisfaction, quality of life, and community and housing environments.
  • Improve data about State long-term care programs to assess the large variation in access to care.
  • Increase national samples to enable separate analyses of the major subpopulations of long-term care users who have very different goals and needs.
  • Increase information about services provided in service-enhanced housing.
  • Capture important transitions in health, levels of care, and social roles to improve our understanding of how decisions are made at transition points.
  • Facilitate access to administrative data and linkages between national surveys and administrative data.

Return to Contents

AHRQ Publication No. 02-0022
Current as of April 2002


The information on this page is archived and provided for reference purposes only.


AHRQ Advancing Excellence in Health Care