Health and Retirement Study (HRS)

01/02/2003

 

TITLE: The Health and Retirement Study

ACRONYM: HRS

AGENCY/PROGRAM: National Institute on Aging

DESCRIPTION:

The Health and Retirement Study provides data to researchers, policy analysts, and program planners to inform major policy decisions affecting retirement, health insurance, saving and economic well-being. These groups have already made extensive use of the HRS to study individual work, savings, and retirement decisions and to model behavioral response to changes in Social Security and Medicare benefits.

Description:

HRS is a national panel study based on biennial interviews. The original sample consists of in-home, face-to-face interviews of persons who were age 51-61 in 1992, and their spouses (7,600 households, over 12,600 persons). Hispanics, Blacks and Florida residents were oversampled. The HRS data collection has incorporated the AHEAD sample (see description of the Asset and Health Dynamics Among the Oldest Old Study below) as of 1998. Follow-ups are done by mail/phone every second year. New baseline cohorts (those born in 1942-1947 or in 1924-1930) were added in 1998.

The HRS is linked with the following administrative data: 1) Health and pension benefit plans collected from employers, 2) Data from the National Death Index, 3) Social Security Administration earnings and benefits data, and 4) Medicare files when respondents reach Medicare age

Questionnaire topics include: 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; disability; demographic background; housing; income and net worth; health insurance and pension plans; experimental modules. Next-of-kin provide information on health changes and distribution of assets for deceased respondents. Additional data from employers, Medicare files, the National Death Index, and Social Security Administration earnings and benefits files are being linked to interview data.

RACE AND ETHNICITY:

The HRS includes 100 percent oversamples of Hispanics and Blacks. The questions used to collect this information are as follows:

  • In what country were you born?
  • In about what year did you come to the U.S.?
  • Do you consider yourself Hispanic or Latino? Inapp, Yes, No, DK, RF
  • Would you say you are Mexican American, Puerto Rican, Cuban American or something else? Inapp, Mexican American/Chicano, Puerto Rican, Cuban American, Other, DK, RF
  • Do you consider yourself primarily White or Caucasian, Black or African American, American Indian or Asian, or something else? Inapp, White/Caucasian, Black/African American, American Indian or Alaska Native, Asian or Pacific Islander, Other, DK, RF

STATUS: Ongoing

HOW TO ACCESS DATA:

Public Use Data Files -- All publicly released data are available to download via our website at http://www.umich.edu/~hrswww. Datasets for Waves 1 and 2 of HRS are publicly available, as is a partial public release of HRS Wave 3. Data from complementary sources such as SSA and Employer Pension Plans are also available through the Restricted Data Process described on the website.

The final public release of HRS Wave 3 is tentatively scheduled for August 1999. A very preliminary dataset from the HRS-98 data collection is expected to be available in June 1999. (Preliminary data releases are available under terms of a Conditions of Use Agreement; for more information contact via email: hrsquest@isr.umich.edu).

Current Activities -- Field activities on Wave 4 (HRS-98) are now coming to a close. This round of interviews included approximately 22,615 sample members consisting of the original HRS and AHEAD cohorts plus two new birth cohorts (1924-30 and 1942-47). From this point forward the HRS study will be a seamless representation of the U.S. non-institutionalized population in their fifties and older.

Future Plans -- A proposal for renewal funding to collect three more waves of data has been submitted to NIA for review. Future data collections will largely replicate the HRS-98 in design, format, coverage, structure and measurement. Proposed plans include the introduction of a new cohort in 2004.

CONTACT NAME AND PHONE NUMBER:

Robert J. Willis, Principal Investigator Richard Suzman, Program Officer

Institute for Social Research National Institute on Aging

University of Michigan (301) 496-3138; fax (301) 402-0051

(734) 936-7261; fax (734) 647-1186

For additional information about the HRS study or datasets contact Health and Retirement Study, Survey Research Center, P.O. Box 1248, Ann Arbor, MI 48106-1248; phone: 734-936-0314; fax: 734-647-1186; or email: hrsquest@isr.umich.edu.

Internet: hrsquest@isr.umich.edu The HRS and AHEAD home page address is http://www.umich.edu/~hrswww/


TITLE: Survey of Asset and Health Dynamics Among the Oldest-Old: HRS Auxiliary Study also known as Aging and Health in America

ACRONYM: AHEAD

AGENCY/PROGRAM: National Institute on Aging

DESCRIPTION:

Purpose -- The course of age-related changes in health can be affected by several broad types of "resources:" economic resources which result from past employment, consumption, and savings; claims on structured programs such as Social Security, Medicare, and Medicaid; family resources of time and money which can be made available to an older relative for caregiving or the purchase of services.

Older individuals have differing levels and mixes of resources and consume them at varying rates as they age. A decline in health is likely to increase consumption of resources. Despite a number of recent studies of late life transitions, none provide detailed coverage of the joint dynamics of health, economic and family resources, and care arrangements.

The AHEAD study is intended to reduce the knowledge gap by providing data to address a broad range of scientific questions focused on the interplay of resources and late life health transitions.

Description -- AHEAD is a panel study of persons born in 1923 or earlier identified in the screening of 69,000 households for the 1992 HRS baseline (see description of the Health and Retirement Study above). AHEAD also includes the same oversamples as HRS and a supplemental sample of persons age 80 and older from the Medicare Master Enrollment File (HCFA). Next-of-kin provide information on health changes and distribution of assets for deceased respondents. More than 8200 persons were interviewed in the first wave of AHEAD in 1993.

Questionnaire topics include:

  • cognitive performance-based testing; physical and functional health; economic status (assets and income);
  • claims on transfer programs and contingent claims; dissaving 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; seven experimental modules, including innovative measures such as a resiliency scale.

Data are being linked with information from Medicare and Medicaid files, the National Death Index, and Social Security Administration earnings and benefits files.

Current Activities -- The third wave of AHEAD was fielded in combination with HRS-98. These field activities are now coming to a close. This data collection included approximately 22,615 sample members consisting of survivors of the original HRS and AHEAD cohorts plus two new birth cohorts (1924-30 and 1942-47). From this point forward the AHEAD study will be merged with the HRS study, and the merged study will be a seamless representation of the U.S. non-institutionalized population in their fifties and older.

Future Plans -- A proposal for renewal funding to collect three more waves of data has been submitted to NIA for review. Future data collections will largely replicate the HRS-98 in design, format, coverage, structure and measurement. Proposed plans include the introduction of a new cohort in 2004.

RACE AND ETHNICITY:

The AHEAD includes 100 percent oversamples of Hispanics and Blacks. The questions used to collect this information are as follows:

  • In what country were you born?
  • In about what year did you come to the U.S.?
  • Do you consider yourself Hispanic or Latino? Inapp, Yes, No, DK, RF
  • Would you say you are Mexican American, Puerto Rican, Cuban American or something else? Inapp, Mexican American/Chicano, Puerto Rican, Cuban American, Other, DK, RF
  • Do you consider yourself primarily White or Caucasian, Black or African American, American Indian or Asian, or something else? Inapp, White/Caucasian, Black/African American, American Indian or Alaska Native, Asian or Pacific Islander, Other, DK, RF

STATUS: Ongoing

HOW TO ACCESS DATA:

Public Use Data Files --All publicly released data are available to download via the HRS website at <http://www.umich.edu/~hrswww>. The AHEAD Wave 1 dataset is publicly available, as is a partial public release of AHEAD Wave 2. The final public release of AHEAD Wave 2 is tentatively scheduled for August of 1999.

A very preliminary dataset from the HRS-98 data collection (which includes AHEAD Wave 3) is expected to be available in Summer of 1999. (Preliminary data releases are available under terms of a Conditions of Use Agreement; for more information contact via email: hrsquest@isr.umich.edu).

CONTACT NAME AND PHONE NUMBER:

Robert J. Willis, Principal Investigator Richard Suzman, Program Officer

Institute for Social Research National Institute on Aging

University of Michigan Phone: (301) 496-3138

Phone: (734) 936-7261 Fax (301) 402-0051

Fax (734) 747-1186

Internet at AHEADASK@umich.edu The HRS and AHEAD home page address is http://www.umich.edu/~hrswww/

For additional information about the AHEAD study or datasets contact Health and Retirement Study, Survey Research Center, P.O. Box 1248, Ann Arbor, MI 48106-1248; phone: 734-936-0314; fax: 734-647-1186; or email: hrsquest@isr.umich.edu.


TITLE: National Long-Term Care Survey

ACRONYM: NLTCS

AGENCY/PROGRAM: National Institute on Aging

DESCRIPTION:

Purpose:

The 1982, 1984, 1989 and 1994 National Long Term Care Surveys (NLTCS) are surveys of the entire aged population with a particular emphasis on the aged who are functionally impaired. The samples drawn from aged Medicare beneficiary enrollment files are nationally representative of both community and institutional residents. As sample persons are followed through the Medicare record system, virtually 100% of cases can be longitudinally tracked so declines, as well as increases in disability may be identified as well as exact date of death. NLTCS sample persons are followed until death and are permanently and continuously linked to the Medicare record system from which they are drawn. Linkage to the Medicare Part A and B service records extends from 1982 to 1995, so that detailed Medicare expenditures and types of service use may be studied. Through the careful application of methods to reduce nonsampling error, the surveys provide nationally representative data on:

  • the prevalence and patterns of functional limitations, both physical and cognitive;
  • longitudinal and cohort patterns of change in functional limitation and mortality over 12 years;
  • medical conditions and recent medical problems;
  • health care services used;
  • the kind and amount of formal and informal services received by impaired individuals and how it is paid for;
  • demographic and economic characteristics like age, race, sex, marital status, education, and income and assets;
  • out-of-pocket expenditures for health care services and other sources of payment;
  • housing and neighborhood characteristics.

Description:

The 1982, 1984, 1989, and 1994 NLTCS 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.

This list sample for the 1982 NLTCS was randomly drawn from Medicare administrative files. 35,008 persons over age 65, Medicare eligible and alive on April 1, 1982 were drawn. A two-stage procedure identified chronically disabled persons for household interviews. In the first stage all 35,008 persons were screened (80% by phone; 20% in person). This identified 6,393 persons in 1982 with at least one chronic (duration of 90 days or more) impairment in seven Instrumental Activities of Daily Living (IADL) or nine Activities of Daily Living (ADL). Since residence was confirmed on the screen, 1,992 institutional residents can be separated from the 26,623 nondisabled community residents (of the 35,008 total) who reported no ADL or IADLs impairments. The 6,393 chronically disabled, community residents in 1982 were interviewed at home to assess health, functioning, social, economic, and other factors. Interviews were completed with 6,088 (95.2%) -- 24.7% were conducted with the help of proxies. No institutional interviews were conducted in 1982.

The 1984 NLTCS had a complete longitudinal design with an institutional component. All persons surviving to 1984 who, in 1982, either reported chronic disability or were in institutions, were reinterviewed in 1984 with either a detailed community or institutional questionnaire. Consequently, not only disabled survivors were tracked, but also previously disabled survivors who had become nondisabled. In addition, 12,100 community residents (45.4% of 26,623) screened in 1982, and who were not chronically disabled, were rescreened with the same instrument in 1984 to determine disability incidence. A new sample of 4,916 persons who became age 65 between the 1982 and 1984 surveys and who survived to 1984 was drawn from Medicare files and screened. Incident cases of chronic disability from either screen sample component received either a detailed in-home or institutional interview. The 1984 sample provides estimates of two-year disability and mortality rates for a longitudinally followed population and representative cross-sectional prevalence estimates of disability and institutional residence of the total 1984 U.S. elderly Medicare enrolled population 65 and over (Manton 1988).

The 1989 NLTCS was designed to estimate disability and institutionalization rates for the 1982-1984 and 1984-1989 periods. Persons in either 1982 (or 1984) who were institutionalized, or who received a detailed community survey and survived to 1989, were reinterviewed -- either in the community or in an institution. An "aged-in" sample of 4,907 persons who became 65 between 1984 and 1989 and survived to 1989 were selected and screened for disability. In addition, 5,000 persons who were over 65 and had "screened out" as not chronically disabled in 1984 were rescreened. All nondisabled persons over age 75 (in 1989) from the 1984 sample were automatically screened in 1989.

The 1994 NLTCS was designed to precisely estimate disability and institutionalization rates for the 1982-1994 period. Persons in either 1982, 1984, or 1989 who received detailed interview in the community organization and survived to 1994 were reinterviewed where ever they might reside in the United States. An "aged-in" sample of approximately 4,500 persons who became 65 between 1989 and 1994 and survived to 1994 were selected and screened for disability. Further, 1994 a supplementary group of about 1,000 community interviews were conducted with non-disabled person (the health supplement). This group will be followed like all other detailed interview persons in subsequent years. These persons augment the approximately 900 persons already in the 1994 detailed survey who report no disability, result of the longitudinal data collection procedure. Further, a group of 540 persons, age 95+ were also drawn in 1994 and screened for disability. This supplemental sample greatly increases the precision associated with estimates of the extremely aged.

In each of the four surveys, large samples (N-2,000) of the oldest-old population (i.e., those 85 and over) are obtained. The survey data (i.e., detailed community and institutional interviews; screening surveys of nondisabled persons) were linked to Medicare service use files for most years 1982 to 1992. The linkage to Medicare enrollment files between 1982 and 1992 was 100%, i.e., there was complete follow-up of all cases (including survey nonrespondents) for Medicare eligibility (and for most years, detailed Part A and B use), mortality, and date of death. Medicare mortality records (and dates of death) were used to estimate mortality rates for the five-year period April 1, 1984 to April 1, 1989 and April 1, 1989 to April 1, 1994 and for the two-year period April 1, 1982 to April 1, 1984. Questions about disability and health were identical in 1982, 1984, 1989, and 1994 and found to be reliable and replicable between surveys (Wiener et al., 1990).

The data were adjusted to ensure consistency across survey years with respect to data collection methods, nonresponse adjustment, mortality, allocation and adjustment, and case classification. Any bias introduced by adjustments was designed to be "conservative," i.e., against showing positive health trends.

One adjustment was for persons automatically interviewed in 1984 and 1989 who were not disabled according to the 1982 screening interview criteria. This is obvious when no disability is reported. However, the ability to perform "heavy" housework was an IADL not included in the screening instrument. It was an item on the detailed household instrument. Thus, a person who reported only problems with heavy housework would not be disabled on the screen (and would not be recruited to the disabled subsample). Thus, we classified persons who automatically screened in, but who reported only problems with heavy housework as nondisabled. This required determining disability from individual items (and not check items).

Nonrespondents were allocated across disability and institutional residence categories. An alternate allocation of nonresponse (less than 5% in each year) might change the distribution of persons on disability level but would not affect the estimated size of the combined community disabled and institutionalized population. Thus, nonresponse allocation is unlikely to produce bias in the total population estimate.

In the new 1982, 1984, 1989, and 1994 NLTCS file linked to Medicare data 1982 to 1995, a number of ancillary analytic variables and special longitudinal weights were appended to aid the investigator in dealing with the complex cross temporal studies of the survey. This will become increasingly important as, with the 1994 file, meaningful cohort analyses can be conducted. The file is beginning to be lengthy enough to identify period effects. The number of deaths (i.e., about 17,000 1982 to 1995) is large enough that detailed analyses can be done. Further, over 12 years spanned by the four surveys a total of 35,848 distinct individuals were followed from and linked to Medicare records. The 1994 survey file will be added to the public use linked file during 1996. Future waves are planned for 1999 and 2004.

RACE AND ETHNICITY:

The following questions were used to collect racial and ethnic information in the NLTCS:

  • Is respondent of Spanish/Hispanic origin? Yes/No
  • What is respondent's race? White, Black, American Indian/Aleut/Eskimo, Asian or Pacific Islander (Japanese, Chinese, Filipino, Korean, Asian Indian, Vietnamese, Hawaiian, Guamanian, Samoan, other Asian), Other

STATUS: Ongoing

HOW TO ACCESS DATA:

The 1982, 1984, 1989, 1994 National Long Term Care Survey Public Use Files are now available on CD-ROM. You may request this CD-ROM by printing out the data request letter from the NLTCS Website: http://cds.duke.edu/ This letter must be notarized before you send it.

Continuously linked Medicare data (1981 thru 1994) for the persons in the 1982, 1984, 1989 and 1994 National Long Term Care Surveys have been placed on CD-ROM. This is the data that was previously available from the Center for Demographic Studies. All Records have a unique sequence number to link to the National Long Term Care Survey respondents. Documentation for all years has been included on the CD for ease of use.

This CD may be obtained through the Health Care Finance Administration at (410) 786-0161.

CONTACT NAME AND PHONE NUMBER:

Larry Corder Richard Suzman

Center for Demographic Statistics National Institute on Aging

Duke University 7201 Wisconsin Avenue

2117 Campus Drive, Box 90408 Bethesda, MD 20892

Durham, NC 27708-0408 Phone: (301) 496-3138

Phone: (919) 684-6126 Fax: (301) 402-0051

Fax (919) 681-3861


TITLE: Longitudinal Studies of Aging

ACRONYM: LSOA

AGENCY/PROGRAM: National Institute on Aging/National Center for Health Statistics

DESCRIPTION:

Purpose -- The LSOA is a multi-cohort study designed to measure change in health status, health-related behaviors and health care, and the causes and consequences of these changes within and across two cohorts of elderly Americans. A second objective of the study is to provide a mechanism for monitoring the impact of proposed changes in Medicare and Medicaid and the accelerating shift toward managed care on the health status of the elderly and their patterns of health care utilization.

Description -- The LSOA is comprised of the following components:

- Supplement on Aging, 1984 (SOA)

- Longitudinal Studies of Aging, 1984-1990 (LSOA)

- Second Supplement on Aging, 1995 (SOA II)

- Second Longitudinal Studies of Aging, 1995, 1997 (LSOA II)

The Supplement on Aging (SOA) was conducted as part of the 1984 National Health Interview Survey (NHIS). The SOA was based on a national sample of 16,148 persons 55 years of age and older living in the community. Interviews for the SOA were conducted in person by the U.S. Census Bureau. The following topics were covered in the interview:

  • Housing characteristics
  • Family structure and living arrangements
  • Relationships and social contacts
  • Use of community services
  • Occupation and retirement (income sources)
  • Health conditions and impairments
  • Functional status, assistance with basic activities
  • Utilization of health services, nursing home stays
  • Health opinions

The 1984 SOA served as the baseline for the Longitudinal Study of Aging (LSOA), which followed all persons who were 70 years of age and over in 1984 through three follow up waves, in 1986, 1988 and 1990. All follow up interviews were conducted over the telephone by interviewers from the U.S. Census Bureau. The major focus of the follow up interviews was on functional status and changes that had occurred between interviews, although information was also collected on housing and living arrangements, contact with children, utilization of health services and nursing home stays, health insurance coverage, and income. The interview data is augmented by linkage to Medicare records, the National Death Index, and multiple cause-of-death records.

The Second Supplement on Aging (SOA II) was conducted as part of the 1994 NHIS. Interviews for the SOA II were conducted during a follow up visit with NHIS respondents between October 1994 and March 1996. The SOA II sample is comprised of approximately 9,447 persons who had participated in the 1994 NHIS and had turned 70 years of age by the time of the SOA II interview. SOA II serves as a comparison cohort to the 1984 SOA, and most of the questions from the SOA were repeated in the SOA II. Topics new to the SOA II include:

  • Use of assistive devices and medical implants
  • Health conditions and impairments
  • Health behaviors
  • Transportation
  • Functional status, assistance with basic activities, unmet needs (expanded)
  • Utilization of health services, nursing home stays (expanded)

The SOA II provides the baseline for the Second Longitudinal Study of Aging (LSOA II). Data collection for the first follow up wave of the LSOA II is scheduled to begin in May, 1997; future follow up waves will be conducted at two-year intervals if funding is secured. Many of the questions included in the SOA II are repeated in the follow up interview. In addition, the follow up questionnaire includes more extensive information on cognitive and affective functioning, health care utilization, unmet health care needs, and health insurance. The amount of health service utilization data provided through linkage to Medicare records will be greatly expanded over what was available in the first LSOA.

RACE AND ETHNICITY:

Racial Categories are: American Indian/Alaskan Native; Asian or Pacific Islander (Chinese, Filipino, Hawaiian, Korean, Vietnamese, Japanese, Asian Indian, Samoan, Guamanian, Other Asian/Pacific Islander); Black; White; and Other.

STATUS: Ongoing

HOW TO ACCESS DATA:

Public Use Data Files -- Data for the full SOA sample of persons 55 years of age and older is available on magnetic tape. The tape can be purchased for $275.00 from the Division of Health Interview Statistics at the National Center for Health Statistics. Members of the Inter-University Consortium for Political and Social Research (ICPSR) may obtain the data free of charge through ICPSR.

The complete set of data from all four interview waves of the LSOA and corresponding documentation is available on magnetic tape and CD-ROM. The CD-ROM, released in November 1993, contains an instruction guide, help screens, complete documentation, the LSOA methodological report and search software. The person file contains all of the data from the four interviews, the most recent information on the fact and date of death, and an indicator of whether there are records of Medicare-covered health service use. Two other files on the same CD-ROM contain information from the Medicare-covered inpatient and outpatient records.

One additional update was released on floppy disk in 1996. This provided information on causes of death and fact of death through 1992.

These data and documentation can be obtained from the Data Dissemination Branch, National Center for Health Statistics, ICPSR, and the National Technical Information Service in Springfield, VA. ICPSR also releases the data on 3.5" disks for those who want to work on personal instead of mainframe computers. The LSOA CD-ROM is also available from the National Institute on Aging.

The study is fully documented in three Vital and Health Statistics monographs: Series 1, Numbers 18, 21, 28.

A public use data tape is scheduled to be released by the National Center for Health Statistics in June 1997. The first release will be on magnetic tape, with a CD-ROM release to follow in early fall 1997. Updated files will be released as record matches from the National Death Index and Medicare utilization records become available.

SOA II, wave 2 data collection will extend through early 1998; release of Wave 2 data is anticipated to occur in March 1999.

CONTACT NAME AND PHONE NUMBER:

Julie Dawson Weeks

Ph: 301-436-5979 x178

Fax: 301-436-8459

jad3@cdc.gov