Describing Death in America: What We Need to Know: Executive Summary (2003)

Institute for Social Research, University of Michigan

Robert J. Willis

to explain the antecedents and consequences of retirement; to examine the relationship between health, income, and wealth over time; to examine life cycle patterns of wealth accumulation and consumption; to monitor work disability; and to examine how the mix and distribution of economic, family and program resources affect key outcomes, including retirement, savings, health declines and institutionalization

health and cognitive conditions, status; retirement plans and perspectives; attitudes, preferences, expectations, and subjective probabilities; family structure and trans

fers; employment status and job history; job demands and requirements; disability, demographic background, housing, income and net worth, health insurance and pension plans

national panel study

in home, face-to-face in 1992 for the 1931-1941 birth cohort and in 1998 for newly added 1924-1930 and 1942-1947 birth cohorts

follow-ups by telephone every second year, with proxy interviews after death

merged with AHEAD in 1998

100% oversamples of Hispanics, Blacks, and Florida residents

Employer Pension Study (1993, 1999); National Death Index, Social Security Administration

publicly available data on www.umich.edu/~hrswww includes HRS Wave 1,2,3 (soon 4)

Longitudinal Study of Aging

National Health Interview Survey (NHIS) 1984 Supplement on Aging

sponsor

PI

to measure transitions in functional status and living arrangements for a national representative cohort of Americans who were aged 70 and older in 1984 and lived in the community; a second survey was field on a cohort derived from the 1994 NHIS

functioning, family structure, housing and living arrangements, health care, economic and retirement indicators, vital status

longitudinal panel

1984 baseline personal interviews called the Supplement on Aging (SOA) were conducted with 16,148 subjects; The LSOA followed 7,527 of these subjects who were 70+

LSOA reinterviews in 1986 (5,151), 1988 (7,541), 1990 (5,978); LSOA II reinterviews were conducted in 1997-1998 and 1999-2000

used personal interviews in 1984; telephone interviews and mailed questionnaires in 1986-1990; plus record linkage

the LSOA followed those participants in the Supplement on Aging (SOA) to the 1984 NHIS who were 70 years of age and over. The SOA was a systematic one-half sample of people 55-64 in the 1984 NHIS and all people 65+; In 1986, the LSOA interviewed all SOA households with participants 80+; all households with Hispanic or black persons and their relatives 70-79; and half of households with whites aged 70-79. In 1988 and 1990, all participants; The LSOA II used the same methodology

coding information available

National Death Index and Medicare administrative records

National Mortality Followback Survey

parent study

National Center for Health Statistics

PI

to supplement information from death certificates in the vital statistics file with information on important characteristics of the decedent which affect mortality. Objectives vary with each survey round

include use of health services prior to death, socioeconomic status, aspects of life style, health care utilization prior to death, and other factors that affect when and how death occurs; 1993 topics: demographics, SES, manner of death, firearm related injury, motor vehicle/driving behavior, problem behaviors, use of alcohol, drugs, and tobacco, medical examiner/coroner abstract file

multicomponent stratified list-based survey

one time survey of next of kin, with linked data obtained from health care facilities used in the last year of life

n/a

complexity of questionnaire necessitated telephone or in person interviews; 83% response rate; 1,000 items focused in 23 domains

drawn from 1993 Current Mortality Sample (10% systematic random sample of death certificates): sampling strata: age, race, gender, and cause of death; 45% of sample se

lected with certainty; oversampling of blacks, females, decedents under age 35 and over age 99 (South Dakota declined to participate)

22,957

data tapes can be purchased through NTIS: four linked data files (death certificate file; proxy respondent survey questionnaire; nursing home, hospital, and hospice questionnaire; and facility abstract record)

Mortality Statistics Branch, NCHS, Hyattsville, MD (301)458-4666

National Health and Nutrition Examination

NHANES

to assess the health and nutritional status of adults and children in the United States; to estimate disease prevalence, awareness, treatment and control; to monitor trends in risk behaviors; to study the relationship between diet, nutrition, and health; to establish a national probability sample of genetic material

demographics, acculturation, SES and education; diet, dietary supplements, and food security; medical conditions, health care utilization, health insurance and prescription drugs used; physical activity, fitness and function; alcohol, drug, and tobacco use; pesticide use;

includes medical examination and special substudy collecting hair

annual survey

“snapshot” only, but now linked to Medicare and NDI for longitudinal and historical purposes

linked to the NDI and Medicare

now linked to the NHIS at the Primary Sampling Unit (PSU); linked to NHIS with regard to questionnaire content; household screening interview; detailed household interview, individual physical exam and health and dietary interviews in mobile examination centers (or at home if necessary)

now each year’s data are representative

link to NDI; passive mortality study 1976-1980; cohort followup study (NHEFS)

National Death Index, Medicare

data can be ordered from NCHS or downloaded from Web (lag time there)

Raynard Kington, Director NHANES, NCHS, 6525 Belcrest Road, Room 1000, Hyattsville, MD 20782

Medical Expenditure Panel Survey

National Medical Care Expenditure Survey (NMCES or NMES)

AHRQ & NCHS

PI

provide comprehensive data that estimate the level and distribution of health care use and expenditures, monitor the dynamics of the health care delivery and insurance systems, and assess health care policy implications

health care use, expenditures, sources of payment, insurance coverage, health care status, and disability

overlapping panel design

baseline

followup

First two surveys (1977 and 1987) were studies of 14,500 households; 1996 (n = 10,000 households) changed to begin phase in of overlapping panel; will evolve to be 4,000 households brought in each year and followed for 2 years (so 10,000 households in any 2 1/2 year frame)

linked to NHIS PSU

number of decedents

linked to NHIS

PUFS on Web

MEPSPD@ahrq.gov

Panel Study of Income Dynamics

NIA, NSF, DHHS/ASPE

University of Michigan Institute for Social Research Sandra Hofferth and Frank Russell

to provide economic and demographic data at the family and individual level, as well as household

income sources and amounts, employment, family composition changes, childbirth and marriage histories, and residential location; NIA funded supplement on health; some sociological and psychological data in some waves

every other year interview

Initial sample = 2,930 households and 1,872 low income families; households formed by earlier panel members are included in subsequent waves; 1996 sample = 8,885 then reduced by suspending some low income and added new post-1968 immigrant sample

annual interviews conducted 1968-1997; every other year interviewing began 1999

long-term cumulative response rate was approximately 60% as of 1994

8,895 households in 1997 head of household interviewed

sample notes

main data files updated with each wave; www.umich.edu/~psid

Center for Medicare and Medicaid Services

agencies submit to CMS

form the core set of data items of a comprehensive assessment of an adult home care patient and form the basis for measuring patient outcomes for purposes of quality improvement

demographics, finances, current illness, therapies, living arrangements, physical assessments, symptoms, function

longitudinal

start of care assessment

follow-up assessment

either incorporated into agency’s data collection forms or can be stand alone data collection

% proxy

coding information available

contacts

Nursing Home Minimum Data Set (Resident)

MDS

IOM Committee on Nursing Home Regulation

CMS (formerly HCFA)

to provide a comprehensive assessment of nursing home residents in a standardized format to transmit to HCFA

demographics, customary routine, cognitive patterns, comunication, mood and behavior, physical functioning, continence, disease, nutrition, medications, procedures, therapy

longitudinal

on admission

quarterly or with significant change in status

research indicates that staff can produce research-quality data, but in practice facilities differ in their commitment to ensuring that staff are trained and adhere to assessment protocols

% proxy

all nursing home residents

can be identified by discharge assessment

availability of data

contacts

Surveillance, Edpidemiology, and End Results Program

SEER

National Cancer Institute

PI

to provide information on cancer incidence and survival in the United States

patient demographics, primary tumor site, morphology, stage at diagnosis, first course of treatment, and follow-up for vital status

registry of all tumors reported in a geographic area

standard for case ascertainment is 98%

followup

in additon to registry information, there is an ongoing program of special studies that collect information through surveys, interviews, record abstraction, and biological materials

geographic areas were selected for inclusion based on their ability to operate and maintain a high quality reporting system

number of decedents

links

updated annually and provided in print and electronic formats

contacts

National Home and Hospice Care Survey

NHHCS

National Health Care Survey

NCHS

PI

to document the availability and utilization of home and hospice care

agency characteristics: ownership, affiliation, geographic region, location, certification status; Patient characteristics: age, race, marital status, living arrangements, caregiver status, services received, service providers, reason for discharge

cross-sectional survey

followup

data collected through personal interviews with administrators and staff

Stratified, two stage probability design (agency type, MSA, region, ownership, certification) In 1996, 13,500 agencies provided services to 2,486,800 patients, with 8,168,900 discharges; in 1996, 1,800 agencies provided hospice care to 59,400 current patients and had 393,200 discharges (322,200 or 82% = deaths)

can be linked to OASIS

public use data files available through 1998

Dr. Barbara Haupt

NCHS

PI

provide information on nursing homes from the perspective of the provider of services and from the recipient

facilities: size, ownership, certification, occupancy rate, number of days of care provided, expenses; residents: demographic characteristics, health status, and services received

cross-sectional survey

1973-1974, 1977, 1985, 1995, 1997, 1999; consist of facility files, discharge files, resident files, and staff files

data obtained through personal interviews with administrators and staff (note: information about residents provided by staff member familiar with care provided to resident)

up to six current residents and up to six charges are selected for individual level interviews

coding information available

public use data files

National Health Care Survey

NCHS

PI

provide information on the characteristics of inpatients discharged from non-federal short-stay hospitals

age, sex, race, ethnicity, marital status, expected sources of payment, admission and discharge dates, discharge status

note: cannot use individual level data because individuals could have multiple discharges in one year and thus be sampled more than once.

500 hospitals and 270,000 inpatient beds

Excludes federal, military, VA, prison hospitals, and hospitals with fewer than six beds. All hospitals with 1,000 or more beds are sampled with certainty. Non-certain hospitals selected randomly from PSU sampling units of NHIS

in 1996, deaths accounted for 3% of discharges

electronic files from FTP server

Healthcare Cost and Utilization Project abbrev HCUP

Agency for Healthcare Research and Quality

PI

to build a multi-state health care data system about discharges from hospital

primary and secondary diagnoses, primary and secondary procedures, admission and discharge status, demographics, expected payment source, total charges, length of stay

coordinated all payer databases

22 states participate, approximating a 20% stratified sample of U.S. community hospitals. National Inpatient Sample = 1,000 hospitals

National Vital Statistics Death Certificates

NVS

National Center for Health Statistics

PI

to provide national mortality statistics

immediate cause of death, intermediate cause of death, underlying cause of death, other coexisting conditions; age, race, marital status, educational attainment, occupation

continuous reporting by states to Bureau of Census

Medicare Claims

Health Care Financing Administration

to track billing for health care services provided under the Medicare benefit to those eligible by age (65 years or older), disability, or end stage renal disease

age, gender, race, geographic region, utilization to include hospitalizations, hospice, services, stays in skilled nursing facilities, outpatient, physician, and home health visits

continuous

approximately 1,700 decedents per year

linked to Social Security Administration records for purposes of capturing decedents

Medicare Continous History File is available on 5% of beneficiaries each year

Medicare Current Beneficiary Survey

Medicare Claims

HCFA

determine expenditures and sources of payment for all health care services used by Medicare beneficiaries (including noncovered services) and to trace changes in health status and spending over time

use of health services, expenditures, insurance coverage, sources of payment, health status and functioning, and a variety of demographic and behavioral information, such as income, assets, living arrangements, family supports, and access to medical care

four-year rotating panel

6,000 new beneficiaires are added each year

interviewed every four months

originally not a rotating panel, but problems with follow up

107 primary geographic sampling units, with oversampling in areas with high growth in population of elders

700 per year

coding information available

Medicare claims and administrative data (in turn linked to Social Security Administration)

purchased through HCFA

Director, Enterprises Databases Group, Office of Information and Systems, HCFA, 7500 Security Boulevard, Baltimore, MD 21244-1850; 410-786-3690

companion to the National Long-Term Care Channeling Demonstration (DHHS)

Office of the Assistant Secretary for Planning and Evaluation (ASPE)

Census administered; Duke University

Kenneth Manton

to learn more about health, functioning, and social and economic factors among community-based disabled. The purpose of the Next of Kin Survey on decedents is to estimate the total long-term care costs and the extent of spend-down to qualify for Medicaid

prevalence and patterns of functional limitations, both physical and cognitive; medical conditions and recent medical problems; use of health care services; kinds and

amounts of formal and informal long-term care services used; demographic characteristics; public and private expenditures for health care services; and housing and neighborhood characteristics

list-based, longitudinal, panel design

In 1982, a random sample of 35,008 Medicare-eligible people >65 yr were screened for disability, yielding 6,393 chronically disabled community residents who qualified for in-person interviews. 1,925 interviews were completed with informal caregivers of disabled elders (of 2,349 identified)

In 1984, all survivors were contacted to be reinterviewed. In addition, a sample of community residents who were not disabled in 1982 was added and a new sample of people who reached 65 years of age between 1982 and 1984 was added. This process was repeated in 1989, 1994, and 1999

Overall response rates ranged from 97.1% in 1982 to 95.4% in 1989; informal care supplement in 1982, 1989, 1999; next of kin decedents 1994; nursing home resident follow-back 1989; nutritional status 1994

55,000 screen interviews; 21,000 community detailed interviews, and 5,000 detailed institutional interviews in total for the 1982-1993 surveys; 1982: Medicare HISKEW file stratified geographically into LTC Primary Sampling Units, which were grouped into 173 long term care strata; one LTC PSU selected from each stratum

17,000 deaths identified from Medicare admin records (1982-1996): Next of Kin Supplement

linked to Medicare Part A and B most years; Medicare mortality records

public use files available from Duke University (http://cds.duke.edu)

National Health and Nutrition Examination NHFES

NHANES I

NCHS/NIA

PI

to investigate the association between factors measured at baseline and the development of specific health conditions. The three major objects are to study morbidity and mortality associated with suspected risk factors, changes over time in participant characteristics, and the natural history of chronic disease and functional impairments

self-reported medical conditions, activities of daily living, health and nutrition habits and weight, physical examinations, laboratory tests, facility medical records, death certificates

longitudinal follow up to a multistage, stratified sample

NHANES I data (1971-1975) included subsample that received detailed study and an augmentation survey 1974-1975.

In 1982-1984, interviews were conducted with subject or proxy for 84.8% of the eligible 14,407 cohort (11,361 alive, 2,022 deceased, 1,024 not traced); in 1986, those over 55 years old (n = 3,980) included 3,132 alive, 635 dead, and 213 not traced. In 1987, all ages cohort = 11,750 with 10,463 alive, 555 dead, and 732 not traced. In 1992, n = 11,195, with 8,687 still alive, 1,392 dead, 1,116 not traced. Interviews in 1982-1984 were conducted in person; remaining years by phone with no physical assessment

four follow up waves: 1982, 1986, 1987, and 1992

12,220 (84.8%) interviewed (or 91.3% of those traced); the NHANES I adults (25-74) civilian noninstitutionalized population who completed a medical examination; included oversampling of persons living in poverty areas, women of childbearing age, and elderly (65+); a subsample of 6,913 were examined in greater depth and asked additional questions; an augmentation survey was conducted in 1974-1975

4,604 decedents: death certificates for 98%; interviews in 1992 with 1,130 (81.2%) of 1,392 decedents

a study ID number links any NHANES files

Mortality Data tape includes death certificate info on all decedents

National Health Interview Survey

NCHS

U.S. Bureau of the Census

to produce statistics on disease, injury, impairment, disability, and related health topics on a uniform basis for the nation

cross sectional (area based household survey)

About 40,000 households are interviewed in person each year (800 representative households selected each week to avoid seasonal bias), representing over 100,000 people Responsible adult interviewed and asked about household members; core set of questions plus varying supplemental questions

substantially redesigned in 1997; changes to capture outpatient surgery, services provided by other than MD providers; mental health services; improvements in measuring health status; symptoms instead of naming conditions; family-level data

area-based households; U.S. civilian, noninstitutionalized population; response rates 94.9% to 96.7%; coordinated with other Census surveys to avoid double survey of same household; Westat, Inc., studied sampling options for 1995 redesign

32,431 decedents 1986-1994 (10,407 died within two years of interview); linked to National Death Index

National Death Index system beginning survey year 1986; vital status known through 1997

Public use files on NCHS Website

Asset and Health Dynamics of the Oldest Old

AHEAD

to provide an understanding of the implications of health dynamics in old age for transtitions in economic well-being, changes in family and martial status, and for reliance on public and private support systems

demographic info; health, cognition, family structure and transfers, use and cost of health services, job status, in

come, net worth, subjective expectations, and insurance; these included physical tasks and personal care activities

Wave 1 conducted in 1993; response rate of 80%; those 70-79 were generally interviewed by telephone; those 80+ were generally interviewed in person

Wave 2 took place in 1995 and 1996; reinterview rate of 95%; combined HRS/AHEAD in 1998

initially noninstitutionalized persons born in 1923 or earlier (aged 70 years or older); Wave 1 = 8,222 respondents (and spouses). Wave 2 interviews were conducted with 7,039 of the 8,222 Wave I respondents and approximately 775 exit interviews with next of kin; multistage sample plan, with oversampling of Blacks, Hispanics, and Florida residents; dual sampliing frame using 81 of the 93 PSUs in HRS, then deleting half and replacing with selections from the HCFA-EDB file; compared to census files, the sample reflected population except for females aged 80 and over (more likely to be in nursing homes)

775 from Wave 1

Medicare (80% gave permission); Medicaid (fewer gave permission)