HRS - sporedata/researchdesigneR GitHub Wiki

General description

The Health and Retirement Study (HRS) surveys approximately 20,000 Americans over 50 years of age every two years. HRS - is a longitudinal study that began in 1992 to provide nationally representative data on the health and economic circumstances related to aging at individual and population levels. Because HRS has data on ZIP codes, it is possible to link it to geographic data. This offers the possibility to investigate a wide range of questions, from environmental health, to health policy.

37,000 individuals had responded to the questionnaires. 22,000 had Medicare records. 15,000 had genotyping profiles from the Illumina platform.

Overall, 11,000 individuals have all data available.

Apart from the “Core” HRS survey, physical health, biomarkers, genotyping, and psychosocial survey, the HRS questionnaires also include:

Off-Year studies
- Consumption and Activities Mail Survey (CAMS)
- Disability Vignette Survey (DVS)
- Health Care Mail Survey (HCMS)
- Health Care and Nutrition Study (HCNS)
- Health Survey
- Human Capital and Educational Expenses Mail Survey (HUMS)
- Internet Survey
- Life History Mail Survey (LHMS)
- Mail Survey
- Veterans Mail Survey
Health studies
- Aging, Demographics and Memory Study (ADAMS)
- Diabetes Study
- Harmonized Cognitive Assessment Protocol (HCAP)
- Health and Well-Being Study (HWB)
- Prescription Drug Study (PDS)

Variable categories

Demographics (for example, education; family history; language; marital history; military history; race; religion; residence)
Assets and Income;
Work (for example, current job, last job, job history);
Pension plans;
Health insurance;
Disability (for example, activities of Daily Living - ADL, Instrumental Activities of Daily Living - IADL);
Physical health (for example, height, weight, waist circumference, and blood pressure) and functioning (for example, grip strength, timed walk, balance tests, and a pulmonary function test);
Cognitive functioning (proxy cognition; self cognition);
Health care expenditures;
Genomic (for example, candidate gene and single nucleotide polymorphisms files, Polygenic Score Data - PGS);
Blood-based biomarker (for example, cholesterol, C-reactive protein, and hemoglobin A1C);
Psychosocial data (for example, well-being, personality, experienced discrimination, job characteristics).

Linkage to other datasets

The HRS can be linked to several complementary datasets:

Social Security earnings and benefits;
Medicare;
Veteran’s Administration;
National Death Index;
Employer-provided pension plan information;
Census Business Register.

The HRS can also be linked to other data sources via subjects' geographic information to study or adjust for regional characteristics or practice variations:

US Census;
Environmental variables:
- NO2 Land-Use Regression Model Estimate Data;
- PM 2.5 Fused Air Quality Surface Using Downscaling (FAQSD) Files;
- O3 Fused Air Quality Surface Using Downscaling (FAQSD) Files;
Health care policy and administration:
- Long-Term Care: Facts on Care in the US (LTCfocus)
- COVID-19 US State Policy Database;
- American Hospital Association;
Social determinants:
- Uniform Crime Reporting Program Data
- USDA Food Access Information;
- Dartmouth Atlas of Health Care;
- Centers for Disease Control;
- RAND Center for Population Health and Health Disparities;
- Decennial Census and American Community Survey (ACS);

Limitations

Some measures in HRS lack excellent psychometric properties such as high internal consistency, reliability, and construct validity. While a few of the measures may lack some content validity and don’t capture the full construct.
HRS content has changed from wave to wave.
The HRS dataset is very large, complex, and can be difficult to navigate.