Estimates of prospective change in self-rated health in older people were biased owing to potential recalibration response shift

H. Galenkamp, M. Huisman, A.J. Braam, D.J.H. Deeg

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Objective: Evidence shows that self-rated health (SRH) remains remarkably stable during aging. Individuals may change their conceptualization of health or revise their standard of good health when facing health decline. Although this "response shift" phenomenon is potentially beneficial to the individual, it also challenges comparison of SRH assessments over time. The present study investigates this response shift. Study Design and Setting: Data come from two waves (T1 and T2) of the Longitudinal Aging Study Amsterdam (N: 1,274; age: 55-89 years; mean follow-up: 3.6 years). Linear regression models were used for predicting SRH at T1 and T2. To capture changes in individual health standards, we administered a then-test at T2, asking respondents to retrospectively rate their health at T1 again. Results: No support was found for a changed conceptualization of SRH after health decline: predictive models for SRH at T1 and T2 were not significantly different. In the subgroup that reported identical SRH at T1 and T2, participants who experienced incident diseases were three times more likely to retrospectively overrate health at T1 with the then-test, suggesting that they had a lowered health standard. Conclusion: Older people's concept of health remains stable when they encounter significant health problems, but they potentially lower their standard of good health over time. © 2012 Elsevier Inc. All rights reserved.
Original languageEnglish
Pages (from-to)978-988
JournalJournal of Clinical Epidemiology
Volume65
Issue number9
DOIs
Publication statusPublished - 2012

Fingerprint

Dive into the research topics of 'Estimates of prospective change in self-rated health in older people were biased owing to potential recalibration response shift'. Together they form a unique fingerprint.

Cite this