Modeling the Effect of Disease on Functional Change in the HRS

Dawn Alley, University of Pennsylvania
Beth J. Soldo, University of Pennsylvania

Previous research on functional decline has largely focused on the onset of disability or short-term changes in ADLs. The purpose of this paper was to model functional change longitudinally using mixed effects models and to estimate the effect of existing disease and disease onset on functional change. Participants came from five waves (1998-2004) of the combined Health and Retirement Survey (HRS) and Assets and Healthy Dynamics among the Oldest Old (AHEAD) survey, providing a nationally representative longitudinal study of Americans aged 50 and over living in the community (N=17,318). Functional change was nonlinear and varied by sex, race, education, and income. Existing chronic conditions and onset of disease were both related to functional decline; arthritis, heart disease, and lung disease caused the largest drops in function.