Trajectories of Physical Disability and Distress: The Longitudinal Effects of Social Support

Miles G. Taylor, Duke University
Scott M. Lynch, Princeton University

Using multiple waves of data from the Duke Established Populations for Epidemiologic Studies of the Elderly (EPESE), we investigate the relationship between trajectories of physical impairment, received and perceived social support, and depressive symptoms. We argue that impairment, depressive symptoms, and social support can be considered growth processes in late adulthood. Growth in impairment is expected to produce growth in depressive symptoms, while social support growth is expected to mediate the effect of impairment. The research makes several contributions. First, we demonstrate how Bayesian random coefficients models can be used to model growth and to handle missing data and unequal timing-of-observations that structural equation-based growth curves cannot. Second, we find that growth in impairment predicts growth in depressive symptoms across age, and that growth in perceived support mediates the effect of growth in impairment on depressive symptom growth. Third, we find that this mediation does not occur for received support.