Methods: The data used were from the Duke University Established Populations for Epidemiologic Studies of the Elderly (EPESE), a longitudinal survey of 4,162 community dwelling adults ages 65 or older. Depression was measured at baseline and at three, six, and ten years post baseline using the CES-D. Individuals were classified at each round as having CES-D depression (16+ on the CES-D), subthreshold depression (depressive symptoms not meeting the CES-D threshold), or nondepressed.

Results: Using categorical logistic regression controlling for sociodemographic variables, cognitive functioning, physical health status, self-perceived health, perceived social support, and the use of psychotropic medications, we found subthreshold depression, as well as CES-D depression, predicted the increased probability of depression at follow-up three to four years later relative to the nondepressed. Individuals with subthreshold depression were one half as likely to become CES-D depressed as were those with CES-D depression to stay depressed. Those with subthreshold depression were less likely to stay at the subthreshold level than were the CES-D depressed to become subthreshold depressed. Estimates across the four waves were aggregated using Markov models.

Significance: Compared to the nondepressed, subthreshold depression, like CES-D depression, is predictive of future depression in a cohort of older adults followed for ten years.