Abstract
OBJECTIVE: Antidepressant use increased substantially among older adults with the introduction of the new-generation medications such as the selective serotonin reuptake inhibitors. The authors analyzed data from two follow-up intervals—1986–1987 to 1989–1990 (interval 1) and 1992–1993 to 1996–1997 (interval 2)—from a community-based cohort of 4,162 older adults to determine predictors of future antidepressant use. METHOD: Information on antidepressant use, demographic and health characteristics, and categories of depressive symptoms—positive affect, negative affect, somatic complaints, and interpersonal problems—were obtained. Logistic regression was used to control simultaneously for multiple variables predicting antidepressant use during the two intervals. Repeated-measures logistic regression (with generalized estimating equations) was employed to model the probability of antidepressant use, with adjustment for the effect of time. RESULTS: Prior antidepressant use and white race were strong predictors of future use during both intervals. Negative affect was the only additional significant predictor of use during interval 1. In contrast, low positive affect scores, cognitive impairment, and poorer health were additional significant predictors during interval 2. In a repeated-measures model, race, prior antidepressant use, poor health, low positive affect scores, and somatic complaints varied as predictors over time. Negative affect and cognitive impairment were consistent predictors over time. CONCLUSIONS: The predictors of antidepressant use by older adults changed over time, with health-related measures of quality of life, such as positive affect, health status, and somatic complaints, becoming more prominent as predictors of use.
