OBJECTIVE: To determine if total out-of-pocket prescription drug (OOPPD) expenditures by the community dwelling elderly differ according to longitudinal patterns depression. METHODS: Secondary data analyses were performed using the population-based study of Assets and Health Dynamics (AHEAD) of the Oldest Old (adults > 65 years). Depression was considered present if 4 or more depressive symptoms were reported on the modified Center for Epidemiological Studies-Depression Scale (CES-D). Three survey waves in 1995, 1998 and 2000 allowed depression to be characterized as persistent, emergent, remittent, and recurrent. ANOVA and regression techniques were used to estimate association between mean total monthly OOPPD expenditures in 2000 based on depression pattern. RESULTS: Of the 7027 elderly residents interviewed in 1995, 19% were lost to follow-up at 2000. More than 50% of respondents never experienced significant depressive symptoms. Mean (SD) OOPPD expenditures for each pattern of depression were: $221(1203) for recurrent (n = 63); $106(598) for emergent (n = 324); $86(227) for remittent (n = 210); $78(138) for persistent (n = 179); and $70($168) for never depressed (n = 3290). Only those with recurrent depression had significantly higher OOPPD expenditures compared to those without depression (ANOVA, p <0.05). After adjusting for age, gender, and comorbidity, recurrent and emergent patterns of depression were associated with significantly higher mean monthly OOPPD expenditures compared to those without depression, while persistent and remittent depression were not. CONCLUSIONS: The community dwelling elderly with fluctuating patterns of depression appear to pay more in monthly OOPPD expenditures than the elderly with stable patterns, including chronic depression. Upon testing of the robustness of the results using non-parametric and longitudinal random-effects models, further investigation into the burden of illness based on longitudinal patterns of depression is recommended.