OBJECTIVE: The purpose of this analysis was to describe and compare patterns of drug selection, dosing, and utilization among depressed patients treated with SSRIs (citalopram, fluoxetine, paroxetine, and sertraline) and venlafaxine in US managed care. METHODS: Symmetry Health Data Systems' Episode Treatment Group methodology was applied to PharMetrics' Integrated Outcomes Database to identify adult subjects (>18 years.) diagnosed with depressive disorder and treated with one of the study medications between January 1, 1998 and June 30, 1999. Prescription claims records were used to identify patterns of drug selection, dosing, and utilization (medication persistence, augmentation, switching, and time to change) for each medication cohort. Descriptive statistics were used to characterize drug selection and dosing patterns, and parametric and nonparametric methods (ANOVA and chi-square) were used to compare utilization indicators across study groups. RESULTS: Twelve thousand twenty six patients met inclusion criteria. The study cohorts were demographically similar. Index antidepressant drug prescriptions were for: citalopram (6.2%); fluoxetine (24.6%); paroxetine (29.5%); sertraline (33.4%), and venlafaxine (6.2%). Citalopram patients more often started (86.7%) and ended (78.4%) on the lowest available dose. Each cohort showed similar patterns of dosage titration/adjustment. The mean number of prescriptions was similar across cohorts; citalopram-treated patients had a significantly higher rate of persistence (mean=118.3 days) than those treated with paroxetine (106.5), sertraline (108.8), and venlafaxine (108.2) cohorts (p < .001). Augmentation and switching rates were similar across cohorts; citalopram patients whose therapy changed had a significantly longer time to change (mean = 80.9 days) than fluoxetine (76.8 days, p = .02), paroxetine (72.1 days, p < .001), sertraline (69.6 days, p < .001), or venlafaxine (66.8 days, p < .001) patients. CONCLUSIONS: Of the antidepressants evaluated, citalopram was most frequently prescribed at the lowest available dose. Persistence rates were also highest with citalopram. Clinicians and payers should consider differences in dosing and persistence to provide optimal care and benefits for depressed patients.