OBJECTIVES: To investigate utilization and persistence in Zoloft patients versus those on other antidepressants. METHODS: Using a 20% sample of California Medicaid claims data from 1999 to 2003, patients on antidepressants were identified and tracked for 12 months from the first fill date of an antidepressant medication. Utilization patterns (discontinuation, gap, switch, and augmentation events), medication possession ratio (MPR), days covered (DC), and persistency (proportion of patients on initial medication at least 80% of the time) were analyzed. Differences were tested using normal approximation with a two-sample test. Odds ratios were computed with respect to Zoloft and tested using logistic regression models with propensity scores. RESULTS: A total of 1,403 patients were initiated on Celexa, 1,309 on Effexor, 10,758 on Paxil, 4,631 on Prozac, and 2429 on Zoloft. Proportionally, more Zoloft patients were event-free than Effexor or Paxil patients (p < .05), but less so than Celexa. Adjusted odds ratios suggest that patients initiated on Zoloft were more likely to persist with their medication than patients on Effexor, Paxil or Prozac (OR = 0.824, 0.732, 0.762, respectively; p < .05). Differences between Zoloft and Celexa (OR = 1.040) were not statistically significant. Zoloft had a higher average MPR than Effexor, Paxil or Prozac, but slightly lower than Celexa (p = 0.0259). All cohorts experienced a decline in days covered (DC) on Day 31, 61, 91, and 181. At the end of follow-up, 29.6% of patients initiated on Zoloft were still taking the medication, which was significantly higher than patients on Effexor (24.3%), Paxil (24.9%) or Prozac (23.7%), but slightly lower than patients on Celexa (32.6%, p=0.0502). CONCLUSIONS: Patients initiated on SSRI's continue to have relatively fast declines in medication adherence and persistence within the recommended timeline for therapy.