OBJECTIVES: Assess the relative performance of new antidepressants in terms of compliance, drug switching and cost. METHODS: A total of 246,116 episodes of antidepressant therapy were abstracted from the fee-for-service paid claims file of the California Medicaid (Medi-Cal) program for year 1999-2002. Episodes of treatment using sertraline, paraxetine, fluoxetine, citalopram, bupropion or venlafaxine were selected for study. Data for each episode cover six-months prior and 12 months post-treatment. RESULTS: Most episodes of antidepressant therapy are for patients who restarted therapy using the same medication (53.2%) or on a second antidepressant (18.9%), followed by augmentation episodes (14.8%) and switching episodes with no break in treatment (13.6%). Antidepressants open formulary access (paraxetine and fluoxetine) are more frequently used in restarting episodes using the same medication. Patients who restart therapy display better adherence and lower switching rates than patients who switch or augment therapy, thus biasing upward the overall treatment compliance performance of open-access antidepressants. Unadjusted data for restart and delay switching episodes suggest that patients treated with sertraline and venlafaxine achieve longer duration of therapy than patients treated other drugs with other drugs. Differences are relatively minor across all of the antidepressants studied with the exception of bupropion. CONCLUSIONS: Unadjusted results indicate little difference in patient outcomes across alternative antidepressants. However, physicians may be selectively prescribing drugs to those sub-populations where each drug may have a clinical advantage. For future research, propensity scoring methods will be used to investigate if clinicians have been successful in prescribing alternative medications for those sub-populations in which each product achieves superior outcomes.