Hospitals, health care systems, and policy makers are striving to seek ways to provide cost-effective care for patients with major depressive disorder. OBJECTIVES: The first objective is to investigate the differences between the newer antidepressants regarding treatment completion, average daily dose, dosage titration, switching, and augmentation behavior. The second objective is to compare direct health service expenditures related to the treatment of depression. The inquiry is guided by the following question: Is there a significant difference between antidepressants in regard to overall health care service expenditures for the treatment of depression? METHODS: Retrospective archival data from computerized claims records of a large managed care organization were analyzed. Treatment completion was defined as receiving at least 180 days of therapy at a minimum therapeutic dose as defined by the AHCPR guidelines for detection, diagnosis, and treatment of depression. Patients were included in the analysis if (1) they had an ICD-9 diagnosis code for depression or if (2) they received an antidepressant prescription. Patients were excluded if (1) they were less than 18 years of age, (2) they had a diagnosis indicating schizophrenia or bipolar depression, (3) there were not at least 6 months of follow up data available, or (4) they were ineligible for coverage by the plan. RESULTS: Patients initiated on fluoxetine were more likely to complete therapy than those on paroxetine, sertraline, nefazodone, or venlafaxine (n = 65,792; p < .01). These differences narrowed over time. Results regarding overall health care utilization related to each antidepressant will be presented. CONCLUSIONS: Based on this sample of patients, it appears that patients initiated on fluoxetine are more likely to complete therapy when compared to the other antidepressants.