This research used paid claims data to investigate the likelihood that patients achieved an adequate course of antidepressant drug therapy and the impact of completed therapy on health care costs. Completed therapy was defined as six or more months of uninterrupted therapy at an adequate dose as determined by AHCPR treatment guidelines. Apparent average daily dose for each prescription filled was calculated from data on the prescription paid claim and allowances were made for titration of dose to therapeutic levels and changes in antidepressant therapy. A total of 1648 new episodes of antidepressant therapy were identified for analysis. The likelihood of achieving an adequate course of antidepressant therapy was 22%. Completion rates varied significantly across antidepressants with fluoxetine achieving the highest completion rate at nearly 51%. Total health care costs were significantly lower for patients who completed therapy (−$1487; P = .0487) due primarily to lower ambulatory care costs (−$1296; P = .0110). Fluoxetine was the only antidepressant therapy which exhibited significantly lower total health care cost per patient relative to the older tricyclic antidepressants (−$3524; P = .0024). The total costs of treating depression in the ambulatory setting were found to vary widely across alternative antidepressants. Most of the cost-savings associated with fluoxetine use were associated with the increased likelihood of completed therapy. Further research is needed to verify if fluoxetine achieves better rates of completed therapy relative to other SSRI antidepressants using data from other settings.