Article AbstractHerein we describe a retrospective intent-to-treat evaluation designed to compare the naturalcourse of antidepressant utilization and direct health service expenditures for the treatment of a singleepisode of major depression among patients enrolled in a multistate network-model health maintenanceorganization and initially prescribed either a tricyclic antidepressant (amitriptyline or nortriptyline)or the serotonin selective reuptake inhibitor (SSRI) fluoxetine. Patient-level paid-claims data forthe period July 1, 1988, through December 31, 1991, were abstracted. During the above time frame,fluoxetine was the only SSRI available in the United States. Patients prescribed amitriptyline weremore than three times as likely to require a change in antidepressant pharmacotherapy (OR = 3.27,95% CI = 2.31 to 5.49), while patients prescribed nortriptyline were nearly four times more likely tochange medication (OR = 3.82, 95% CI = 2.74 to 6.83) relative to patients initially prescribed fluoxetine.Consistent with our intent-to-treat design, all accrued health service expenditures were assignedto the pharmacotherapeutic option initially prescribed. Multivariate analyses revealed that initiationof antidepressant pharmacotherapy with amitriptyline resulted in a 25.7% increase in per capitadepression-related health service expenditures per year, while initiation of antidepressant pharmacotherapywith nortriptyline resulted in a 28.1% increase in per capita depression-related health serviceexpenditures per year relative to patients initially prescribed fluoxetine. A financial break-even pointwas achieved at the conclusion of Month 5, at which time all three intent-to-treat cohorts had comparablehealth service expenditures in total. From a financial perspective, results stemming from thisinquiry suggest that the initiation of antidepressant pharmacotherapy with an SSRI is warranted.