Inquiry into racial and gender differences have rarely focused on prescribing of new drug therapies. OBJECTIVE: We examined the effect of race and gender on the likelihood of being prescribed a new antidepressant (fluoxetine versus existing antidepressants) in the South Carolina Medicaid population. METHODS: Patients were included in the study if they had a primary physician or hospital diagnosis of depression (based on ICD-9-CMs) between January 1, 1988 and December 31, 1990 and if they had filled an antidepressant prescription within four months following the diagnosis. Selection of the study period was based on the year fluoxetine entered the market (1987). A total of 1,447 patients met the inclusion and exclusion criteria. Racial and gender differences in prescribing patterns were determined using logistic regression, controlling for age, type of county, provider type, Medicaid eligibility, physician visits and early filling of prescriptions. RESULTS: Blacks were 30% less likely than whites (OR = 0.699, CI = 0.540,0.904. p <0.01) and males were 48% less likely than females to receive a fluoxetine prescription. CONCLUSION: Results demonstrated that there were significant racial and gender differences in the utilization of fluoxetine in the period immediately following its launch. The findings should be interpreted with caution as potential confounding by disease severity, patient behavior, and certain regional characteristics were not controlled.