Objective. To compare the direct mental health care costs between individuals with Seasonal Affective Disorder randomized to either fluoxetine or light therapy. Methods. Data from the CANSAD study was used. CANSAD was an 8-week multicentre double-blind study that randomized participants to receive either light therapy plus placebo capsules or placebo light therapy plus fluoxetine. Participants were aged 18-65 who met criteria for major depressive episodes with a seasonal (winter) pattern. Mental health care service use was collected for each subject for 4 weeks prior to the start of treatment and for 4 weeks prior to the end of treatment. All direct mental health care services costs were analysed, including inpatient and outpatient services, investigations, and medications. Results. The difference in mental health costs was significantly higher after treatment for the light therapy group compared to the medication group-a difference of $111.25 (z = -3.77, P = 0.000). However, when the amortized cost of the light box was taken into the account, the groups were switched with the fluoxetine group incurring greater direct care costs-a difference of $75.41 (z = -2.635, P = 0.008). Conclusion. The results suggest that individuals treated with medication had significantly less mental health care cost after-treatment compared to those treated with light therapy.