Randomized trial evidence and expert guidelines are mixed regarding the value of combined pharmacotherapy and psychotherapy as initial treatment for depression. This study describes long-term results of a randomized trial (N = 393) evaluating telephone-based cognitive-behavioral therapy (CBT) plus care management for primary care patients beginning antidepressant treatment versus usual care. In a repeated measures linear model with adjustment for baseline scores, the phone therapy group showed significantly lower mean Hopkins Symptom Checklist (HSCL) Depression Scale scores (L. Derogatis, K. Rickels, E. Uhlenhuth, & L. Covi, 1974) from 6 months to 18 months versus usual care, F(1, 336) = 11.28, p = 0.001. Average HSCL depression scores over the period from 6 months to 18 months were 0.68 (SD = 0.55) in the telephone therapy group and 0.85 (SD = 0.65) in the usual-care comparison group. Addition of a brief, structured CBT program can significantly improve clinical outcomes for the large number of patients beginning antidepressant treatment in primary care.