In a study in the August 2000 American Journal of Psychiatry (Vol. 157, No. 8), Walsh and colleagues including Wilson randomized patients who didn't improve with either cognitive behavioral therapy or interpersonal psychotherapy, to either Prozac or a placebo.