In their article titled "A Placebo-Controlled, Randomized Clinical Trial Comparing Sertraline and Imipramine for the Treatment of Dysthymia," Thase et al1 come to the conclusion that pharmacotherapy provides considerable relief from the symptoms of dysthymia, with both sertraline and imipramine being more effective than placebo.
However, they do not mention—and, therefore, I calculated for the readers of your journal—that the mean 17-item Hamilton Depression Rating Scale (HAM-D) end point score is (12.7-5.6=) 7.1 in the sertraline-treated group, (13.4-5.9=) 7.5 in the imipramine-treated group, and (12.7-3.9=) 8.8 in the placebo group. These active drug-placebo differences (1.7 and 1.3, respectively) may be statistically significant, but in my opinion they are clinically not relevant. The same may be said for the differences found in the mean Montgomery-Asberg Depression Rating Scale end scores: 18.5-9.6=-8.9 in the sertraline-treated group, 18.6-9.5=9.1 in the imipramine-treated group, and 19.0-6.7=12.3 in the placebo group. In short, Thase et