Sample size calculation

The sample size (100 in the active treatment groups and 40 in the placebo group) was estimated according to the following variables: primary endpoint change from baseline on Hamilton depression total score; assumption of improvement in the placebo group by nine points, improvement in the active drug groups by 12 points, and standard deviation of changes by 6 points; a minimum important difference for comparison between hypericum extract v placebo: d>0.5 (d is difference between both treatments divided by standard deviations; non-inferior difference for comparison between hypericum extract v imipramine: d£ 0.35; and a =0.025 (one sided), 1-b =0.80.

Flow of participants