Mean drug effect is possibly a bad metric to measure. If the distribution is bimodal, or has very high variance, the FDA could take that into account when appr

Mean drug effect is possibly a bad metric to measure. If the distribution is bimodal, or has very high variance, the FDA could take that into account when approving a drug. I believe SSRIs often vary in effectiveness, for example; some people are just very good responders. Fortunately, their mean effectiveness is high enough. (Not a doctor, etc)