Use of placebo control groups in evaluating efficacy of treatment of unipolar major depression.

Double-blind, random-assignment placebo-controlled trials are routinely used in evaluating efficacy of potential antidepressant agents. In recent years concern has risen that placebo response rates in unipolar depression are too high, and this has eroded confidence in both old and new agents. At the same time, the use of placebos has been questioned by patients and their advocates. We review factors that have been commonly explored as associated with placebo response (e.g., length of episode, severity, subtype), as well as issues in methodology (e.g., interrater reliability, statistical artifacts). We discuss possible methods of dealing with the problem of placebo response, emphasizing reconceptualizing effect sizes to design more powerful single trials and a systematic sequence of trials to achieve the right answers.