Is there a placebo problem in antidepressant trials?

In psychiatry, particularly in antidepressant clinical studies, placebo-controlled trials often yield results that are very difficult to interpret because of robust placebo responses. Meta-analyses of trials in major depressive disorder (MDD) suggest that drug-placebo differences in response rates range from 11% to 18%. However, in trials of marketed antidepressants present in the FDA databases, antidepressant drugs were superior to placebo in only 45 out of 93 RCTs (48%), and the placebo response overall appears to have increased over time. This gradual increase in placebo response rates may lead to delays in bringing new antidepressant treatments to the market, increased costs of antidepressant drug development and, in some cases, decisions to stop the development of certain compounds, or FDA decisions to not approve new treatments. A number of possible contributing factors to this significant placebo response in MDD have been identified, but further studies are needed. Many of the remedies used by researchers to minimize the placebo response, such as lead-in periods or shortening the duration of study visits, have failed to show consistent benefits. From our analysis of published studies, it appears that expectations about the speed of response may be shaped by the duration of the trial and that most of the placebo response occurs in the first half of the trial, regardless of its duration. These observations have led us to develop a novel approach to the placebo response problem called the Sequential Parallel Comparison Design.