Is High Placebo Response Really a Problem in Depression Trials? A Critical Re-analysis of Depression Studies.

Objective: We investigated the accuracy of the often-stated assumption that placebo nonadditivity and an increasing placebo response are major problems in clinical trials and the cause of a trend for smaller treatment effects observed in clinical trials for major depressive disorder (MDD) in recent years. Method of research: We reviewed data from 122 MDD trials conducted between the years 1983 and 2010 (analyzed originally by Undurraga and Baldessarini in 2012) to determine whether the data support the assumption of placebo additivity. Statistical techniques, such as conventional least squares regression, orthogonal least squares regression and locally weighted loess smoothing, were applied to the data set.
Results: Re-analysis of the data set showed the active and placebo responses to be highly correlated, to the degree that would be expected assuming placebo additivity, when random variability in both active and placebo response is considered. Despite the placebo responses in MDD trials increasing up to approximately the year 1998, we found no evidence that it has continued to increase since this date, or that it has been the cause of smaller reported treatment effects in recent years.
Conclusion: Attempts to reduce the placebo response are unlikely to increase the treatment effect since they are likely to reduce drug nonspecific effects in the treatment arm by a similar amount. Thus, it should come as no surprise that trial designs set up with the sole purpose of reducing placebo response fail to discernibly benefit our ability to identify new effective treatments.