Placebo response rates in trials of antidepressant drugs in adults with clinical depression: Increasing, decreasing, constant or all of the above?

Abstract

Objective: A seminal study quantified an increase in placebo (and drug) response rates in antidepressant trials during the eighties, with subsequent key systematic reviews reporting placebo response rates as either having continued to increase or stabilize in trial subjects with major depression. We therefore undertook an additional study examining a more recent period.

Method: We analyze response rate data from 121 studies published over the 2001-2015 period and sourced from a previous meta-analysis of antidepressant drugs.

Results: Our analyses indicated trends for decreasing placebo response rates over the whole study period. Analyses of consolidated year blocks quantified a significant linear decrease and a significant cubic pattern in placebo response rates. Visual inspection of the yearly data patterns revealed wide fluctuations as well as distinct and correlated peaks and troughs for both placebo and drug responses.

Limitations: The key studies we analyzed differed in a number of ways, including selection of inclusion criteria (especially in relation to analyzing published studies alone or together with unpublished studies) and analytic techniques.

Conclusions: The markedly fluctuating placebo response patterns evidence an unstable 'signal' so leading principal studies to generate three mutually contradictory conclusions. We suggest that the increase in rates in the 1980s studies reflected the uptake and impact of 'major depression' being used as the diagnostic criterion, while the striking peaks and troughs observable over our study period argue for varying trial characteristics over time or rater biases in judging depression.

Keywords: Antidepressive drugs; Depression; Randomized control trial.