Placebos, active control groups, and the unpredictability paradox

Randomized controlled trials often rely on placebo control groups to estimate treatment differences. Recently, the high frequency of negative trials and ethical concerns surrounding the use of placebos have brought the use of placebo control groups under increased scrutiny. Although many psychiatric researchers argue that placebo control groups should be replaced with active control groups, we argue that preferential use of active control groups will not reduce the number of negative trials. Rather, we suggest that some of the variation and contradiction in randomized controlled trial results arises from the clinical heterogeneity of patient characteristics, disease severity, comorbidity, and cotherapies. Further characterization of patient heterogeneity, through improved disease taxonomies, severity indices, and classification of comorbid diseases, will serve to reduce clinical heterogeneity among patients and reduce the number of negative trials produced by wide variation in treatment and control response rates.