Jeffrey A. Welge

Rationale. Variation in placebo response within and among clinical trials involving patients with schizophrenia can substantially affect conclusions about the efficacy of new antipsychotic medications. Therefore, it is of great importance to identify factors that moderate response to placebo in such trials. Objective. The objective of this meta-regression analysis was to estimate the effect of potential moderators of placebo response in randomized, short-term clinical trials involving patients with schizophrenia, schizoaffective disorder or schizophreniform disorder. Methods. Mean placebo response and potential moderators were extracted from 35 placebo-controlled, randomized trials of antipsychotic medications in patients with schizophrenia. Placebo response was defined as the absolute change in the Brief Psychiatric Rating Scale total score. Fixed-effects meta-regression was used to investigate between-trial variation in placebo response. Results. Trial duration accounted for a substantial proportion of the between-trial variation in response (27%), with greater improvement on placebo observed in shorter trials. Other variables showed insufficient variation across trials to permit any inferences regarding their relationships with placebo response. Conclusions. Placebo-controlled trials of short duration (<6–8 weeks) are vulnerable to substantial placebo response. Recruiting patients with more severe pathology to mitigate placebo response does not appear to offer benefits and may even be counterproductive. Meta-analyses based on individual patient data offer the potential for much more detailed and inferentially sound exploration of factors affecting placebo response and are highly recommended.