Negative Effects During Placebo Treatment

Importance Analyzing effects within placebo groups allows for transdiagnostic comparisons, as placebo is the only substance systematically studied across all major psychiatric diagnoses. Recently, the study team meta-analytically showed that improvement under placebo varies across major psychiatric diagnoses. However, comprehensive transdiagnostic comparisons of negative effects under placebo (nocebo effects) are lacking. Objective To compare premature study termination rates in placebo groups of high-quality randomized clinical trials (RCTs) across 9 major psychiatric disorders, focusing on total dropouts, dropouts due to adverse events, and dropouts due to lack of effect. Data Sources This analysis is part of a broader research project using a systematic approach to identifying the most recent high-quality systematic review for each diagnosis (Open-Science-Foundation preregistered: u469a ). Study Selection From these reviews, the 10 highest quality and most recent placebo-controlled RCTs were selected for each diagnosis, totaling 90 RCTs. Data Extraction and Synthesis Searches and data extraction were conducted according to the Cochrane Handbook. Pooled dropout rates (DRs) with 95% CIs were calculated. Main Outcomes and Measures The primary outcome was total DR per diagnosis, determined in random-effects meta-analyses. Diagnostic differences were tested for statistical significance using Q tests. Potential confounders were examined in multivariable meta-regression analyses. Results Eighty-six of the 90 studies reported total DRs (10 056 participants). DR differed between diagnoses ( Q = 82.2; df = 8; P < .001), with schizophrenia (DR, 0.41; 95% CI, 0.35-0.48), panic disorder, and mania showing the highest rates, and posttraumatic stress disorder, major depressive disorder (MDD), and attention-deficit/hyperactivity disorder (ADHD) (DR, 0.17; 95% CI, 0.11-0.25) had the lowest. Schizophrenia and mania also had the highest DR due to lack of effect, while ADHD and MDD had the lowest ( Q = 71.3; df = 8; P < .001). Most dropouts due to adverse events occurred in obsessive-compulsive disorder (DR, 0.07; 95% CI, 0.05-0.09) and panic disorder studies and the fewest occurred in MDD and ADHD trials ( Q = 32.1; df = 8; P < .001). Meta-regression revealed no additional associated factors on DRs. Conclusion and Relevance These findings indicate that placebo treatment is associated with adverse effects that differ among psychiatric diagnoses. The main negative effect was lack of effect, with diagnostic variations corroborating findings on positive placebo response from earlier analyses. Schizophrenia had the least favorable course under placebo, while MDD demonstrated the most favorable.