Masked by the mean: how aggregating placebo responders and non-responders distorts clinical trial outcomes

The placebo effect, often dismissed as statistical noise, is a real mind-body phenomenon caused by conditioning and expectation. We should distinguish between the placebo effect (the psychobiological process) and the placebo response (the outcome). This commentary argues that placebo effects are often underestimated in randomized controlled trials (RCTs) because trial outcomes are reported as averages of both responders and non-responders. Averaging can conceal meaningful effects, distort comparisons with active treatments, and generate inaccurate conclusions about treatment efficacy. To illustrate this, I use data from my research to show how averaging across participants diminishes the statistics-based placebo effect. Therefore, ignoring response heterogeneity can reduce accuracy and mask actual effects in responders. This issue extends beyond methodology. Underestimating placebo responses can lead to Type I errors in drug approvals, while in clinical practice, it can prevent practitioners from leveraging placebo mechanisms as part of holistic care. Ethical concerns also arise when the total placebo response approaches that of a drug, raising questions about exposing patients to unnecessary medication risks. Future trials should capture individual differences, such as expectancy assessments, prior-experience measures, stratified randomization, latent-class models, and biomarker research. Although these approaches pose challenges, they could incorporate heterogeneity into analysis and practice. In conclusion, embracing variability in placebo and treatment responses is more than a technical change; it is a shift toward greater validity, reliability, and personalized medicine. Moving beyond group averages and recognizing individual differences can help research produce more reliable evidence to guide treatment decisions.