James E.G. Charlesworth, Grace Petkovic, John M. Kelley, Monika Hunter, Igho Onakpoya, Nia Roberts, Franklin G. Miller, Jeremy Howick

To investigate the clinical efficacy of open-label placebos compared with no treatment in a systematic review and meta-analysis. We searched the Cochrane Injuries Group's Specialised Register, The Cochrane Central Register of Controlled Trials (CENTRAL), Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations (OvidSP), EMBASE (OvidSP), and clinical trials registers and screened reference lists. The search was run on April 27(th) 2015. We included all randomised controlled trials of any medical condition with open-label placebo and no-treatment groups. Authors independently assessed records and extracted data. We excluded non-randomised trials and non-clinical studies. Risk of bias was assessed using Cochrane criteria. We used random-effects model for meta-analysis. We screened 348 publications, assessed 24 articles for eligibility and identified 5 trials (260 participants) that met inclusion criteria. The clinical conditions were: irritable bowel syndrome (IBS), depression, allergic rhinitis, back pain and attention deficit hyperactivity disorder (ADHD). The risk of bias was moderate. We found a positive effect for non-deceptive placebos (standardized mean difference 0.88, 95% CI 0.62 to 1.14, P < 0.00001, I(2) = 1%). Open-label placebos appear to have positive clinical effects compared to no treatment. Caution is warranted when interpreting these results due the limited number of trials identified, lack of assessor blinding, and the fact that positive messages were included alongside open label placebos. Larger definitive trials are now warranted to explore the potential patient benefit of open-label placebos, to investigate the relative contributions of positive suggestions, and ethical implications. This article is protected by copyright. All rights reserved.