Background.The claims made for the clinical effects of homeopathy are controversial. The results of several meta-analyses of clinical trials are positive, but they fail in general to highlight specific medical conditions that respond well to homeopathy.

Aims.This review examines the cumulative research from randomised and/or doubleblind clinical trials (RCTs) in homeopathy for individual medical conditions reported since 1975, and asks the question: What is the weight of the original evidence from published RCTs that homeopathy has an effect that is statistically significantly different from that in a comparative group?

Method.Analysis of the 93 substantive RCTs that compare homeopathy either with placebo or another treatment.

Results.50 papers report a significant benefit of homeopathy in at least one clinical outcome measure, 41 that fail to discern any inter-group differences, and two that describe an inferior response with homeopathy. Considering the relative number of research articles on the 35 different medical conditions in which such research has been carried out, the weight of evidence currently favours a positive treatment effect in eight: childhood diarrhoea, fibrositis, hayfever, influenza, pain (miscellaneous), sideeffects of radio- or chemotherapy, sprains and upper respiratory tract infection. Based on published research to date, it seems unlikely that homeopathy is efficacious for headache, stroke or warts. Insufficient research prevents conclusions from being drawn about any other medical conditions.

Conclusions.The available research evidence emphasises the need for much more and better-directed research in homeopathy. A fresh agenda of enquiry should consider beyond (but include) the placebo-controlled trial. Each study should adopt research methods and outcome measurements linked to a question addressing the clinical significance of homeopathy’s effects. Homeopathy (2003) 92, 84–91

Keywords: homeopathy, research design, evidence-based medicine, clinical trials