Clinical Question: Are the effects of homeopathy merely placebo effect?

Setting: Various (meta-analysis)

Study Design: Systematic review

Synopsis: The authors searched 19 databases, including some that specialize in complementary and homeopathic registries, for clinical trials of homeopathy and conventional medicine. They enhanced the search by consulting the reference lists of included articles and by contacting experts in homeopathy. Additionally, they searched the Cochrane Controlled Trials Register to identify trials of conventional therapies. They only included studies that had a parallel group design with placebo control, used a random or quasirandom assignment to treatment and placebo groups, and had a written report with sufficient data to allow for calculating the effects of the intervention. The authors did not report on measures that allow for determining the quality and objectivity of the application of their inclusion and exclusion criteria. Two investigators extracted the data independently using a tested data collection method. They resolved any discrepancies by consensus. The investigators also assessed the methodologic quality of each study but do not describe any steps to ensure the reliability and accuracy of these assessments.

They included 110 trials of homeopathy and 110 matching trials of conventional therapies. The trials covered a spectrum of conditions including respiratory tract infections, pollinosis and asthma, obstetrics and gynecology, surgery and anesthesia, gastroenterology, musculoskeletal disorders, and neurology. The homeopathic trials were less likely to be published in English (53 versus 85 percent) or in MEDLINE-referenced journals (41 versus 86 percent) and were more likely to be described as double-blinded (92 versus 87 percent) and to report concealed allocation (45 versus 19 percent). Only 19 percent of the homeopathy papers and 8 percent of the conventional papers were of high quality (double-blinded, randomized, with concealed allocation). In homeopathy and conventional trials, smaller and lower-quality studies were more likely to show benefit with the intervention. Looking at only the higher-quality studies, the trials of homeopathy were no more effective than placebo (odds ratio [OR] = 0.88; 95% confidence interval [CI], 0.65 to 1.19); the matched conventional therapy trials were more effective (OR = 0.58; 95% CI, 0.39 to 0.85). Ideally, the authors would have evaluated studies directly comparing conventional and homeopathic therapies.

Bottom Line: High-quality studies demonstrate that homeopathic therapies are no more effective than placebo. (Level of Evidence: 1a–)