Performing systematic reviews of clinical trials of acupuncture: problems and solutions

Abstract

Systematic reviews of randomized controlled trials, rather than narrative reviews, generally provide the approach that is least subject to bias for assessing the efficacy and effectiveness of a therapy. But despite their increasing use, systematic reviews are not free of problems, particularly when used to assess the evidence for acupuncture. Common weaknesses are inadequate literature searches and unclear inclusion criteria, including absence of a definition of ‘acupuncture’ for delineating the scope of the review. In addition, the adequacy of the acupuncture performed in the trials is often not addressed. Overall, there is little consistency in the criteria used for critiquing the included trials. The validated five-point, and expanded nine-point, Jadad scales for assessing trial quality are discussed, especially in regard to their emphasis on double blinding that can be applied to acupuncture trials if ‘double blind’ is defined as patient and assessor blinding. Suggestions are made for avoiding each of the above-cited problems for future systematic reviews, which should include an acupuncture specialist in the author team. For the near future, however, there is a greater need for new high-quality RCTs of acupuncture than for additional systematic reviews of existing trials.