Cochrane systematic reviews in acupuncture: methodological diversity in database searching.

BACKGROUND Since the early 1970s, the efficacy of acupuncture for treating clinical conditions has been evaluated in several hundred randomized trials. Results from these trials have been synthesized in systematic reviews. A well-designed systematic review provides the highest level of evidence for establishing the efficacy of a clinical intervention. OBJECTIVES The present study assesses the source of original literature contributing to Cochrane reviews on acupuncture. Databases searched to retrieve original studies are evaluated. The distribution of controlled trials in acupuncture across different topic areas and journals, the ability of the reviews to provide conclusive results, and the proportion of original studies indexed with MEDLINE are evaluated. METHODS Systematic reviews on acupuncture were extracted from the Cochrane Database of Systematic Reviews. The key search term used was "acupuncture." When more than one systematic review was retrieved on the same topic, the most recent review was included. Indexing of individual clinical trials with MEDLINE was searched using the Single Citation Matcher in PubMed. RESULTS A total of 94 papers were retrieved from the Cochrane database, of which 10 were included in the analysis. The most common subject areas were related to chronic pain. Considerable heterogeneity was observed in the number of databases searched (median 5, range 3-12). A total of 69% (74/108) papers were indexed with PubMed. Only 13% (14/108) of the papers were published in the primary acupuncture journals. Conclusive statements about the efficacy of acupuncture were made in only 2 of the 10 systematic reviews. CONCLUSIONS Considerable methodological diversity exists in the comprehensiveness of database searches for Cochrane systematic reviews on acupuncture. This diversity makes the reviews prone to bias and adds another layer of complexity in interpreting the acupuncture literature.