This single author review addressed a clear research question that was defined in terms of the participants, study design and intervention; inclusion criteria for the outcomes were not explicitly reported but the primary review outcome was clearly stated. Three relevant databases were searched, no real attempts were made to locate unpublished studies, and no language restrictions were applied. The methods used to select studies, assess validity and extract the data were not described, so it is not known whether any efforts were made to reduce reviewer errors and bias.
Given that this was a single-author review, it is possible that only one reviewer was involved in all stages of the review, which could introduce selection and reporting bias. There was also little attempt to locate unpublished work, which could introduce publication bias. This is a real possibility given the nature of the studies and the fact that it is likely that many studies in this area are not published in journals. However, by limiting the review to only randomised controlled studies and assessing the quality of the studies using the Jadad scale, the author has tried to limit his review to the best-quality evidence available. Validity was assessed using specified criteria but only the composite score was reported, making it difficult for the reader to judge the potential sources of bias.
No details of the participants were given, so the population results may not be generalisable. The data were pooled by meta-analysis. However, although statistical heterogeneity was assessed it was not reported, thus it was not possible to judge the appropriateness of the study pooling. Overall, the author's cautious conclusions appear valid given the limitations and the paucity of evidence available.