This review addressed a clear question supported by appropriate inclusion criteria. Relevant medical databases were searched and search terms were reported. It appeared that no attempts were made to identify unpublished studies and it was unclear whether language limitations were applied. The risk that some relevant studies were not included and of publication and language biases was, therefore, unclear. It was unclear whether steps to reduce reviewer bias and error were undertaken for study selection, data extraction or validity.
The decision to present included studies narratively and not to pool studies in a meta-analysis was appropriate given the heterogeneity between studies. It was unclear whether the significance tests for individual studies were poorly reported or not done at all. The author appeared to calculate intra-group and between-group differences to derive the results. Intra-group differences appeared straightforward, but there were a number of units missing in the table, which made interpretation difficult. Between-group differences were less clear and there was insufficient explanation for how these were calculated. Again, in many cases there were no units were associated with the differences. The author distinguished between statistical and clinical significance, but the finding of clinical significance was said to be when statistical significance occured and the magnitude of difference was 20% or more; without further information on the meaning of magnitude of difference, the results were difficult to interpret.
Limited data available for analysis from the primary studies, small study sizes and poor reporting of key aspects, including statistical significance, raised concerns about the reliability of the authors' conclusions and the appropriateness of developing a rehabilitation protocol.