The review question was clearly stated and indicated that CBT was to be the focus of the review. Inclusion criteria were specified for study design, intervention, control and participants, and apparently indicated that cognitive or behavioural interventions were to be compared to a non credible control. Some experimental interventions did not include cognitive treatment and some controls were credible treatments. Thus, it appeared that inclusion criteria were not adhered to and inclusion criteria did not appear to reflect the stated review aim. Two relevant databases were searched and authors were contacted, but it was not clear if attempts were made to minimise language bias. Methods were used to minimise reviewer errors and bias in the assessment of validity and extraction of data, but it was not clear whether similar steps were taken in study selection. The authors stated that validity was assessed, but results were not reported and so it was not possible to comment on the reliability of the results presented.
Other than diagnostic criteria, no information was provided about participants. In addition, very few details were provided about interventions and no information was given about measures used to assess outcomes, so it was not possible to determine the clinical heterogeneity among studies and hence the appropriateness of pooling data. Reducing apparently outlier effect sizes in one study to the closest relevant non-outlier values in that study appeared questionable.
Data were pooled using meta-analysis and significant statistical heterogeneity was found. Various potential sources of heterogeneity were examined and some potential reasons for differences between studies were discussed. Inadequate information about participants and interventions, lack of reporting of study quality and differences between studies meant that it was difficult to assess the reliability of the authors’ conclusions.