A meta-analysis of fibromyalgia treatment interventions

CRD COMMENTARY:

This was a reasonable review of the area, which was limited by the inappropriate pooling of the study's results. A thorough literature search was conducted and it is unlikely that many important publications were missed. The inclusion criteria were stated, but the study designs eligible for inclusion were not discussed. The authors do not report how the studies were assessed for inclusion and no formal validity assessment was conducted. Only one in five papers were double-checked for data extraction; the authors reported a mean rate of agreement of 98% which seems very high. Limited details of the studies were tabulated, but it is unclear from this what type of study design was used. The studies were classified as 'between', 'crossover' (controlled design) and 'within' (uncontrolled designs), which is not a very helpful classification. It was not specified whether any of the studies used a randomised design.
To carry out a meta-analysis, the authors grouped the studies into intervention and outcome categories. They then pooled the results within these categories using effect sizes. The heterogeneity of the studies was not reported, so it is unclear whether it was appropriate to pool the results. Even if the effect sizes were not statistically heterogeneous, it would appear that, given the nature of the topic and the wide variation of interventions and outcomes within the reported categories, and the fact that controlled and uncontrolled studies were included, a meta-analysis is inappropriate. A more informative analysis would have been to use the intervention and outcome categories to present a narrative synthesis of the results. The authors' conclusions are not supported by the results presented.