The review answered a well-defined question that clearly specified the participants that might be included in the incorporated studies, the intervention to be studied and the outcomes of interest. The literature search appears to have been appropriate, but restricting the search to English language publications indexed in MEDLINE could mean that significant numbers of studies may have been missed. This particularly applies to research conducted outside of North America.
In addition, the search strategies devised by the NHS Centre for Reviews and Dissemination (see Other Publications of Interest) mention topic-specific terms, but the authors did not report such terms added to the search strategy specific to this review.
The authors assessed the validity of the included studies and compared those studies which met all the criteria with all others. Unfortunately, it is difficult to reconcile which studies and which patients were incorporated in this comparison. Also, only five studies proved sufficient data to calculate a 2x2 contingency table. As this level of data reporting is required to calculate pertinent diagnostic indices, it might have been useful to have incorporated a criterion into the validity assessment which pertained to the completeness of the reported data.
The study details given were appropriate, but it is important to note that important data were omitted and important analyses were not reported. In particular, a table detailing the sensitivity of the intervention was provided whereas the specificity of the intervention was not reported. Even in those studies where a 2x2 table was reported to have been constructed, the specificity was not included in the review. As the sensitivity details are meaningless in the absence of specificity data, this omission seriously weakens the study. Nevertheless, the conclusions drawn, and the clinical recommendations, appear to follow from what information is presented.