First, the extent to which a Cochrane review can draw conclusions about the effects of an intervention depends on whether the data and results from the included studies are valid. For our analysis, the most valid studies would have been prospective, double-blind studies with a large sample, which would have had the same gold standard for comparison. However, there were only a few studies which complied with these criteria. All of the 20 studies we included in our analysis complied with at least seven of 10 quality of diagnostic accuracy studies tool items (shown in e-Table 1 of our article1), and the study published in 1995 complied with nine of 10 quality of diagnostic accuracy studies tool items. The possible sources of heterogeneity across the studies were explored using metaregression analysis, which implied that heterogeneity resulted from factors other than the way in which a study was designed.