Scoring the quality of clinical trials.

To the Editor: Dr Jüni and colleagues compared 25 checklists from systematic reviews. We agree that readers should be critical of the heterogeneity of the content and results of checklists. Therefore, empirical studies in this field are useful. However, by using the same collection of checklists as Moher et al, Jüni et al portray an unfair representation of the scientific development of research groups. Our list, which Jüni et al included in their analysis, was developed in 1990 and published in 1991. Thereafter, however, we have changed and hopefully improved our checklist, according to the new insights provided by Moher et al and others. This has resulted in an updated version of our checklist, which has been published in the method guidelines for systematic reviews within the Cochrane Back Review Group. The updated checklist has already been used in several protocols and reviews in the module of the Back Review Group, as well as in related journal articles. We agree that, based on current knowledge, the use of checklists is problematic. However, we argue that replications of studies like that of Jüni et al should evaluate the latest versions of checklists used by the research groups at issue, not lists from the relatively early years of systematic review methodology. Otherwise the results of such empirical studies will be of predictably limited value.