Even where reviews led to conclusions,

these were typically couched in terms such as ‘moderate effect’, ‘few high quality trials’ and ‘there is a need for further, well-designed trials.’ The equivocation shown by so many authors is, of course, understandable. That further information and evidence is desirable is a truism and a non-committal conclusion has become almost obligatory in systematic reviews. Ku-0059436 supplier Is it, however, always appropriate to conduct a systematic review? A systematic review is a timeconsuming matter, not uncommonly taking from six to 12 months to complete. Where it becomes clear that minimal evidence exists (as opposed to a substantial number of wellconducted trials leading to an unclear result) one wonders whether the reviewer’s energy might have been better spent in other ways. Perhaps inconclusive systematic reviews of randomised trials, where the reader is left with no idea whether a treatment works, should include an expanded ‘Discussion’ section with a broader gathering of information from the literature and from clinical reasoning and other study designs to at least provide a synopsis of the evidence as it exists. What then of the other high level

source of evidence, the randomised controlled clinical trial? Here too, publication rates in the major physiotherapy Selleckchem 3-MA journals have increased over the years, with this journal leading the way. It is certainly extremely encouraging to see such growth in this type of research, yet there are traps for the reader and the researcher here, too. out One danger is that the reader travels no further than the authors’ conclusions with, perhaps, a nod in the direction of the methodological rating through the PEDro score. Often this is the message the reader takes away. However, in

one investigation of outcome studies, 70% were found to have conclusions related to causation that were unjustified by the research design used (Rubin and Parrish 2007). Even in randomised trials, the authors’ conclusions may not always be valid. The PEDro score provides a service of enormous value, but is constrained to assess to what extent the design of the trial threatens the internal validity of the study, not the overall validity of the question or choice of design and, as the originators of the instrument themselves note, they can only rate what the authors are prepared to disclose (Moseley et al 2008). In many randomised trials the primary hypothesis is the only hypothesis tested or reported. There are few examples in which subsequent analysis has been published or where further exploration of the data seems to have occurred. The researchers often seem to consider that, when a randomised trial is published, they can draw a line under that and move on to the next study.