Epidemiology, the cherry on the cake.

Purpose: In guideline development the evidence is more and more coming exclusively from randomized-controlled trials (RCTs), while all other evidential levels are too easily brushed aside. This adopted creed is based on the radical ideas of Archibald Cochrane. Randomize until it hurts—which should presumably be read as a stimulus to perform better research—was the initial suggestion of Cochrane. Methods: This commentary is based on quotes from Cochrane's original work. Results: Cochrane's statements were figured out in a long-gone era in which medical and social inequality prevailed. Adhering to the orthodoxy nowadays hurts both clinicians and patients. I doubt that this was ever Cochrane's intention. Conclusion: In my opinion, the most important part of guideline development should be making inferences of the total medical content (all available evidence including expert opinion); a process that can only be done by subject experts. Methodological assessment, which is undoubtedly the most essential point in the planning of future studies, should come only second place in guideline development and should be used for grading of the evidential level, not for the decision to reject studies completely. Otherwise, far too much relevant evidence is ignored. [ABSTRACT FROM AUTHOR]