Abstract
Evidence-based medicine assimilates available data to direct decisions regarding the care of patients. Extrapolations from epidemiologic observations and laboratory or clinical markers of disease severity have supported new treatments that were subsequently found to be without value in randomized, controlled clinical trials.1 In addition, therapies considered to be beneficial and safe owing to their effect on surrogate markers have been found to be ineffective or harmful in such trials.1,2 Even the results of randomized, controlled trials can be misleading, especially when the trials are not adequately powered.3