The greatest achievement of modern medicine is arguably not any single therapeutic advance, but the development of a highly sophisticated framework for distinguishing a true advance from what merely looks like one. Evidence based medicine has completely transformed our profession to the extent to which no doctor—not even the most cavalier one—would countenance a change to current practice that has not been justified by a rigorous comparison between the old and the new. What constitutes a rigorous comparison is well established; indeed, agreement on the principles of designing and reporting therapeutic trials is so widespread that all good journals refuse to publish any study that does not fit the standard template. These principles are simple and incontestable: all comparisons between treatments should be fairly made, with meticulous attention to potential sources of bias, with clear outcome measures, and with inferences limited to only what is statistically justified by the data.

One might therefore have thought that the response of any competent physician to what I am about to describe would be predictably derisive. Imagine that the government proposed a radically new treatment (let's call it Effupin) for a complex and important condition that has hitherto been treated in an imperfect but largely satisfactory way. Effupin's mode of action is …