What Evidence in Evidence‐Based Medicine.doc

1 What Evidence in Evidence-Based Medicine? JOHN WORRALL 1?1 Introduction The usual reaction from outside observers on being told that there isa (relatively) new movement called "Evidence-Based Medicine" is "What on earth was medicine based on before?" Telling clinicians that they ought to operate in accordance with EBM sounds, then, about as controversial as telling people that they ought to operate in accordance with virtue. But the more obviously true the general idea of EBM is normatively speaking, the more important that idea is if, descriptively speaking, it has not, or has not uniformly, been put into clinical practice. Moreover, just as everyone agrees that people should act in accordance with virtue, but disagrees about what virtue precisely is, so in the case of EBM disagreements soon emerge once we get to the details. We all presumably accept that clinicians ought to base practice on best evidence, but what counts as best evidence? How persuasive are different kinds of evidence (or rather how persuasive ought they to be)? What evidential role, if any, is played by 'clinical experience' or 'clinical expertise'? EBM needs a fully coherent, articulated and detailed account of the correct relationship between the evidence and various therapeutic and causal claims that would answer questions such as these from general first principles. This seems to me an area where philosophers of science can, for once, beof real practical value. After all, the topic of the relationship between theory and evidence in general, and that of the relationship between causal claims and evidence in particular, have, of course, long been recognised as central issues in philosophy of science. There are, I believe, two main areas in which EBM has yet to produce a fully defensible account of the view of evidence that it mends. The first concerns the role and evidential power of randomization; the second the role and evidential power of clinical judgment and expertise. In the present pap