There can be few better illustrations of why the term ‘evidence-based medicine’ is misleading than the title of Kenneth W Goodman's (no relation) thought-provoking book. The question has been asked many times, but what other sort of medicine is there? Medicine that is intentionally divorced from evidence is unworthy even of ethical consideration. It is the subtitle of the book that is the author's real subject—fallibility and responsibility in clinical science—though even that is wrongly phrased. Had the title been ‘The Ethics of Uncertainty in Clinical Medicine’, I would have started reading with less prejudice.
Evidence-based medicine (EBM) should have been called epidemiology-based medicine. Its basic flaw is an over-reliance on meta-analysis, which, as many but especially Bruce Charlton (whose papers, like mine on this subject, go uncited in Goodman's work) have pointed out, is an insecure basis for treating individuals. It also proposes ‘levels of evidence’ upon medicine without due allowance for the fact that different questions in different circumstances demand different sorts of evidence. Had its advocates used the term epidemiology-based medicine, this approach might have made less impact but would have generated less antipathy. As Kenneth Goodman realizes (p. 58), the arguments are not whether we should use evidence but what sort of evidence we should use.
Goodman's message is that the evidence-based medicine movement has the welfare of patients at heart, that it is our duty to pay attention to the movement (though not slavishly), and that we must also strive to reduce uncertainty in medical practice. In the 140 pages of text he also canters through the ethics of using secondary evidence in databases, the difficulties of public health policy, and some reflections on informed consent. Some of his ideas are challenging, and many will have readers nodding agreement; but it is hard not to think that he is chasing a platitude. And ultimately that platitude comes up against a truism, which he recognizes (p. 139)—there is no general statement or rule that governs the reliability or uncertainty of medical evidence, and no algorithm to tell us what we should do when we don't know what to do. Medicine is never going to be easy; moreover, it may never get any easier.
What really disappoints me is his treatment of metaevidence (p. 44 et seq.). He picks it up between his teeth but then sets it down gently. He describes the process of systematic review as ‘scientific’, but it is not (although perhaps he is not rigorous in his definition of science since management and marketing is earlier referred to as one of the ‘sciences’ that first embraced meta-analysis). I would have liked to see the views of Bruce Charlton and Jonathan Rees considered here. Why doesn't he keep hold of the notion that there is ‘a lot wrong with the gold-standard randomised controlled trial’ and run with that idea? He quotes John Bailar (p. 59) granting meta-analysis ‘a potentially useful role in carefully selected situations’, but doesn't quote Bailar (as I have) saying that the problems of meta-analysis are so deep that he worries should their results be too readily applied clinically. Nor does he cite Bailar's preference for the old-fashioned review by the knowledgeable expert explaining and defending judgments. Goodman consistently discards narrative reviews as biased: ‘the inadequacies of narrative reviews left us no choice but to develop these epistemic engines’ (p. 29); ‘narratives are biased... because the nature of the review was such that reviewers could not find a... neutral vantage point’ (p. 44).
Goodman is good on the debate about science versus art in medicine. He cleverly argues that the art is really tacit experiential knowledge, which is what underlies ‘clinical judgment’. He is also realistic about the difficulties of incorporating patient preferences into treatment plans (p. 135): ‘Sometimes they just do not make any sense [and become] a silly waste of public money’. And, a couple of pages further on, while stressing the ethical imperative that lies with individual clinicians, ‘We should be forgiven the weariness that accompanies the need to sort out individual duties in an environment in which society has dropped the ball’.
The book discusses important issues, but under the wrong title and with a hollow centre. It is still worth reading, but a second, retitled, expanded edition ought to be better. I can thank Goodman at least for not describing evidence-based medicine as a ‘new paradigm’.