The grammar of interpretive medicine.

Can the practice and science of medicine ever be reconciled? Several years ago, The Lancet introduced a series of occasional duets entitled, ambitiously, “science and practice.” The idea was to commission 2 writers to review a subject — for example, Parkinson’s disease, heart failure, chronic renal failure — in these constituent parts. Once the underlying biology and epidemiology were set out, we thought, the practical aspects of management would naturally follow. The series proved a failure and fizzled out through neglect. There were technical difficulties, to be sure. Timing 2 review articles to run in parallel, an exercise multiplied many times over for an elaborate series, required a logistical rigour beyond the juggling skills of the commissioning editor (me). But often the 2 papers were so at odds with one another in style, approach, depth and interpretation that it must have been hard for the reader to discern any connection between them at all. Olli Miettinen, in his careful analysis of how physicians use evidence in their clinical practice, tackles this issue head on. He invites us to separate evidence into 2 categories. First, there is specific evidence that comes directly from the patient: symptoms, signs, results of investigations, age, risk profile and so on. Second, there is general evidence that reflects the physician’s past experience with similar patients in similar settings and an acquaintance with the available scientific literature on the subject. The coherent and convincing medicine that Miettenin defends is one in which these types of evidence are applied systematically to diagnosis and treatment. But this project, by Miettinen’s own admission, is also destined to fail. Of the grand utopian approach to medical evidence that he demands of clinicians, he asks (and answers): “Does he or she possess the competence for it? Does he or she have the time for this? Broadly speaking, no and no.” Miettinen tries a different, and currently rather fashionable, tack. Is meta-analysis the answer to the impossible diversity and complexity of evidence? Again, it seems not: “In fact, trusting a ‘meta-analyst’ is more difficult for a practitioner than trusting a practitioner is for a patient.” He goes on to take a swipe at those conducting large-scale programs of systematic reviews (within the Cochrane Collaboration, for example) and at clinical professors whose experience remains “very narrow” and “off-focus.” The prospects for relief seem bleak: “Enormous efforts and expenditures have, thus, resulted in evidence which guides presumed experts to highly divergent conclusions and leaves nonexpert readers to judge for themselves!” I share some of Miettinen’s pessimism — his diagnosis of the problems facing physicians is surely correct — but not all of it. We can do more to educate physicians in the analytical and humanistic skills needed for a reflective clinical practice, although I readily admit that to do so will be a struggle. At the same time, I believe that Miettinen, and others who have tried to build fanciful theoretical structures for managing evidence, are making the problem far more difficult than it really is. Indeed, the challenges Miettinen identifies (and perhaps too easily gives up on) are yielding to persistent scrutiny. Clinical strategies and methodologies that link evidence to practice are making substantial progress. The skill that physicians lack above all is the ability to reason successfully. By “to reason” I mean interrogating a clinical argument to discover its weaknesses or the basis for its validity. Reasoning is not the skill of switching on a computer, typing in a few key words and printing out several abstracts of randomized trials Editorial