Domhnall MacAuley: NAPCRG and the relevance of evidence based medicine

Domhnall MacauleyTrisha Greenhalgh (London) was always going to challenge the current paradigm. And, she didn’t disappoint. Her keynote at NAPCRG was elegant, persuasive, and beautifully crafted. Drawing from literature and philosophy she explored aspects of care beyond medicine and, in particular, the limitations of evidence based medicine (EBM). Her first assertion was that we get so fixed on a paradigm that new ideas struggle. Such is the tyranny of conformity that any new idea is often rejected by accepted medical journals. She used Helicobacter as an example, pointing to the journey from radical new idea to Nobel prize. One of the major weaknesses she identified, is that EBM is based on reasoning from populations to the individual – but most patients in primary care don’t quite fit with the evidence. Citing Aristotle, she asked for phronesis- practical wisdom… intuition. In her view, EBM has a rationalistic stranglehold and she railed against the reductionism of policy making with its unassailable set of rules and expectations. Her main complaint is that EBM plays to a vision of logic and certainty- and life is not like that. If it were then, using principle component analysis, art could be reduced to a series of coloured shapes and music to a neatly stacked pile of notes.

Trisha got a standing ovation. It was thoroughly justified from the scholarship, breadth of research on the topic, and quality of her presentation. But, I worried that it may have been based on emotion- a collective knee jerk reaction against EBM. I was reassured when a statistician in the audience applauded her talk. Rather than respond to her challenge to EBM, he explained that most statisticians recognise its limitations because they appreciate that they use a small set of variables to look at uncertainties; anyone doing analytical work appreciates the narrowness of Bayesian thinking. Later I heard Joachim Sturmberg (Melbourne) say that medicine has always recognised complexity. We may now understand EBM in terms of mathematical linearity but, he believes, we will come to appreciate that medicine is non linear. Reductionist EBM, he said, is just a phase.

If we had begun to think that EBM was redundant and that we should trust only our intuition, Richard Hobbs (Oxford) underlined its importance through his work on atrial fibrillation. His BAFTA trial examined the relative benefits of anticoagulation in disabling stroke. They recruited 260 practices and a relatively unselected population and tested aspirin (75mg) against warfarin titrated within a therapeutic range. There was a 50% risk reduction at 4 years in disabling stroke- a NNT of 50 patients to avert one event and, it was  just as safe as aspirin. (Mant Lancet 2007). Looking at the BAFTA data, risk scores didn’t perform very well (ROC of .61)  so a more pragmatic policy is to assume that all older patients with atrial fibrillation are at high risk and treat them all. In individual circumstances, it is for the patient to decide but aspirin is not a useful treatment option. Aspirin, he described as a placebo with risk.

Next morning, Martin Fortin (Chicoutimi, Québec) again illustrated the limitations of the linear single disease paradigm in his keynote on multiple morbidity. He pointed out that the classic gold standard research method for appraising an intervention- the RCT- produces good evidence within the parameters it measures but that multiple morbidity needs a more complex design and a more pragmatic approach. Multiple morbidity is a permanent condition with patients adding to it all the time. It increases with age and is associated with a lower quality of life, an increased proportion with psychological distress and, interestingly, an increased risk of medical error. Its prevalence depends on how we define it. In this context, I particularly enjoyed the observation of a geriatrician from the floor who said that normal ageing did not exist and that the length of a patient’s problem list is limited only by the time available in the consultation.

New research ideas and methods are fairly rare at conferences so I enjoyed the distinguished paper on leg cramps – a classic primary care condition but rarely recorded as a diagnosis in a consultation. Scott Garrison (Vancouver), looked at new prescriptions for quinine as a proxy for diagnosis and, using sequence symmetry, identified drugs that could possibly be causally related. He found a relationship with long acting beta agonists and diuretics, with a huge difference depending on the particular diuretic. Statins, perhaps surprisingly, were probably not a factor. It was a fairly preliminary observational study but it was novel and addressed a common but rarely researched area in medicine. One of the audience asked about the quinine level in tonic water (a small dose approximately 1/7 of 200mg tablet) and wondered if a gin and tonic every evening  might be a helpful solution. An interesting idea for a RCT- and, happily, both intervention and control groups should get the gin.

Domhnall MacAuley is primary care editor, BMJ