Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
I am often (pleasantly) surprised by the spread of evidence-based medicine (EBM) around the world. On recent visits to Oman, Iran, Jakarta, and elsewhere I've found a high level of interest in and practice of EBM. The same issues and questions seem to occur there as in English-speaking countries: is it practical? Does it neglect clinical experience and expertise? Are randomised controlled trials always necessary? Can guidelines replace EBM? What about patients’ unique circumstances and values? Of course, there are additional problems because of limited access to full-text research papers, lack of translation, and greater concerns about the applicability of research conducted elsewhere, but more issues are in common than different.
Last week in Hanko, Norway a group of almost 100 Norwegians, Danes, Swedes, and Finns gathered for the 10th annual 5-day workshop in EBM. I suspect I learnt more than I taught. The problem of rigid guidelines was illustrated with the story of a patient in rural Norway with non-ST elevation myocardial infarction who was eligible for urgent angiography but who did not want to leave their local small hospital. Before trying to persuade the patient of the need for a long journey, the treating physician wanted to know the evidence and the size of benefit—which appeared inconsistent and small, respectively. Of course, after this week he could have “GRADE”d the evidence. Andy Oxman—one of the founders of the GRADE group—spoke on the latest version of the GRADE tool, which is currently featured in a series in the BMJ. Readers should take a look. The main idea of GRADE is to use the traditional design-based hierarchy of evidence as a starting point, but then upgrade or downgrade the level based on quality of the study, risk of publication bias, size of the effect, etc. The methods look a little daunting at first, but are probably simpler than calculating a post-test probability.
The trial in this month’s issue of providing smokers with their “lung age” based on spirometry was the topic of one of the appraisal sessions we did in Hanko. My group was enthused by the number needed to treat of 14, but we also wondered about the downside for the 84% who failed to stop smoking. Did they feel anxious or take on illness roles because of the information? This brought to mind Muir Gray’s saying that “All screening does harm; some also has benefits.” I still think it worthwhile overall, but it will make me more selective in use of the tool. But we’d be interested in readers’ thoughts and experiences with it also.