Practice guidelines by specialist societies are surprisingly deficient

Abstract

The term evidence-based implies that it is supported by data from randomised controlled trials (RCTs). RCTs, if conducted and reported correctly, are accepted as the best way to confidently know the benefits and harms of an intervention. When purportedly evidence-based guidelines and recommendations stray from this principle, they should no longer be able to call them evidence-based. So much of the evidence-based paradigm is missing from the Joint British Societies Guidelines on Prevention of Cardiovascular Disease in Clinical Practice (JBS-2) that the author critiquing them has entitled his article ‘Eminence Based Guideline’ (1). The use of the word ‘eminence’ seems particularly appropriate, as it connotes the ‘arrogance’ that seems to be an invariable part of guidelines based on expert opinion (2). The JBS-2 guidelines score low on most of the main quality criteria according to the AGREE instrument (3): stakeholders’ involvement, rigour of development, applicability and editorial independence. This is not the first time that specialist societies’ clinical practice guidelines have been found deficient; however, it surprising that guideline rigour and quality is not improving. Clinical practice guidelines came about when specialists were asked to give their opinions and guidance as to best practice in an attempt to decrease unacceptable variability and cost inefficiencies in clinical practice. Not surprisingly, this proved inadequate and often irrelevant to the problems encountered in primary care. The evidence-based medicine movement arose to provide some evidential basis behind guideline recommendations. The result has been some improvement in the validity and reliability of most guidelines, but unfortunately this has not had much impact on clinical practice (4). Why is that the case? In my opinion one of the reasons is because guideline writers are unable to overcome their own ‘arrogance’ that they can provide ‘aggressively assertive’ guidance despite lack of evidence (2). The attempt by guideline writers to make the rationale behind the decision process more transparent by grading the recommendations has not worked. Readers of guidelines do not distinguish between grade A, level 1, grade B, level 2 recommendations and grade C, level 4 recommendations. In my experience, when doctors find any recommendations in guidelines inconsistent with their own clinical practice, they become sceptical about all of them. Furthermore as long as guidelines continue to be funded by companies and written by individuals with competing interests, they lack credibility (5). This sad situation is certainly not helped when guidelines, such as JBS-2, are produced that after evaluation are judged as in Minhas’ article: ‘of low quality and should not be recommended for clinical practice’ (1). How can guidelines be improved? I suggest that the best way forward is to return to the first principle. Guidelines should limit their recommendations to the interventions and patient populations that are supported by high quality evidence from RCTs. This approach has many advantages: (i) recommendations would be supported by RCTs and systematic reviews; (ii) there would be no necessity to grade recommendations; (iii) recommendations would be fewer and less subject to bias; (iv) recommendations would be more likely to be followed, because they are fewer and better supported; (v) time and effort to produce guidelines would be less and (vi) clinical settings with no recommendations would be identified as requiring RCTs. Following this first principle, cardiovascular guidelines would not make recommendations regarding lipid targets or blood pressure targets as was done by JBS-2; RCTs are badly needed to determine optimal treatment targets for both lipids and blood pressure. I am not suggesting that it is always easy to determine when RCT evidence is clear enough to make a strong recommendation. There will still be recommendations that are open to debate, and this would be healthy. I am also aware that the methods of evaluating guidelines, including the AGREE instrument (3) used by Minhas could be improved. I am suggesting that physicians and patients deserve truly evidence-based guidelines and not ‘eminence’-based guidelines, and I am hopeful that by putting forward this provocative proposal that I can help to discourage guideline recommendations based on expert (eminent) opinion.