 |
 |
Thanks Ash for drawing our attention to this new system for rating/classifying the evidence used to support recommendations. Some quick comments. FORM, like GRADE, is a step in the right direction. FORM and GRADE are useful for systematic reviews with well defined PICOTs (Population, Intervention, Outcome, Time). But they cannot take into account the complexities of evidence and its role in decision-making. For example, let's look at the treatments used to lower the risk of osteoporotic fragility fractures. There is evidence that these drugs are more effective than placebo in terms of fracture risk reduction. But, there are no good head-to-head comparisons, the drugs vary in the amount and quality of evidence for fracture risk reduction at different sites (vertebrae, femoral neck, other, all, ...). Raloxifene has a marketing edge over other drugs with some evidence that it may reduce the risk of breast cancer, but the relevance of this is contested. HRT would not be used solely for fracture risk reduction, but it could be considered for treating both menopausal symptoms and risk of fragility fracture. For each harm or adverse effect, the evidence varies with drug, drug class, dose, and duration of treatment. There is varied evidence for the different drugs about how long treatment should be given for, or how long drug holidays should be. There are economic studies, but their methods vary, reporting varies in transparency, drug costs have changed, and there is no overall set of head-to-head comparisons. There is no evidence from clinical trials of the effects of treatment strategies eg the UK National Osteoporosis Guideline Group (NOGG) and NICE have very different approaches to deciding when to start treatment, but we do not know (although we can guess) what effect this has in practice. People vary in the values they place on different outcomes (chances of benefits vs risks of harms). Thus different recommendations/decisions will be appropriately made - or will be inappropriately not made. Finally (until I think of another complication), it makes little sense to talk about the strength of recommendations - for starters you can say the same thing in too many different ways. It might make sense to talk about the strength of evidential support for recommendations. But, a simple classification or ranking of evidence (even if it has 16,384 classes) will not be useful for vague or complex recommendations such as "the choice of agent is determined by the spectrum of anti-fracture effects across skeletal sites, side effects and cost" or "In women who are intolerant of alendronate or in whom it is contraindicated, other bisphosphonates, denosumab, strontium ranelate or raloxifene may provide appropriate and cost-effective treatment options" (quoted from NOGG guideline). I think that the perfect system is still a long way off. I see systems like FORM and GRADE as steps towards the development of an agreed vocabulary and set of concepts for describing evidence and how it is used to justify recommendations in guidelines AND decisions in clinical practice. And, I wonder if the time has not come to turn attention away from the bureaucratization of evidence appraisal which tries to turn people into automatons, and to develop the human skills including vocabulary, concepts, arguments needed to appraise evidence and justify its application. Michael From: Ash Paul [mailto:[log in to unmask]] Sent: 17 February 2011 14:23 Subject: The Australian 'FORM' Approach to Guideline Development: The Quest for the Perfect System Dear colleagues, Just to let you know that Ben Djulbegovic and Philipp Damm from our Group have recently published a very good article on EB guideline development. The Australian 'FORM' Approach to Guideline Development: The Quest for the Perfect System You can access the article freely at: http://www.biomedcentral.com/content/pdf/1471-2288-11-17.pdf Regards, Ash Dr Ash Paul Medical Director NHS Bedfordshire 21 Kimbolton Road Bedford MK40 2AW Tel no: 01234897224 Email: [log in to unmask]<mailto:[log in to unmask]>