3. Practice guidelines and other support for clinical innovation

The ®rst two articles discussed how to formulate and pursue clinical questions and how to build a clinical reference library. Looking up clinical questions during or immediately after patient care can be an effective method of learning: however, you must ®rst realize your ignorance, then formulate a searchable question, ®nd the answer and change your practice as a result. Even when there is widespread knowledge about an important innovation in clinical practice, such as aspirin for patients with myocardial infarction (MI), many patients are still incorrectly managed. Practice guidelinesÐ`systematically developed advisory statements created according to validated methodolo-gies' 1 Ðact as a compact summary of the evidence and other factors guiding patient management so could be helpful in this context, but they are many and various 2,3. Your consultant asks you to identify a guideline to increase the local use of aspirin after MI As a perfectionist, you want to select the very best from the multitude of guidelines on management of MI. But what makes a good-quality guideline? Italian workers have proposed three criteria: whether the guideline (a) reports the range of professionals involved in development, (b) reports the strategy used to identify primary evidence and (c) explicitly grades the recommendations 2. Inspection of 431 specialty society guidelines published from 1988 to 1998 showed that 54% met none of the criteria, 34% met one, 7% met two and only 5% met all three. Seemingly, specialty guidelines have been dominated byèxperts', with little involvement of GPs and other users. Equally, most guidelines result from an informal me Âlange of opinion in the charged social atmosphere of a committee. One encouraging result in the Italian review was that the more recent guidelines were more likely to include details of literature search and graded recommendations. A UK group has developed a checklist of thirty-seven questions exploring three dimensions (rigour of development , clarity of content and context, documentation of methods for application and monitoring) [www.sghms. ac.uk/phs/hceu/]. In a study in which six clinicians used this`St George's checklist' to assess 60 guidelines the inter-rater agreement was excellent 4. However, even if a guideline is of high scienti®c quality, doctors may still not follow it. An observational study of Dutch clinicians suggested that key determinants of whether recommendations would be followed were that they were uncontro-versial, speci®c, evidence-based and required no change to existing routine 5. Even when we locate a satisfactory published guideline, it …