Abstract
Most of the guidelines in Grimshaw and Russell's review were not based on systematically reviewed evidence; however, it is a reasonable assumption that guidelines that are accurately based on evidence of effective treatment will benefit patients more than guidelines developed in an ad hoc manner or through informal consensus. The patchy nature of evidence, even in the best researched subjects in clinical practice, means that all guidelines in the conceivable future will be hybrid documents, with recommendations based on varying degrees of evidence and consensus. Good guidelines will clearly label recommendations according to strength of supporting evidence.