Levels of evidence in cardiovascular clinical practice guidelines

BACKGROUND: Clinical practice guidelines (CPGs) can be helpful in distillin g the medical research literature for clinicians; however, the guidelines s hould acknowledge the variable methodological quality used in clinical rese arch by tempering their recommendations with a 'levels of evidence' scale. OBJECTIVE: To evaluate the proportion of English-language cardiovascular CP Gs that provide the user with recommendations graded according to a defined levels of evidence scale. In addition, to evaluate other key aspects impor tant in the critical appraisal of CPGs. METHODS: CPGs for atrial fibrillation, congestive heart failure and myocard ial infarction were identified by searching MEDLINE, a reference text of CP Gs and the Internet. Each CPG was evaluated using a priori-defined criteria based on the Evidence-Based Medicine Working Group's paper on critical app raisal of CPGs, including use of a reproducible search strategy, method of obtaining consensus, peer review and testing in practice. RESULTS: A total of 95 CPGs were evaluated. Only 13% graded their recommend ations using a defined levels of evidence scale. In addition, few CPGs docu mented a reproducible search strategy or peer review process, and none had been formally tested in practice. CONCLUSIONS: Reporting the levels of evidence for recommendations is an imp ortant component of CPGs, yet this system is not widely used.