Grading of scientific evidence--a discussion paper from SHPIC

This paper examines some of the issues in the grading and use of evidence, and suggests that the (US) Agency for Health Care Policy and Research (AHCPR) system used by the Scottish Intercollegiate Guidelines Network (SIGN) and others is useful, but that it needs development to take account not only of type of study, but also of quality. It also considers the need to complement evidence of efficacy from randomised controlled trials, which may involve selected patients in selected centres, with evidence of "real-life" effectiveness. Furthermore, there are occasions where randomised controlled trials are not necessary. A grading system to take account of these issues is proposed for discussion.