Over two million biomedical papers are published each year, making it impossible for health care workers to keep up to date on current knowledge. Reviews are needed to provide manageable information on which decisions on health policy, and individual treatment, can be based. But how can the quality of these reviews be judged?