Discussion: Methodologic guides for the conduct of systematic reviews are continuing to evolve. One recent contribution took place at the Postdam Consultation on Meta-analysis in March 1994. Beginning with prior guidelines and submissions from conference participants, we developed the next generation of guidelines for the conduct of systematic reviews of randomized control trials of health care. They are directed toward those who conduct, or are learning to conduct systematic reviews. Four principles guided this generation: that systematic reviews are retrospective research, and are potentially, subject to many of the same biases that affect case-control studies; that the health care question posed by the review will determine which studies and data are relevant, and how they should be synthesized; that teams with expertise in both the content area and methodology are required to conduct valid, useful systematic reviews; and that collaboration with the investigators who conducted the primary studies is necessary. In addition to considering the usual steps in preparing a systematic review these guidelines address new issues of special relevance to the Cochrane Collaboration related to the interpretation, dissemination and updating of systematic reviews, as well as suggestions regarding future methodologic developments.