What counts as reliable evidence for public health policy: the case of circumcision for preventing HIV infection

BACKGROUND: There is an ongoing controversy over the relative merits of randomized controlled trials (RCTs) and non-randomized observational studies in assessing efficacy and guiding policy. In this paper we examine male circumcision to prevent HIV infection as a case study that can illuminate the appropriate role of different types of evidence for public health interventions. DISCUSSION: Based on an analysis of two Cochrane reviews, one published in 2003 before the results of three RCTs, and one in 2009, we argue that if we rely solely on evidence from RCTs and exclude evidence from well-designed non-randomized studies, we limit our ability to provide sound public health recommendations. Furthermore, the bias in favor of RCT evidence has delayed research on policy relevant issues. SUMMARY: This case study of circumcision and HIV prevention demonstrates that if we rely solely on evidence from RCTs and exclude evidence from well-designed non-randomized studies, we limit our ability to provide sound public health recommendations.