Journal clubs (JCs) are a common form of interactive education in health care aiming to promote the uptake of research evidence into practice, but their effectiveness has not been established. This systematic review aimed to determine whether the JC is an effective intervention in supporting clinical decision making. We searched for studies which evaluated whether clubs promote changes in learner reaction, attitudes, knowledge, skills, behaviour or patient outcomes. We included undergraduate, postgraduate and practice JCs and excluded studies evaluating video/internet meetings or single meetings. Eighteen studies were included. Studies reported improvements in reading behaviour (N = 5/11), confidence in critical appraisal (N = 7/7), critical appraisal test scores (N = 5/7) and ability to use findings (N = 5/7). No studies reported on patient outcomes. Sixteen studies used self-reported measures, but only four studies used validated tests. Interventions were too heterogeneous to allow pooling. Realist synthesis identified potentially 'active educational ingredients', including mentoring, brief training in clinical epidemiology, structured critical appraisal tools, adult-learning principles, multifaceted teaching approaches and integration of the JC with other clinical and academic activities. The effectiveness of JCs in supporting evidence-based decision making is not clear. Better reporting of the intervention and a mixed methods approach to evaluating active ingredients are needed in order to understand how JCs may support evidence-based practice.