Objective To evaluate the effects of standalone versus clinically integrated teaching in evidence based medicine on various outcomes in postgraduates. Design Systematic review of randomised and non-randomised controlled trials and before and after comparison studies. Data sources Medline, Embase, ERIC, Cochrane Library, DARE, HTA database, Best Evidence, BEME, and SCI. Study selection 23 studies: four randomised trials, seven non-randomised controlled studies, and 12 before and after comparison studies. 18 studies (including two randomised trials) evaluated a standalone teaching method, and five studies (including two randomised trials) evaluated a clinically integrated teaching method. Main outcome measures Knowledge, critical appraisal skills, attitudes, and behaviour. Results Standalone teaching improved knowledge but not skills, attitudes, or behaviour. Clinically integrated teaching improved knowledge, skills, attitudes, and behaviour. Conclusion Teaching of evidence based medicine should be moved from classrooms to clinical practice to achieve improvements in substantial outcomes.