A number of concerns have recently been raised about whether or not antidepressant medications are associated with suicidal thoughts and behaviour in children and adolescents. These concerns are based largely on results of meta-analyses of randomized, controlled trials (RCTs). Controversy exists about generalizing evidence from short-term RCTs, designed primarily to test efficacy outcomes, to routine practice settings. Pharmacoepidemiological studies complement RCTs by using observational methods to examine safety and effectiveness of medications in the general population. This article reviews the contribution of pharmacoepidemiology to the controversy surrounding suicide risk in children and adolescents taking antidepressants, noting how variations in study design and adjustment for potential confounding factors influence outcome.