Pharmacovigilance, which is the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other possible drug-related problems, generates knowledge to facilitate the rational use of medicines. When a medicine is first marketed, there is limited information on adverse drug reactions (ADRs), especially in paediatrics, where medicines are less likely to have been extensively studied. Knowledge in drug safety is built up over time when more (in number, and more heterogeneous) patients are treated than were studied in the randomised controlled trials preceding the marketing of a medicine. Previously not recognised ADRs are often initially described in case reports and case series. Prospective cohort studies are useful in determining the incidence and risk factors of common ADRs. Case series and pharmacovigilance reporting systems have been useful in identifying previously unknown uncommon ADRs and risk factors for specific ADRs. This brief review provides examples that illustrate how various study designs and data sources contribute to the evolving knowledge of ADRs that is essential to help develop guidelines and improve the rational use of medicines.