Before being launched on the market, a medicinal product must have been demonstrated of good quality, effective and safe. However, rare adverse drug reactions or those occurring after long-term use may not always be detected during the clinical development of a drug. Therefore, continuous drug safety monitoring, or pharmacovigilance, is essential during the whole lifecycle of a drug, including after the market authorization. This thesis provides an overview of different tools available for the postmarketing surveillance of drug safety. Classic pharmacovigilance is based on spontaneous reports of individual cases. Postmarketing studies are also useful to evaluate safety in certain populations (e.g. children, pregnant women, elderly) or to assess the impact of certain risk factors (e.g. pharmacogenetics). On the other hand, drug-drug interaction can lead to toxicity or to therapeutic failure when the systemic exposure of the victim drug is modified by a perpetrator drug. Finally, data and text mining strategies are useful when a large amount of data have to be explored, such as global pharmacovigilance databases, electronic health records, or social media.