Pharmacoepidemiology for oncology clinical practice: Foundations, state of the art and perspectives.

Since the early 2000s, the arrival of the so-called targeted therapies and immunotherapies have prolonged survival rates in many cancers. In parallel, post-marketing surveillance of anticancer drugs through pharmacoepidemiology has gradually developed. This paper provides (i) a detailed argumentation of the foundations for pharmacoepidemiology of anticancer drugs, (ii) an overview of pharmacoepidemiological studies currently available in this field, and (iii) some perspectives to improve pharmacoepidemiology for oncology practice. First of all, according to the existing literature, the development of pharmacoepidemiological studies for the clinical evaluation of anticancer drugs appears particularly justified based on common limitations of clinical trials in oncology regarding essential methodological principles such as adequate control groups, randomisation or double blinding. Many descriptive field cohort studies have investigated together treatment patterns, effectiveness, and safety to compare results from clinical trials with those of everyday practice. The utilisation of anticancer drugs has also been extensively described through cross-sectional or cohort studies by often using medico-administrative or medical databases. Such studies are useful to quantify and characterise use over time in the population, including clinically unvalidated use, and to evaluate adherence and persistence to increasingly available oral anticancer drugs. Despite their importance to increase knowledge, comparative effectiveness or safety studies remain uncommon. In a context of rapidly emerging therapies and personalised treatments, this may be due to methodological challenges especially related to the choice of a comparator or the consideration of confounding by indication. In the future, efforts must be pursued to provide real-time access to high-quality, large-scale clinical, biological and treatment data, and to improve record-linkage between hospital and outpatient databases. More research is also needed to better evaluate all medications, not only anticancer, as part of an overall cancer care pathway and to bring the evaluation of anticancer drugs closer to patients and society (social pharmacology).