This thesis aims to contribute to improving the management of adverse drug events of hospitalised older patients, especially for drugs with anticholinergic properties, by identifying risk factors to develop and validate automatic prediction models. The thesis consists of four subprojects (Chapters II to IV) and is framed by a general introduction (Chapter I) and an overall discussion, conclusion and outlook (Chapter V). In the first subproject, we identified all published scales that score drugs from 0 to 3 depending on their anticholinergic activity. Further, we assessed their quality and impact on clinical outcomes. In the next projects, we used data from patients’ electronic health records from a Swiss tertiary teaching hospital to evaluate the association of all published scales with in-hospital mortality, length of stay and delirium during hospitalisation. Finally, we developed and validated prediction models to study their potential usefulness in clinics for tailored decision-making.