Analyse et amélioration des outils d'aide à la décision en lien avec les médicaments pour les infirmiers

Details

Supervisors

Dalleur, Olivia

Faculty

Faculté de pharmacie et des sciences biomédicales

Degree label

Master [120] en sciences pharmaceutiques, à finalité spécialisée

Abstract

enThis thesis examines the clinical relevance of clinical decision support systems (CDSS) designed for nursing staff at the Cliniques universitaires Saint Luc, with a particular focus on the phenomenon of alert fatigue, which may undermine the effectiveness of these systems in securing the medication use process. In a context of increasing digitalisation of healthcare, the high rate of alert overrides raises concerns about how well these systems are adapted to the constraints and realities of nurses’ work in hospital settings. This single centre retrospective observational study is based on the analysis of 2,646 general pharmaceutical alerts (January 2025) and 7,814 “too close dosing” alerts (October–November 2025), generated from the electronic patient record used at the Cliniques universitaires Saint Luc. The fully anonymised data were processed using various Excel functionalities (filters, sorting, calculations, pivot tables), and the clinical relevance of the alerts was assessed in collaboration with a hospital pharmacist. The study protocol received approval from the Hospital Faculty Ethics Committee, ensuring compliance with ethical and regulatory requirements. The results show an overall alert override rate of 80.8%, of which 91.1% had no documented justification. Electrolytes account for 59.8% of general pharmaceutical alerts, while the analysis of “too close dosing” alerts indicates that 64.7% were clinically justified, alongside 18.7% technical errors (intervals < 5 minutes) and 13.4% false positives. For certain high risk medications, such as unfractionated heparin, the proportion of technical errors reaches particularly high levels, highlighting a misalignment between specific system settings and actual nursing practice. On this basis, the thesis proposes a progressive optimisation strategy specifically aimed at supporting nursing practice: prioritised removal of technical artefacts, refinement of alert thresholds according to therapeutic class and care context, and, in the longer term, the development of more intelligent and context aware systems that better reflect the clinical situation and the role of nurses in the medication use process. The overall objective is to reduce alert fatigue, improve the usability of CDSS for nurses, and strengthen medication safety at the patient’s bedside.