Démarche pour implanter un système d'aide à la décision pharmaceutique : détecter et résoudre les problèmes liés à la pharmacothérapie.

Goals. To present for implementation of a Pharmaceutical Decision Support System (PDSS) that improves the detection and resolution of Drug-related problems (DRP). The aim is to improve the relevance of the patient's drug management. Methods. Over 4 years in two health facilities, pharmacists and IT professionals supported by the company Keenturtle (France) formalized the PDSS according to the active triangulation of a CDSS. Guidelines are defined to represent knowledge into pharmaceutical algorithms including human supervision. Terms not only in AI-pharmacy are defined in a glossary in order to contribute to users' training. Results. Since 2018, the PDSS is operational; it associates patient health data to pharmacotherapy knowledge in the deductive reasoning software Pharmaclass®. A guideline defined in 12 steps, helps the pharmacist to transpose clinical recommendations into 201 pharmaceutical algorithms which modeled patient situations as a part of the PDSS. A specific template of the pharmaceutical algorithms supports the detection and resolution of DRP. Moreover 41 terms are defined in a glossary. Conclusion. Defining a framework to implement and use a PDSS reduces its complexity. The knowledge representation strengthens the -pharmacists' expertise through its pedagogical side. This representation is a central point of the PDSS. Symbolic artificial intelligence approach will help pharmacists in their practice. [ABSTRACT FROM AUTHOR]