Représentation et gestion des connaissances pour une aide informatisée à la prescription et la surveillance

This paper describes SEPIA whose goal is to provide an intelligent support to monitoring and treatment of a patient enduring chemotherapy protocols in hemato-oncology. SEPIA is able to memorize prescriptions at different levels of abstraction, and has critiquing facilities with respect to drug and test prescriptions. SEPIA performs an assessment of the underlying patient state, in order to detect trends or complications that it will try to prevent by providing appropriate recommendations. Intelligent user interfaces are needed to facilitate access and maintenance of the large knowledge bases used.