Utilisation d'une approche Data Mining pour la spécification des durées en milieu hospitalier

Levecq, Philippe [FUCAM] Artiba, Abdelhakim [FUCAM] Meskens, Nadine [UCL]

Many authors agree to show the great variability of the length of stay in the operating room. Our researches aim at trying to explain this variability by taking into account factors linked to the patients (administrative data, medical information, …) and to the operating environment (surgeon, anaesthetist, …). A process of knowledge discovery based on data mining methods (rough sets, decision trees,…) is implemented. The improving of the prediction of the length of stay in the operating room is linked to the necessity to better schedule the operating activity. In this way, an architecture including scheduling algorithms and heuristics and simulation is presented.