Analysis and modeling of the key performance indicators in the emergency department

International audienceImproving emergency department (ED) processes requires consideration of multiple variables and objectives. EDs are known as a highly dynamic and unpredictable environment, which makes decision making extremely difficult. The use of different methodologies and tools to support the decision-making process is therefore essential. Nevertheless, medical staff are not adequately trained or prepared to solve such problems. They are completely destitute in methodologies and tools for decision support and management adapted to their future activities. The management of emergency systems generally involves three issues: their conception, planning and control. Conception is the definition/prediction of the future characteristics of the hospital system. Planning involves identifying the various resources used to perform care operations and defining how and when these operations should be performed. Finally, the control aims at correct the inherent disturbances (deterioration in the patient's state of health, new patients to treat). These three key steps are performed according to the frequency of problems encountered during the execution of the scheduling. In this paper, we focus on the first point, which is a presentation of a prediction model and a modeling of the patient pathway in the adult ED (AED) of the Lille University Hospital Center (LUHC). First, we present a statistical real data analysis. We use a real database of the AED provided by the LUHC, which is our partner. This Data collected over a period of four years, from June 2016 until June 2020, thanks to ResUrgence, a software implemented in LUHC. This analysis allows us to determine the most interesting aspects of the functioning of the AED