Date : Dec. 21, 2020, 10 a.m. - BATATA Oussama - Visio-conférence
NonePrédiction et simulation de l’épuisement des aidants et organisation des services de répit
In France, 11 million people are caregivers for a relative for reasons of health or disability. The respite of caregivers becomes a social challenge. However, the number of respite services are still insufficient to address the needs of caregivers. The ``France Répit'' Foundation tackles this challenge and contributes to the development of solutions adapted to caregivers and their patients.
Our objective is to propose novel mathematical tools (i) to model, to predict and to simulate the exhaustion of caregivers and (ii) to evaluate the performance of respite services using these tools.
We first introduce the caregiver burnout assessment as a classification-based supervised learning problem to categorize caregivers according to their exhaustion. In the context of this study, classification models allowed us to list the risk factors for depression for caregivers. Then, to model the dynamic evolution of caregiver exhaustion, we introduce a mathematical modeling framework based on Markov chains. Such model allowed us to evaluate the performance of respite services. After validation of the model, we enrich the Markov chain with a discrete-event simulation algorithm to illustrate and evaluate caregivers' respite pathways according to several indicators: health (level of exhaustion of carers), respite resources usage and costs generated by caregivers exhaustion and respite services. Finally, we propose a predictive admission policy based on machine learning models. An agent-based simulation based on the previous Markov chain model is developed to evaluate this admission policy. A real case study based on the Lyon respite house is proposed to illustrate this research.
Andrea MATTA, Politecnico di Milano, Rapporteur
Feng CHU, Paris Saclay, Rapporteure
Farouk TOUMANI, LIMOS, ISIMA, Examinateur
Evren SAHIN LGI, CentraleSupélec, Examinatrice
Lionel PERRIER GATE, Centre Léon Bérard, Examinateur
Vincent AUGUSTO LIMOS, Mines Saint-Etienne, Directeur de thèse
Xiaolan XIE LIMOS, Mines Saint-Etienne, Co-directeur de thèse.