Predicting the Risk of Readmission in Patients With Chronic Heart Failure

Heart failure (HF) is a frequent, serious, and costly chronic disease: it leads to 150,000 hospitalizations each year in France at a cost of 525 million Euros.

It is estimated that 20-40% of these hospitalizations are preventable by known interventions: home telemonitoring, care coordination, therapeutic intensification and therapeutic education. But these interventions only work if patients at high risk of rehospitalization are targeted to individualize management. In these patients, the risk of rehospitalization depends on clinical, biological, socioeconomic, care pathway, and location-related data. Existing predictive tools perform poorly due to three important limitations: non-use of unstructured clinical data, lack of integration of multimodal data, and weakness of the algorithmic approach.

The objective is to design and validate a predictive algorithm for the risk of rehospitalization in heart failure patients, using multiple data sources