Implementation of a new approach of decision-support tools for the hospitalization of patients to GPs - 22/05/21
Résumé |
Introduction |
The objective of our study was to evaluate a suite of Machine Learning algorithms under R and TADA which is a predictive analysis tool developed by our company MyDataModels, for the implementation of a new approach to decision support tools for the hospitalization of patients to general practitioners. The idea is to proceed by extracting relevant knowledge with the help of GPs, for the prediction of a hospitalization event.
Methods |
We compare the impact of variables selected by experts compared to automatic learning methods through feature selection. The first step consists in creating a test cluster, then building several predictive models and then comparing these models in order to select a short list to be cross validated to evaluate their performance on hospitalization prediction.
Results |
Our approach has been tested on data from the PRIMEGE PACA database (www.primege.org/) that contains more than 600,000 consultations carried out by 17 general practitioners (GPs).
Conclusion |
Protocol in progress.
Le texte complet de cet article est disponible en PDF.Keywords : Algorithms, Machine Learning, Hospitalization, Predictive model, TADA-R
Plan
Vol 69 - N° S1
P. S53 - juin 2021 Retour au numéroBienvenue sur EM-consulte, la référence des professionnels de santé.