UCAM develops an AI-based method that predicts relapses in ER patients
This software predicts that 7% of people treated in the emergency services of the Region of Murcia will return to them within one month of discharge, and can calculate the individual risk of relapse.
A multidisciplinary UCAM research led by doctors Juan José Hernández and Horacio Pérez, has developed a methodology based on Artificial Intelligence (specifically on machine learning) capable of predicting the relapse of patients who have been treated in the emergency rooms of the Region of Murcia with 95% accuracy.
This achievement has been possible thanks to the data provided by the Murcian Health Service. Said data has made it possible to create a system that estimates that around 7% of the patients treated in the ER, and who meet certain characteristics according to the algorithm, will need to be re-treated within 30 days of being discharged. In addition, the model can calculate the individual risk of relapse for each patient.
The system has been built upon a wide range of information, including clinical and demographic data. These include interventions performed during the hospital stay, time of admission and discharge, medical history, postcode, gender, age and habits such as smoking or alcohol consumption. The combination of all these factors allows the model to identify patterns that influence the likelihood of relapse after ED care.
The UITA, BIO-HPC and Hydro MRLab research groups have developed this tool, which will be made available to the Murcian Health Service or any other health service that wishes to use it to improve hospital management and citizen care.