SchizoReadPredict: A Tool for Prediction Readmission Risk of Schizophrenia Patients in a Spanish Region

Hospital readmission is currently a worldwide problem, specifically in health services such as psychiatry. Patients with mental disorders such as schizophrenia have high rates of hospitalization, which leads to increased costs in health systems and decreases the life quality of these patients. In this sense, the main objective of this study is development and evaluation of a web tool for prediction readmission risk of schizophrenia patients in the public hospital of Castilla y León. The development of the SchizoReadPredict tool is based on a Random Forest model optimized with the Particle Swarm Optimization algorithm. A database of hospitalizations of patients with schizophrenia in the public hospitals of Castilla and León was used. The database has 11126 records corresponding to 5412 patients. The results obtained show a high performance of the model with values of accuracy = 0.805, AUC = 0.848, recall = 0.800, F1-score = 0.790. The user interface of the tool is intuitive and easily accessible to determine the readmission risk of a patient with this disorder. This tool can help healthcare personnel in decision making and reduce the high costs associated with hospitalizations.