Prediction of Readmission Risk in Hospital Patients Using Artificial Intelligence Techniques

Hospital readmissions are a significant management challenge, with both clinical and economic implications. In this study, we apply artificial intelligence techniques to predict hospital readmissions in adult patients within 30 days of discharge using a dataset from Hospital Alma Máter de Antioquia, that includes demographic, clinical, and hospitalization information. The project tested several machine learning algorithms including logistic regression, decision trees and deep learning as neural networks. The results show that the neural network model achieved the best balance between the metrics of Specificity=84.0%, Precision=41.5%, Recall=98.7%, F1-Score=58.4%, accuracy=85.5% and AUC=94.7%. This study highlights the usefulness of artificial intelligence techniques for improving hospital readmission management.