Prediction of Hospitalization Cost and Length of Stay for Patients with Heart Failure Using Deep Learning

Heart failure is a leading cause of hospitalization in Japan, and the related length of hospital stay is much longer than that in western countries. This brings a heavy burden on society. To improve the allocation of medical resources, a deep learning model with weighted predictors is developed for the prediction of hospitalization cost and length of stay using electronic health record. Through 5-fold cross validation, the proposed model demonstrates a better prediction performance than that of linear regression.