Abstract | The readmission rate is an important indicator of the hospital quality of care. With the upsetting increase in readmission rates worldwide, especially in geriatric patients, predicting unplanned readmissions becomes a very im-portant task, that can help to improve the patient’s well-being and reduce healthcare costs. With the aim of reducing hospital readmission, more atten-tion is to be paid to home healthcare services, since home healthcare pa-tients on average have more compromised health conditions. Machine Learning and Artificial intelligence algorithms were used to develop predic-tive models using MIMIC-IV repository. Developed predictive models ac-count for various patient details, including demographical, administrative, disease-related and prescription-related data. Categorical features were en-coded with a novel customized target encoding approach to improve the model performance avoiding data leakage and overfitting. This new risk-score based target encoding approach demonstrated similar performance to existing target encoding and Bayesian encoding approaches, with reduced data leakage, when assessed using Gini-importance. Developed models demonstrated good discriminative performance, AUC 0.75, TPR 0.69 TNR 0.67 for the best model. These encouraging results, as well as an effective feature engineering approach, can be used in further studies to develop more reliable 30-day readmission predictive models. |
---|