Machine learning-based risk prediction of mild cognitive impairment in patients with chronic heart failure: A model development and validation study

Abstract

Accurate identification of individuals at high risk for mild cognitive impairment (MCI) among chronic heart failure (CHF) patients is crucial for reducing rehospitalization and mortality rates. This study aimed to develop and validate a machine learning model to predict MCI risk in CHF patients. 602 CHF patients were included in this cross-sectional analysis. We constructed four machine learning models and assessed the models using the area under the receiver operating characteristic curve (AUC), calibration curve, and clinical decision curve. Results showed that scores of psychological and social adaptation management, age, free triiodothyronine, Self-rating Depression Scale scores, hemoglobin, sleep duration per night and gender were the best predictors and these factors were used to construct dynamic nomograms. Among all models, eXtreme Gradient Boosting (XGBoost) with an AUC of 0.940 performed the best in predicting the risk of MCI in CHF patients. Dynamic nomogram helps clinicians perform early screening in large populations.