Machine Learning-Based Risk Prediction Model for Fatigue in Chronic Heart Failure Patients

Background Accurately identifying high-risk individuals with fatigue among patients with chronic heart failure (CHF) is crucial for improving their quality of life. This study aimed to construct a risk prediction model for fatigue in patients with CHF based on machine learning (ML) algorithms. Method The study population consisted of patients diagnosed with CHF at two tertiary hospitals in Yunnan from May 10, 2024, to October 31, 2024. LASSO (Least Absolute Shrinkage and Selection Operator) and logistic regression were employed for variable selection. Prediction models were developed and validated using five ML algorithms, and the model’s performance was assessed using several metrics, including the area under the receiver operating characteristic curve (ROC AUC), accuracy, sensitivity, specificity, F1 score, and brier score. SHAP (SHapley Additive exPlanations) plots were utilized for model interpretation. Results A total of 1171 CHF patients were included. Among the five ML models, Random Forest (RF) had the best predictive performance and was the optimal prediction model for fatigue in CHF patients. The best predictors identified included New York Heart Association (NYHA) classification, anxiety, sleep quality, depression, and activities of daily living (ADL). Conclusion The RF model demonstrated robust performance in predicting fatigue risk in CHF patients, providing a valuable tool for healthcare professionals to identify high-risk individuals and implement timely interventions. Trial registration: Not applicable