Integrating Heart Rate Variability and Exercise Load for Early Warning of Overtraining Syndrome

Abstract

Objective To develop an early warning model for overtraining syndrome (OTS) and provide OTS patients with a targeted and practical risk assessment tool, thereby reducing the incidence of OTS among athletes undergoing long-term excessive training and improving the quality of personalized care. Methods Seventy-two patients with OTS were selected. Eight indicators, including training load and heart rate variability, were chosen as model inputs. Relevant risk factors were identified using LASSO regression analysis, and early warning models were constructed using machine learning methods such as logistic regression and gradient boosting classifiers. Model performance was evaluated using metrics such as AUC, accuracy, sensitivity, and specificity, while calibration curves and receiver operating characteristic (ROC) curve analysis were employed to assess the model’s calibration and clinical utility.Results: Variables were screened via LASSO regression analysis, ultimately identifying eight factors—including training load and heart rate variability—as potential risk factors. The logistic regression model demonstrated the best performance (AUC = 0.841; 95% CI: 0.780–0.898) and demonstrated good clinical utility in both calibration curve and decision curve analyses. Conclusion: The OTS early warning model developed in this study facilitates the long-term, systematic collection of physiological data from athletes undergoing prolonged excessive training, promotes the clinical translation of early warning models, screens for high-risk OTS patients, guides nursing practice, and effectively reduces the incidence of OTS.