Development of a web platform for predicting fall risk in cardiovascular patients using machine learning.

The aim of this study was to develop and validate a machine learning-based fall risk prediction model for middle-aged and elderly individuals with cardiovascular disease. Data were sourced from the China Health and Retirement Longitudinal Study. Key predictive variables were identified using Least Absolute Shrinkage and Selection Operator regression. Six machine learning algorithms were employed to construct predictive models. Model performance was assessed through both internal and external validation, with SHapley Additive exPlanation values used to interpret the optimal model. A total of 1784 participants were analyzed, of which 434 (24.3%) experienced falls during a two-year follow-up period. Nine predictive factors were incorporated into the model, with the Light Gradient Boosting Machine model demonstrating the best performance, achieving areas under the receiver operating characteristic curve of 0.839 (95% CI 0.804-0.874) for internal validation and 0.816 (95% CI 0.798-0.833) for external validation. This model provides a scientific basis for personalized fall prevention strategies aimed at reducing fall incidence and enhancing the quality of life for patients with cardiovascular disease. Finally, we formed a web platform based on the best model to predict the probability of falls in cardiovascular patients.