Building and validation based on machine learning methods: Predictive model for falls risk among community patients with chronic obstructive pulmonary disease in China

Background Falls in patients with chronic obstructive pulmonary disease (COPD) can have potentially devastating consequences; however, there is still a lack of accurate fall risk prediction models for community-dwelling patients with COPD in China. The aim of this study was to develop a risk prediction model for falls in COPD patients applicable to the Chinese community. Methods The clinical data of 809 Community COPD patients were analyzed by using the 2020 China Health and Retirement Longitudinal Study (CHARLS) database. The least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression were used to analyze predictors. Multiple machine learning (ML) classification models are integrated to analyze and identify the optimal model, and Shapley Additive exPlanations (SHAP) interpretation was developed for personalized risk assessment. Results The following eight variables: Memory_disease, Cardiology, Hyperlipidemia, Hypertension, Gender, Sleeping_time_at_night, ADL_score, and Age are predictors of falls in community-based COPD patients. Logistic classification model was the optimal model, test set area under curve (AUC) (95% confidence interval, CI): 0.693 (0.621–0.765), accuracy: 0.638, sensitivity: 0.627, and specificity: 0.642. Conclusions The model constructed in this study has relatively reliable predictive performance, which helps clinical doctors identify high-risk populations of community COPD patients prone to falls at an early stage.