Development and Validation of a Frailty Risk Prediction Model for Community-Dwelling Elderly in Shanghai

Abstract

Background: China is rapidly aging, increasing the burden on families, society, and public health services. The health of elderly individuals tends to deteriorate with age, and chronic conditions like frailty become more prevalent, driving up the use of healthcare services. Early screening and intervention for frailty are crucial in managing this demographic shift. While tools like the Fried Frailty Phenotype and Frailty Index assess frailty in communities, they are resource-intensive and only indicate frailty status without predicting risk or providing management recommendations. This study aims to develop a risk prediction model for frailty using real-world data, which can support the early detection of high-risk individuals in community settings. Objectives: To analyze the prevalence of frailty and its influencing factors in community-dwelling elderly, to construct a frailty risk prediction model and develop a nomogram, and to validate the model and assess its clinical utility. Methods: A cross-sectional survey of 420 elderly individuals in a Shanghai community health center was conducted (August 2022–March 2023). Data from various assessment tools were used to build a frailty prediction model through logistic regression, with validation conducted on 180 additional participants. The model’s predictive performance was evaluated using the ROC, AUC, calibration curves, and decision curve analysis (DCA). Results: The frailty prevalence was 7.4%. Independent risk factors included social support, malnutrition, fatigue, sarcopenia, reduced grip strength, and sleep duration. The prediction model achieved an AUC of 0.968 in the training set and 0.939 in the validation set, indicating high discrimination and calibration. DCA confirmed the model’s clinical utility. Conclusion: This study highlights a frailty prevalence rate of 7.4% among elderly individuals in Shanghai, with key risk factors identified. The validated frailty risk prediction model provides accurate and clinically effective frailty risk assessment, supporting targeted early interventions to prevent frailty in community settings.