Meeting the 24‐Hour Movement Guidelines and Frailty Risk: Evidence From Two Nationally Representative Population Studies and Machine‐Learning Models
Jiulong Song, Yuehao Chen, Yuxin Fang, Guoguo Zhao, Yuxin Zhang, Jian FuABSTRACT
Aim
This study aimed to examine the association between meeting the 24‐Hour Movement Guideline components—moderate‐to‐vigorous physical activity (MVPA), sleep duration, and sedentary behavior (SB)—and frailty across two nationally representative populations.
Methods
We analyzed 11 893 National Health and Nutrition Examination Survey (NHANES) participants (2011–2018) and 4823 China Health and Retirement Longitudinal Study (CHARLS) participants (baseline 2011 with follow‐up through 2018). The reference category was meeting none of the three guideline components. Cross‐sectional associations were examined using multivariable logistic regression, and longitudinal associations were examined using Cox proportional hazards models. Exploratory machine‐learning analyses were conducted to evaluate frailty classification and feature importance.
Results
In cross‐sectional analyses, meeting all three components was associated with lower odds of frailty in NHANES (OR = 0.21; 95% CI, 0.14–0.32) and CHARLS (OR = 0.29; 95% CI, 0.16–0.46). Meeting two‐component combinations was also associated with lower odds of frailty in both cohorts. In CHARLS longitudinal analyses, over a median follow‐up of 7.0 years, 1457 incident frailty cases occurred. Compared with participants meeting none of the three components, those meeting all three components had a lower hazard of incident frailty (HR = 0.33; 95% CI, 0.27–0.41). In exploratory machine‐learning analyses, the Random Forest model showed the highest overall performance, and SHAP analyses identified 24‐Hour Movement Guideline adherence and MVPA among relevant features for frailty classification.
Conclusion
Meeting multiple 24‐Hour Movement Guideline components was associated with lower frailty odds and incident frailty hazard. These findings support considering movement behaviors within an integrated 24‐h framework.