Frontiers in public health · 2026
Dynamic risk prediction of physical fitness failure in college students: a four-year longitudinal landmark supermodel study.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Lu GuanYu, Zhang XuPeng, Han Shuo, Jia ZhenYu, Shi DanPing
Аннотация
PURPOSE: To develop and validate a dynamic risk prediction model (Landmark Supermodel) for identifying college students at risk of physical fitness failure by senior year, and to determine the optimal intervention window through sequential risk updating.
METHODS: A total of 4,006 college students (2,745 males; 1,261 females) with four-year longitudinal physical fitness test data were included. Ten physical fitness indicators were assessed annually. Sex-stratified Landmark Supermodels were constructed at three landmark time points using Random Forest and XGBoost classifiers with 57 engineered features encompassing static snapshots, cumulative summaries, change trajectories, variability, and interaction terms. Model performance was evaluated by AUC, Brier score, calibration curves, and decision curve analysis.
RESULTS: The Random Forest Landmark Supermodel achieved AUC of 0.842 (95% CI: 0.831-0.853) for males and 0.773 (95% CI: 0.737-0.818) for females. AUC improved from 0.751 (post-freshman) to 0.858 (post-sophomore) to 0.894 (post-junior) in males. At the junior-year landmark, risk stratification separated male students into high-risk (44.3% failure rate), medium-risk (6.7%), and low-risk (1.3%) groups. Vital capacity, endurance running performance, and BMI trajectories emerged as the most important predictive features.
CONCLUSION: Dynamic landmark risk prediction can accurately identify students at risk of physical fitness failure, with the sophomore year representing the optimal intervention window. The model provides an evidence-based tool for individualized early warning in physical education.
METHODS: A total of 4,006 college students (2,745 males; 1,261 females) with four-year longitudinal physical fitness test data were included. Ten physical fitness indicators were assessed annually. Sex-stratified Landmark Supermodels were constructed at three landmark time points using Random Forest and XGBoost classifiers with 57 engineered features encompassing static snapshots, cumulative summaries, change trajectories, variability, and interaction terms. Model performance was evaluated by AUC, Brier score, calibration curves, and decision curve analysis.
RESULTS: The Random Forest Landmark Supermodel achieved AUC of 0.842 (95% CI: 0.831-0.853) for males and 0.773 (95% CI: 0.737-0.818) for females. AUC improved from 0.751 (post-freshman) to 0.858 (post-sophomore) to 0.894 (post-junior) in males. At the junior-year landmark, risk stratification separated male students into high-risk (44.3% failure rate), medium-risk (6.7%), and low-risk (1.3%) groups. Vital capacity, endurance running performance, and BMI trajectories emerged as the most important predictive features.
CONCLUSION: Dynamic landmark risk prediction can accurately identify students at risk of physical fitness failure, with the sophomore year representing the optimal intervention window. The model provides an evidence-based tool for individualized early warning in physical education.