Related Experiment Video
Updated: Jan 9, 2026
Author Spotlight: Innovations in iTUG Test for Enhanced Risk Assessment and Cognitive Insights
Published on: October 25, 2024
Performance of risk assessment algorithms recommended by the World Falls Guidelines (WFG) to predict fall and
Weiqiang Li1, Min Zhao1, Yanhong Fu1
1Department of Epidemiology and Health Statistics, Hunan Provincial Key Laboratory of Clinical Epidemiology, Xiangya School of Public Health, Central South University, Changsha, Hunan Province, China.
Introduction:
The World Falls Guidelines (WFG) Task Force published a falls risk stratification algorithm in 2022, but its predictive performance was reported only in Ireland, the United States, the Netherlands, Australia, and Malaysia.
Methods:
Using a nationally representative dataset, the China Health and Retirement Longitudinal Study (CHARLS), we analyzed data from six follow-up cohort visits (2, 3, 4, 5, 7, 9 years). The Cochran-Armitage trend test examined trends in fall and fall-related injury rate across the WFG algorithm. Multivariable logistic regression models examined associations between the WFG algorithm and fall and fall-related injury incidence rates. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and 95% confidence intervals (95% CIs) were calculated to assess predictive performance of the WFG algorithm. Sensitivity analyses were performed to assess the impact of missing values on principal findings.
Results:
We included 9,735, 5,377, 4,092, 9,426, 7,776, and 3,355 eligible older adults across the six follow-up time periods, with sample sizes varying due to the study's dynamic recruitment strategy. Fall risk categorized by WFG algorithm was significantly associated with falls and fall-related injuries at all six follow-up cohorts (p < 0.05). However, its predictive performance for both falls and fall-related injuries was unacceptable, with sensitivity ranging from 20.2% to 32.5% for both outcomes across the six follow-up visits. Sensitivity analyses displayed highly similar results.
Conclusion:
The WFG algorithm is valuable for predicting future falls and fall-related injuries among older Chinese community-dwellers, but its predictive performance is unacceptable for practical use without considering other contributing factors.
Practical Applications:
Further methodological modifications of the WFG algorithm are recommended to improve its predictive performance.
