Related Experiment Video
Updated: Mar 27, 2026
05:26
Author Spotlight: Innovations in iTUG Test for Enhanced Risk Assessment and Cognitive Insights
Published on: October 25, 2024
1.9K
A Systematic Review of Multivariate Models for Predicting Fall-Related Injuries in Older Adults
Yan Cai1,2,3, Wei Zhu4, Xue Zhang1
1Center of Gerontology and Geriatrics, West China Hospital, Sichuan University/West China School of Nursing, Sichuan University, Chengdu, Sichuan, China, scu.edu.cn.
Journal of Nursing Management
|March 25, 2026
Summary
Prediction models for fall-related injuries in older adults show varied accuracy and high risk of bias. Future research must prioritize developing robust, externally validated models for better clinical decisions and fall prevention.
Area of Science:
- Gerontology
- Biostatistics
- Epidemiology
Background:
- Numerous prediction models for fall-related injuries in older adults exist.
- Their clinical utility and research applicability remain uncertain due to quality concerns.
Purpose of the Study:
- Evaluate the quality, risk of bias, and clinical applicability of existing prediction models for fall-related injuries in older adults.
- Assess adherence to reporting guidelines and identify areas for improvement in model development.
Main Methods:
- Systematic literature search across multiple databases (Medline, Embase, Cochrane, CINAHL, Web of Science, Scopus) up to May 23, 2024.
- Included observational and experimental studies on multivariable prediction models for fall-related injuries in older adults.
- Assessed risk of bias and applicability using PROBAST; evaluated reporting quality with the TRIPOD+AI checklist.
Main Results:
- Fifteen studies reported 31 prediction models; discriminative statistics ranged from 0.54 to 0.89.
- All studies exhibited a high risk of bias (100%), particularly in the analysis domain.
- Median adherence to TRIPOD+AI reporting items was 56.4%, with low risk of applicability issues.
Conclusions:
- Existing fall-related injury prediction models have high bias and inconsistent accuracy, limiting clinical utility.
- Future models require rigorous internal and external validation, sufficient events per variable, proper missing data handling, and transparent reporting.
- Improved models are essential for data-driven clinical decisions and effective fall prevention strategies in older adults.
