Related Experiment Video
Updated: May 15, 2026
09:51
The Transition to an Anterior-Based Muscle Sparing Approach Improves Early Postoperative Function but is Associated with a Learning Curve
Published on: September 7, 2022
MMPro-HIP: multimodal progressive fusion model for elderly HIP fracture risk prediction
Songyuan Chen1, Ziqi Liu2,3, Yifan Cao1,2
1School of Artificial Intelligence, Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing, Beihang University, Beijing, China.
Frontiers in Medicine
|May 14, 2026
Summary
A new model, MMPro-HIP, accurately predicts hip fracture risk in older adults, even with missing data. Bone mineral density and demographics are key predictors, improving upon traditional methods for better patient care.
Area of Science:
- Gerontology
- Orthopedics
- Biomedical Informatics
Background:
- Hip fractures carry high mortality and disability risks in older adults.
- Accurate risk prediction is crucial for preventive measures and individualized care.
- Incomplete clinical data in older adults complicates risk modeling.
Purpose of the Study:
- Develop a robust hip fracture risk prediction model for older adults.
- Address challenges posed by missing clinical and imaging data.
- Improve generalizability and feasibility of risk assessment in diverse settings.
Main Methods:
- Retrospective analysis of 1,287 elderly patient records.
- Development of a global machine learning model using complete data.
- Proposal of a progressive fusion model (MMPro-HIP) for incomplete data.
Main Results:
- Global model achieved 84.67% accuracy and 0.8064 AUC.
- MMPro-HIP model demonstrated superior performance with 90.94% accuracy and 0.9423 AUC.
- Bone mineral density and demographic variables were significant predictors.
Conclusions:
- MMPro-HIP shows strong predictive performance for hip fracture risk in older adults with incomplete data.
- Progressive multimodal fusion is a practical strategy for clinical prediction under missingness.
- External validation of the MMPro-HIP model is recommended.
