![[2602.17962] Improving Generalizability of Hip Fracture Risk Prediction via Domain Adaptation Across Multiple Cohorts](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.17962] Improving Generalizability of Hip Fracture Risk Prediction via Domain Adaptation Across Multiple Cohorts
Summary
This article presents a study on improving hip fracture risk prediction models through domain adaptation techniques, demonstrating enhanced generalizability across diverse cohorts.
Why It Matters
The research addresses a critical issue in clinical risk prediction, where models often fail to perform across different patient populations. By enhancing model generalizability, this work has implications for better patient outcomes and more effective healthcare interventions, particularly in osteoporosis management.
Key Takeaways
- Domain adaptation methods significantly improve hip fracture risk prediction across different cohorts.
- Combining multiple adaptation techniques yields the best performance, with AUC scores reaching 0.95.
- The study introduces outcome-free approaches that enhance model selection under realistic deployment conditions.
Computer Science > Machine Learning arXiv:2602.17962 (cs) [Submitted on 20 Feb 2026] Title:Improving Generalizability of Hip Fracture Risk Prediction via Domain Adaptation Across Multiple Cohorts Authors:Shuo Sun, Meiling Zhou, Chen Zhao, Joyce H. Keyak, Nancy E. Lane, Jeffrey D. Deng, Kuan-Jui Su, Hui Shen, Hong-Wen Deng, Kui Zhang, Weihua Zhou View a PDF of the paper titled Improving Generalizability of Hip Fracture Risk Prediction via Domain Adaptation Across Multiple Cohorts, by Shuo Sun and 10 other authors View PDF Abstract:Clinical risk prediction models often fail to be generalized across cohorts because underlying data distributions differ by clinical site, region, demographics, and measurement protocols. This limitation is particularly pronounced in hip fracture risk prediction, where the performance of models trained on one cohort (the source cohort) can degrade substantially when deployed in other cohorts (target cohorts). We used a shared set of clinical and DXA-derived features across three large cohorts - the Study of Osteoporotic Fractures (SOF), the Osteoporotic Fractures in Men Study (MrOS), and the UK Biobank (UKB), to systematically evaluate the performance of three domain adaptation methods - Maximum Mean Discrepancy (MMD), Correlation Alignment (CORAL), and Domain - Adversarial Neural Networks (DANN) and their combinations. For a source cohort with males only and a source cohort with females only, domain-adaptation methods consistently showed improved...