-
1
Vertebral compression fractures (VCFs) significantly increase mortality risk among elderly patients, necessitating improved prognostic tools.
-
2
A predictive model using machine learning was developed, achieving a C-index of 0.753, indicating strong predictive performance for long-term mortality.
-
3
Key predictors of mortality identified include age, sex, previous fractures, cancer history, and co-morbidities, enhancing patient risk stratification.
-
4
The model's effectiveness was validated through Kaplan–Meier survival analysis, demonstrating significant differentiation between high- and low-risk groups.
-
5
The explainable machine learning model aids clinicians in making individualized treatment decisions for elderly patients with VCFs.
