XGB model predicts mortality in elderly patients with vertebral compression fractures
This single-center retrospective cohort study developed and validated a predictive model for long-term mortality in 440 patients aged 65 years and older diagnosed with vertebral compression fractures between 2017 and 2020. The study compared five survival analysis models using a training set (n=296) and a validation set (n=144).
The XGB model demonstrated superior predictive performance compared to other models, with a C-index of 0.753 in the validation set. SHAP analysis identified age, sex, previous fracture, history of cancer, and co-morbidity as significant predictors of mortality. Kaplan-Meier survival analysis showed significant stratification of high- and low-risk groups (p < 0.05).
Safety and tolerability were not reported, as no interventions were tested. Key limitations include the single-center retrospective design, lack of external validation, unreported follow-up duration, and no information on mortality rates or absolute risk reductions.
The model may help identify high-risk elderly patients for targeted interventions, but it is not ready for clinical implementation without external validation. This is a predictive modeling study, not an interventional trial, and does not establish causation.