Skip to content
Open access

Development and internal validation of a clinical prediction model for postoperative venous thromboembolism after neurosurgical intervention in patients with traumatic brain injury

Oct 2026 · Frontiers in Neurology · Vol 17 · 0 citations · 29 references
Medicine

Abstract

Background Venous thromboembolism (VTE) is a critical complication after traumatic brain injury (TBI) surgery. Conventional risk assessment tools were not developed for neurosurgical trauma populations, limiting their applicability. Objective To develop and internally validate an interpretable clinical prediction model for postoperative VTE in TBI surgery patients. Methods This retrospective cohort study enrolled 1,450 TBI patients who underwent surgery (2016–2024). Seven predictors were selected using Recursive Feature Elimination. Six machine learning algorithms were compared, with logistic regression (LR) selected as the final model. Performance was evaluated using AUC-ROC, calibration, and decision curve analysis. Results Among 1,450 patients, 205 (14.1%) developed postoperative VTE. The LR model achieved AUC-ROC of 0.868 (95% CI: 0.815–0.920) with good calibration (Brier score: 0.085). Seven predictors were identified: age, Barthel index, ASA classification, multiple surgeries, general anesthesia, limb AIS score, and serum magnesium. Compared with Caprini score (AUC: 0.686), the model showed improved discrimination (ΔAUC = 0.182; NRI: 0.559). Conclusion An interpretable model based on seven routinely available variables demonstrated promising performance for postoperative VTE risk stratification in TBI surgery patients. A nomogram and web-based calculator are provided for clinical use.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.