Abstract
PURPOSE: Thoracic trauma is a frequent cause of emergency department admission in both civilian and military settings however injuries sustained in conflict zones differ in mechanism and severity. The STUMBL Score is a clinical prediction tool originally developed for blunt thoracic trauma in civilian populations. This study aimed to externally validate the STUMBL Score in a wartime context. METHODS: We conducted a retrospective, single-centre validation study of adult patients with thoracic trauma admitted to the ED of a frontline Ukrainian hospital between January 2023 and December 2024. The primary composite outcome was defined as in-hospital mortality, pulmonary complications, or ICU admission. Model performance was assessed by discrimination (area under the receiver operating characteristic curve) and calibration (Hosmer-Lemeshow test and calibration belt). RESULTS: A total of 374 patients were included (87% male; median age 38 [32-44]). Blast injury was the predominant mechanism. The median STUMBL Score was 30 [24-33], and 92% of patients developed the composite outcome. The area under the receiver operating characteristic curve (AUC) was 0.96 (95% CI 0.94-0.98), and calibration assessed by the Hosmer-Lemeshow test yielded a p-value of 0.812. Using a threshold score of 23, the sensitivity was 0.89 and specificity 0.97. CONCLUSION: The STUMBL Score demonstrated excellent predictive performance in a military population with high-acuity, war-related thoracic trauma. These findings support its potential utility in conflict-zone emergency care, although prospective validation in broader military settings is warranted.