Comparative Diagnostic Accuracy of the STUMBL (Battle) Score and Chest Trauma Score for Predicting In-Hospital Mortality in Adult Blunt Thoracic Trauma: An Age-Stratified STARD-Compliant Study at an Indonesian Tertiary Referral Center

Abstract

Background: The Chest Trauma Score (CTS) evaluates anatomical injury, whereas the STUMBL (Battle) score also incorporates physiological and comorbidity parameters.

Objective: To compare their accuracy for predicting in-hospital mortality in adults with blunt thoracic trauma.

Methods: This retrospective study included 106 adults hospitalized at Dr. Mohammad Hoesin Hospital, Indonesia, from January 2024 to September 2026. Admission scores were evaluated using receiver operating characteristic curves, Youden thresholds, and age-stratified analyses (18–60 versus >60 years).

Results: Thirteen patients (12.3%) died. Non-survivors had higher median CTS (6.00 versus 4.00) and mean STUMBL scores (28.77 versus 14.89; both p < 0.001). STUMBL had an area under the curve (AUC) of 0.974 and CTS an AUC of 0.966 (pairwise p = 0.546). At thresholds of ≥25 and ≥6, respectively, sensitivity was 92.31% for both scores. Specificity was 91.40% versus 90.32%, positive predictive value was 60.00% versus 57.14%, negative predictive value was 98.84% versus 98.82%, and accuracy was 91.51% versus 90.57%. Among adults aged >60 years (n = 23), STUMBL ≥28 had specificity of 95.00%, compared with 75.00% for CTS ≥6; sensitivity was 100.00% for both.

Conclusion: Both scores showed excellent mortality discrimination in this cohort. STUMBL had numerically higher specificity in older adults; prospective validation is needed.