A retrospective validation study of the STUMBL score for emergency department patients with blunt thoracic trauma

Introduction: Blunt thoracic trauma (BTT) is a leading cause of emergency department (ED) trauma-related attendance. Risk prediction tools are commonly to predict patients' outcomes and assign them to the most appropriate care setting. The STUMBL score is a prognostic model for BTT, derived and validated in the United Kingdom; items comprising the score are age, number of rib fractures, use of pre-injury anticoagulants, chronic lung disease and oxygen saturation levels. This study's aim was to validate the STUMBL score in an Italian ED.

Methods: This single-centre retrospective validation study was conducted in the ED of Santa Croce and Carle hub hospital in Cuneo, north-western Italy. All patients with an ED attendance for isolated BTT from 2018 to 2021 were included. Exclusion criteria were an age of under eighteen and the presence of any immediately life-threatening lesion. The primary outcome was the development of trauma-related complications, defined by the occurrence of one or more of the following: in-hospital mortality, pulmonary complications (infection, pleural effusion, haemothorax, pneumothorax, pleural empyema), need for intensive care unit admission, hospital length of stay equal to or greater than seven days. The performance of the STUMBL score was analysed in terms of discrimination with the evaluation of the receiver operating characteristic curve and calibration with the Hosmer-Lemeshow test and with the calibration belt.

Results: 745 patients were enroled (median age 64 [25th;75th percentile: 50;78], male/female ratio 1:4, median Charlson comorbidity index 2 [1;4], median STUMBL score 11 [6;17]). 65.2% of patients were discharged home after ED evaluation. 203 patients (27.2%) developed the primary outcome. The STUMBL score was significantly different in patients with complications compared to those without complications (9 [5;13] vs 21 [17;25], p < 0.001). The C index of the score for the primary outcome was 0.90 (95% CI 0.88-0.93), and the result of the Hosmer-Lemeshow test was 9.01 (p = 0.34). STUMBL score = 16 has a sensitivity of 0.80 (95% CI 0.75-0.85), specificity of 0.87 (95% CI 0.84-0.90), a positive predictive value of 0.70 (95% CI 0.64-0.76), and a negative predictive value of 0.92 (95% CI 0.90-0.94).

Conclusion: In this validation study, the STUMBL score demonstrated excellent discrimination and calibration in predicting the outcome of patients attending the ED with a BTT.