IDDF2024-ABS-0007 Establishment and evaluation of a predictive model for acute bowel injury after multiple trauma

<h3>Background</h3> The aim of this study is to construct a prognostic model for predicting the incidence of acute bowel injury (ABI) in patients with multiple injuries through data analysis. The performance of the model will be assessed and validated for clinical application. <h3>Methods</h3> This study retrospectively analyzed the clinical data of 389 polytrauma patients from the mimic-IV database between 2008 and 2019. Patients were categorized into ABI and N-ABI groups based on the presence of acute bowel injury. Logistic regression analysis was performed to construct a predictive model by incorporating multiple independent risk factors. The model’s predictive accuracy was evaluated using the receiver operating characteristic curve (ROC), and the optimal cutoff value was determined by maximizing the Youden index. Furthermore, the model’s performance was assessed using the Omnibus test and the Hosmer-Lemeshow test. <h3>Results</h3> Among the polytrauma patients included in the study, 74.29% (289 patients) had concurrent ABI, while 25.71% (100 patients) had uncomplicated non-ABI cases. The univariate analysis revealed significant differences (P &lt; 0.05) in the Glasgow Coma Score (GCS), Sequential Organ Failure Score (SOFA), Acute Physiology and Chronic Health Score (APACHE II), shock index, mechanical ventilation, length of stay in the ICU, procalcitonin (PCT), and lactate (Lac) levels between the ABI and non-ABI groups. Logistic regression analysis confirmed that APACHE II score, mechanical ventilation, PCT, and Lac were independent risk factors associated with the occurrence of acute bowel injury post-polytrauma (P &lt; 0.05). A predicted logit(P) value &gt; 0.51 indicated the likelihood of ABI occurrence. The Omnibus test, χ2 = 93.1, P &lt; 0.05, demonstrated statistical significance and the Hosmer-Lemeshow test, χ2 = 3.7, P = 0.9 &gt; 0.05, indicated a good model fit. The area under the ROC for the predicted probability was 0.88, with a 95% confidence interval of (84.1%-94.1%). <h3>Conclusions</h3> The prevalence of ABI is elevated in polytrauma patients. The proposed model for prognosticating the occurrence of acute bowel injury after polytrauma can serve as a valuable clinical tool for predicting its incidence.