# Prognostic accuracy of the CMPMIT-ICD-10, APACHE Ⅱ, SOFA, ISS, and AIS for in-hospital death among patients with traumatic hemorrhagic shock

## TL;DR

A new trauma scoring system called CMPMIT-ICD-10 was found to be more accurate than existing systems in predicting in-hospital death for patients with traumatic hemorrhagic shock.

## Contribution

The CMPMIT-ICD-10 trauma scoring system is introduced and shown to outperform established models like APACHE II, SOFA, ISS, and AIS in predicting mortality in traumatic hemorrhagic shock patients.

## Key findings

- CMPMIT-ICD-10 had the highest AUC (0.8757) for predicting in-hospital death among traumatic hemorrhagic shock patients.
- CMPMIT-ICD-10 showed high sensitivity (85.4%) and specificity (79.9%) at a cut-off value of 59.5.
- CMPMIT-ICD-10 was independently associated with in-hospital death (OR 1.057, p < 0.001).

## Abstract

The China Mortality Prediction Model in Trauma, which is based on the International Classification of Disease Disorders (ICD)-10-CM lexicon (CMPMIT-ICD-10),is a new trauma scoring system. Our objective was to compare the prognostic performance of the CMPMIT-ICD10 with that of the Acute Physiology and Chronic Health Evaluation II (APACHEII), Sequential Organ Failure Assessment (SOFA), Injury Severity Score (ISS), and Abbreviated Injury Scale (AIS) for in-hospital mortality in patients with traumatic hemorrhagic shock(THS).This retrospective observational cohort study was conducted at a tertiary teaching hospital from May 1, 2013, to May 31, 2023.The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, accuracy, and associations with outcomes of theCMPMIT-ICD-10, APACHE II, SOFA, ISS, and AIS scores for the prediction of in-hospital death were assessed. A total of 420 patients with THS were included. Forty-one (9.8%) patients died during hospitalization. For the prediction of in-hospital death, the CMPMIT-ICD-10 (0.8757) and APACHE II(0.8709) had greater AUCs compared with the AIS (0.6243), SOFA (0.7669), and ISS (0.6601). With the best cut-off value of 59.5, the CMPMIT-ICD10 had a highest sensitivity (85.4%) and good specificity(79.9%) and overall accuracy (80.4%). The CMPMIT-ICD10 (OR 1.057, 95% CI 1.028–1.087, p < 0.001) and APACHE II (OR 1.125, 95% CI 1.045–1.211, p = 0.002) were independently associated with in-hospital death. Comparable to the APACHE II but significantly better than the SOFA, ISS, and AIS, the CMPMIT-ICD-10 performed well in predicting the short-term mortality of patients with THS. These findings suggest that the CMPMIT-ICD-10 may have superior utility for predicting short-term death in THS patients.