A machine learning-based Coagulation Risk Index predicts acute traumatic coagulopathy in bleeding trauma patients.

Area of Science:

• Trauma resuscitation
• Medical artificial intelligence
• Coagulation disorders

Background:

• Acute traumatic coagulopathy (ATC) is a critical condition following injury, increasing the risk of massive transfusion and mortality.
• Early identification of ATC is crucial for effective hemorrhagic shock management.

Purpose of the Study:

• To develop and validate a machine learning-based Coagulation Risk Index (CRI) for early prediction of ATC in bleeding trauma patients.
• To assess the accuracy of CRI in identifying patients at risk for critical administration thresholds and massive transfusion.

Main Methods:

• The CRI was developed using continuous vital signs (VSs) from photoplethysmographic and electrocardiographic waveforms, oximetry, and blood pressure trends over the first 15 minutes post-admission.
• Machine learning models were trained on 2 years of data and tested on 2 years of data from a single trauma center, including 17,567 patients.
• ATC was defined by international normalized ratio (INR) >1.2 and >1.5, and model accuracy was evaluated using area under the receiver operator curve.

Main Results:

• The CRI demonstrated excellent predictive accuracy for ATC, with true positive rates of 95.6% and 94.9% for INR >1.2 and >1.5, respectively, in critical administration threshold patients.
• For patients receiving massive transfusion, the CRI showed high true positive rates of 92.8% and 100% for INR >1.2 and >1.5, respectively.
• The study analyzed data from 17,567 patients, with 52.8% sustaining blunt injury and a mean age of 44.6 years.

Conclusions:

• The Coagulation Risk Index (CRI) effectively predicts early acute traumatic coagulopathy in bleeding patients using continuous vital signs and machine learning.
• Clinical implementation of CRI can guide timely hemostatic resuscitation strategies.
• Expanding CRI to prehospital settings may enable earlier intervention for ATC, potentially improving patient survival and outcomes.