TCCC Decision Support With Machine Learning Prediction of Hemorrhage Risk, Shock Probability

Area of Science:

• Military Medicine
• Medical Technology
• Artificial Intelligence in Healthcare

Background:

• Extended field care is required for combat casualties due to evacuation delays in remote conflicts.
• Complications such as shock, sepsis, and hemorrhage increase treatment complexity beyond initial stabilization.
• Existing systems may lack real-time monitoring and predictive capabilities for these complications.

Purpose of the Study:

• To develop and evaluate the Trauma Triage Treatment and Training Decision Support (4TDS) system for real-time casualty health monitoring.
• To integrate machine learning models for predicting shock, hemorrhage, and transfusion needs.
• To assess the usability and clinical relevance of the 4TDS system for medical personnel.

Main Methods:

• A mixed-methods approach including literature review, rapid prototyping, and agile development.
• Development of machine learning models using vital sign data to predict critical complications.
• Usability assessments conducted with military medics of varying expertise levels.

Main Results:

• Machine learning models for shock and hemorrhage/massive transfusion protocol achieved high validation accuracy (0.83 for shock).
• Usability assessments showed participants could accurately assess simulated casualties using the 4TDS prototype.
• Participants reported strong satisfaction with the system's fit with Tactical Combat Casualty Care (TCCC) and ease of use.

Conclusions:

• The 4TDS system, developed with participatory design, aligns with user workflows and mental models.
• Validation results indicate the system's readiness for potential FDA 510(k) clearance as a Class II medical device.
• The system shows promise for improving casualty care in prolonged field settings.