Trauma THOMPSON: Clinical Decision Support for the Frontline Medic

Area of Science:

• Medical technology
• Military medicine
• Artificial Intelligence

Background:

• Military healthcare providers face challenges in prolonged casualty care due to extended evacuation times.
• Inexperienced providers may deliver critical care in austere environments, necessitating enhanced support.
• Telemedicine, AI, and AR offer potential solutions to improve competence and confidence in critical care delivery.

Purpose of the Study:

• To develop the Trauma TeleHelper for Operational Medical Procedure Support and Offline Network (THOMPSON) application.
• To create a voice-controlled mobile app with video references, audio narration, and AI-guided AR interventions.
• To provide offline procedural guidance for prolonged casualty care in communication-limited settings.

Main Methods:

• Development of a voice-controlled mobile application with video and audio procedural guidance.
• Integration of an AI-guided intervention system using augmented reality (AR) overlay and voice commands.
• Collection of procedural videos for reference materials and machine learning algorithm training.

Main Results:

• Preliminary video collection initiated for critical procedures like tube thoracostomy and cricothyrotomy.
• Machine learning algorithm achieved 20.1% action recognition and 11.4% anticipation accuracy in initial 'in the wild' datasets.
• The system demonstrates performance over 100 times better than random prediction.

Conclusions:

• The THOMPSON platform aims to deliver real-time, offline clinical decision support.
• It addresses the gap in available tools for remote and communication-limited environments.
• The technology has the potential to significantly benefit military medical providers during prolonged casualty care scenarios.