COACT: Collaborative Objective AI-Assisted Clinical Team Assessment in Emergency Medicine

Ioan-Sorin Comsa, Ivan Moser, Tanja Birrenbach, Thomas C. Sauter, Per Bergamin

Confirmation: I have read and agree with the IEEE BSN 2025 conference submission's policy on behalf of myself and my co-authors.

Keywords: Team Performance Assessment, Wearable Signals, Pervasive Computing, Machine Learning, Pattern Analysis

Abstract: Effective teamwork is vital in emergency medicine, yet training interprofessional teams is challenging due to resource constraints, time limitations, and inconsistent assessment methods. Traditional performance evaluation relies on subjective ratings, often biased, and lacking objective validation. To address this, we introduce COACT (Collaborative Objective AI-Assisted Clinical Team Assessment), a data-driven framework for the evaluation of team dynamics in emergency medical team training. COACT enables debriefing by using wearable devices that capture physiological and motion data during training. The system processes raw signals in real time across multiple timescales to detect team dynamics. A Machine Learning (ML) model then analyzes the processed data, identifying synchronization patterns related to team coordination and effectiveness. We introduce the COACT index, a novel metric that quantifies how closely team patterns align with high-performing teams. The system operates on a continual ML paradigm, enabling self-monitoring, self-training, and self-adaptation as new data is collected. The results show up to 90% accuracy between the COACT index and subjective team performance ratings, validating the potential of the proposed framework as an objective AI-driven team assessment tool for emergency medicine training.

Track: 3. Signal processing, machine learning, deep learning, and decision-support algorithms for digital and computational health

NominateReviewer: Ivan Moser, [e-mail]

Submission Number: 4