Trust Formation Associated with Algorithmic Triage Acceptance in Emergency Service Users

Abstract

The rapid development and integration of clinical decision support systems have made algorithmic triage a viable solution for managing patient overflow in emergency medical services. However, the deployment of such systems is severely bottlenecked by the challenge of user acceptance, which depends heavily on the dynamics of trust formation. This study examines how emergency service users form, calibrate, and maintain trust when interacting with algorithmic triage applications. Utilizing a randomized, high-fidelity trial simulation, we evaluated the responses of emergency department service users who interacted with varying configurations of an automated triage system. The simulation systematically manipulated algorithmic transparency, feedback response times, and error rates across diverse clinical urgency scenarios. Quantitative outcomes measured through standardized trust scales and behavioral compliance metrics were combined with qualitative user experience data. The findings indicate that trust is a highly malleable, multi-dimensional construct that is heavily influenced by the explanation interfaces provided during the triage process. We show that explainable AI features significantly improve trust calibration, helping to prevent both over-reliance and under-reliance. Furthermore, cognitive load and baseline technological trust act as critical mediating variables in the trust-acceptance pathway. The insights gained from this simulation trial provide concrete, evidence-based recommendations for healthcare designers and clinical administrators aiming to implement safe, transparent, and highly accepted algorithmic triage systems.