ARTIFICIAL INTELLIGENCE, PATIENT BEHAVIOR, AND HEALTHCARE ACCESS: A SIMULATION STUDY OF QUEUE ABANDONMENT AND TRIAGE

Abstract

This research explores the effect of AI-based triage process on queue performance in presence of reneging and balking phenomenon. There are long queues of patients in all the health systems around the world that results in patient dissatisfaction and delays in treatment leading to complications. The issue of the reneging (patients starting and leaving a queue) and balking (patients do not join long queues) will also make inefficiency to increase. Although classical queuing theory has provided some understanding of hospital processes (Green, 2006; Cochran & Bharti, 2006), new artificial intelligence (AI) toolsets have begun to offer a means to further enhance triage and prioritization among patients (Topol, 2019). Two options were compared with discrete-event simulation: (1) standard human triage and (2) AI-assisted triage for dynamic prioritization. Results indicate that AI-based triage does indeed compacts average wait times, by 18%, and reduces reneging rates, by 23%, preserving quality of service. The results of this study suggests that the introduction of AI into triage can improve patient flow and resource utilization, especially in busy hospital emergency departments (Wiler et al., 2010). We add to healthcare operations literature by offering simulation-based proof of concept that AI tools can be effective at managing patient queue dynamics. Implications include increased patient satisfaction, robustness of emergency workflows and guidance for policy on AI for the design of healthcare services.

Keywords: Healthcare queues, reneging, balking, AI-driven triage, simulation, emergency department efficiency