A Human-Centered AI Coordination Layer for Healthcare Supply Chains During Crises: A Comparative Analysis and Pilot Framework
TL;DR
The paper argues that instead of replacing existing systems, the proposed AI solution should function as a human-monitored coordination layer built on LMIS data, using demand forecasting, shortage prediction, route optimization, and anomaly detection.
Abstract
Healthcare resource allocation during crises fails because decisions are made by separate systems operating in silos, limiting the ability to adapt quickly as demand changes. This problem disproportionately affects low- and middle-income countries (LMICs), where a lack of infrastructure leads to more logistical failures. This paper asks how healthcare distributors can improve resource allocation during crises, comparing existing solutions of government stockpiling, mutual aid agreements, logistics management information systems (LMIS), blockchain, localized 3D printing, and a new proposed solution of artificial intelligence (AI) using Ben Shneiderman’s reliability, safety, and trust (RST) framework, as well as considerations regarding LMIC feasibility. The paper argues that instead of replacing existing systems, the proposed AI solution should function as a human-monitored coordination layer built on LMIS data, using demand forecasting, shortage prediction, route optimization, and anomaly detection. The paper’s original contribution is a system architecture and pilot framework for a regional network of hospitals. This framework includes measurable outcomes such as shortage days, forecast error, response time, emergency shipments, waste, equity gaps, and human overrides. Because no live deployment has been tested, the paper’s conclusions are conceptual and not empirical. Yet, the comparisons suggest that a coordinated, transparent, and human-centered AI model may be more realistic and defensible than treating AI as a standalone replacement in an industry where trust is so vital.