Predictive Analytics for Pharmaceutical Supply Chain Resilience: AFramework for Reducing Medication Shortages in Rural andUnderserved U.S. Healthcare Facilities

Abstract

Medication shortages remain a persistent threat to continuity of care, patient safety, and health equity in the United States. Rural and underserved facilities are especially exposed because they operate with smaller inventories, weaker purchasing leverage, thinner pharmacy staffing, longer replenishment routes, and fewer clinically equivalent alternatives. This paper develops a predictive analytics framework for pharmaceutical supply chain resilience and evaluates it in a synthetic simulation environment calibrated to publicly reported shortage patterns and rural healthcare constraints. The framework integrates drug-level shortage-risk prediction, facility vulnerability assessment, inventory and substitution decision support, regional pooling, and equity-sensitive allocation governance. A multi-facility synthetic environment comprising 12 rural or underserved facilities, 60 essential medicines, and a 52-week planning horizon was constructed. Five policies were compared across 250 Monte Carlo replications: reactive baseline management; rule-based inventory triggers; predictive analytics; predictive analytics with regional pooling; and predictive analytics with equity-weighted pooling. The strongest policy reduced shortage-unit severity by 26.6%, the treatment-delay index by 30.0%, emergency orders by 8.5%, and disruption cost by 25.6% relative to reactive management. These results should be interpreted as simulation-based evidence of feasibility rather than as external validation of real-world effectiveness. The study contributes an operationally specified framework, a transparent mathematical formulation, and a reproducible synthetic evaluation architecture for future empirical testing using hospital pharmacy, wholesaler, and electronic health record data.