Machine Learning and Medical Authority Engagement for Antimicrobial Resistance Management: A Review of Surveillance, Prediction, and Stewardship

Abstract

Antimicrobial resistance (AMR) is a critical global health challenge, with particularly severe consequences in low and middle-income countries (LMICs) where surveillance infrastructure, diagnostic capacity, and stewardship resources remain constrained. This review synthesises recent advances in AMR surveillance, data integration, community-level antibiotic use, and the growing role of machine learning (ML) in resistance prediction and clinical decision support. We examine integrated digital platforms such as India’s i-DIA and international initiatives like the Fleming Fund that are bridging data fragmentation across healthcare systems. We survey ML approaches from supervised classifiers to ensemble methods, with particular attention to resource-appropriate frameworks operating under minimal data requirements. As a contextual case study, we describe AMRX—a calibrated probabilistic decision-support system designed to address the structural gap in empiric prescribing support for resource-limited environments. We further discuss socioeconomic barriers to antimicrobial access, community stewardship challenges, and evidence-based policy recommendations, aiming to assist researchers, clinicians, and policymakers in building effective, data-driven AMR control strategies.