Digital Twins in Antimicrobial Stewardship: A Multi-Scale Review of Applications, Enabling Technologies and Implementation Barriers

Abstract

Objective: To synthesise contemporary and ongoing evidence on digital twin (DT) applications in antimicrobial stewardship (AMS) across the patient, ward, and health-system levels, and to differentiate digital twins conceptually from traditional artificial-intelligence (AI)-based stewardship decision support.

Methods: Narrative review of peer-reviewed and important and significant preprint literature issued up to July 2026, identified through PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, ScienceDirect, and medRxiv. Findings were organised thematically into patient-, ward-, and system-level applications, permitting technologies, and implementation obstacles.

Results: At the patient level, digital twins built on population pharmacokinetic-pharmacodynamic models and Bayesian forecasting support individualised antibiotic dosing, real-time exposure observation, and previous de-escalation. At the ward and hospital level, agent-based and graph-structured digital twins simulate nosocomial pathogen transmission, permitting infection-prevention interventions to be tested virtually before physical implementation. At the system level, emerging hospital- and population-scale digital twins integrate monitoring data to predict opposition trends and optimise stewardship policy. Persistent hurdles contain data quality and compatibility, model validation and clinician trust, computational cost, regulatory uncertainty, cybersecurity, and the risk of widening resource-based inequity.

Conclusion: Digital twins constitute a conceptually distinct and complementary advance over existing decision-support tools, with the potential to unify diagnostic, therapeutic, and policy dimensions of stewardship within a single adaptive framework. Rigorous prospective validation, governance structures, and equitable deployment strategies remain prerequisites for their responsible translation into routine clinical practice.