AI for Early Patient Screening and Risk Stratification in Antimicrobial Stewardship: Combating Antimicrobial Resistance Through Real-World Implementation

Antimicrobial resistance (AMR) is a major global health threat, with delayed or inappropriate antimicrobial therapy contributing significantly to morbidity, mortality and resistance development. This presentation explores how artificial intelligence (AI) can enhance antimicrobial stewardship (AMS) by enabling early patient screening, predictive risk stratification, optimised empirical therapy and real-time decision support. Drawing on global AMS frameworks—including the WHO AWaRe classification, the UKHSA Start Smart–Then Focus guidance, and the South Centre GUIDE framework—the session outlines how AI tools can operationalise evidence-based stewardship in clinical practice. Examples include sepsis early-warning systems, MDR risk prediction, dose optimisation algorithms, resistance forecasting and automated 48–72-hour antibiotic review prompts. Implementation challenges such as data quality, algorithmic bias, clinician trust, interoperability and equity considerations are discussed. A roadmap is presented for integrating AI into AMS workflows to improve patient outcomes, support policy alignment and reduce inappropriate antimicrobial use.