Machine learning for predicting antimicrobialresistance in critical and high-prioritypathogens: A systematic review consideringantimicrobial susceptibility tests in real-worldhealthcare settings

This systematic review evaluates the effectiveness of machine learning (ML) in predicting antimicrobial resistance (AMR) in critical and high-priority pathogens (CHPP) in real-world healthcare settings. The review included 21 studies assessing 688,107 patients and 1,710,867 antimicrobial susceptibility tests. The top-performing ML models were GBDT, Random Forest, and XGBoost. GBDT showed the highest AuROC values compared to Logistic Regression, while Random Forest showed better AuROC values. However, limitations such as retrospective methodology, nonstandard data processing, and lack of validation in randomized controlled trials must be considered before applying these models in clinical practice.