Interpretable Machine Learning-based Decision Support for Prediction of Antibiotic Resistance for Complicated Urinary Tract Infections

Abstract

AbstractUrinary tract infections are one of the most common bacterial infections worldwide; however, increasing antimicrobial resistance in bacterial pathogens is making it challenging for clinicians to correctly prescribe patients appropriate antibiotics. In this study, we present four interpretable machine learning-based decision support systems for predicting antimicrobial resistance. Using electronic health record data from a large cohort of patients diagnosed with complicated UTIs, we demonstrate high predictability of antibiotic resistance across four antibiotics – nitrofurantoin, co-trimoxazole, ciprofloxacin, and levofloxacin. We additionally demonstrate the generalizability of our methods on a separate cohort of patients with uncomplicated UTIs. Through our study, we demonstrate that machine learning-based methods can help reduce the risk of non-susceptible treatment, facilitate rapid clinical intervention, enable more personalized approaches for treatment recommendation, while additionally allowing model interpretability that explains the basis for predictions.