AI Algorithm Improves Antibiotic Decision-Making in Urinary Tract Infection

Urinary tract infections (UTIs) are among the most common bacterial infections worldwide and are a major driver of antibiotic use. Inappropriate or overly broad prescribing contributes to antimicrobial resistance, a growing global health threat responsible for millions of deaths annually. Clinicians must balance the need for effective treatment with the risk of promoting resistance, a decision that is often complex. Now, a new artificial intelligence (AI) system can help guide antibiotic selection for UTIs while prioritizing both patient safety and resistance prevention.

Researchers at the University of Liverpool (Liverpool, UK) have developed an AI-based algorithm that integrates clinical expertise with data-driven predictions to support treatment decisions for urinary tract infections. The system uses a mathematical utility function to weigh the benefits and risks of different antibiotic options for individual patients. It accounts for factors such as treatment effectiveness and resistance potential, blending human judgment with predictive modeling.

In simulation studies using real healthcare data, the AI’s recommendations performed as well as those made by physicians. However, the algorithm was less likely to select antibiotics associated with higher resistance risk and more likely to recommend oral treatments instead of intravenous options when appropriate. The study, published in npj Digital Medicine, also demonstrated a built-in safety feature: when patients were severely ill, the model prioritized selecting the most effective antibiotic to ensure optimal treatment outcomes.

Researchers suggest the tool could reduce unnecessary use of broad-spectrum antibiotics and slow the spread of antimicrobial resistance. By supporting precision prescribing, the system aims to improve both safety and convenience for patients. Further research across diverse healthcare settings is needed to validate the model globally, particularly in regions where antibiotic resistance has the greatest impact. The approach represents a step toward integrating AI with clinical expertise to strengthen infection management strategies.

“In an era where antimicrobial resistance continues to increase, innovative solutions to facilitate precision use of antimicrobials are required – our utility-based system may present such a solution,” said Dr. Alexander Howard, lead author of the study.

Related Links:
University of Liverpool