Antimicrobial Resistance Prediction in Urinary Tract Infections

Urinary tract infections (UTIs) represent a major global health burden, with rising rates of antimicrobial resistance (AMR) among common uropathogens undermining the efficacy of standard treatments. Traditionally, clinicians have relied on culture and sensitivity testing, which introduces delays in selecting effective empiric therapy. In response, computational approaches—especially machine learning and probabilistic modelling—have been developed to predict resistance profiles before laboratory results are available. These methods draw on electronic health records, patient demographics, prior antibiotic exposures and microbial genomic features to forecast nonsusceptibility and guide precision prescribing. By offering rapid, data-driven estimates of resistance risk, such tools promise to reduce inappropriate antibiotic use, improve patient outcomes and slow the emergence of multidrug-resistant organisms in both hospital and community settings.

Research from Nature Portfolio

Recent studies have applied machine learning and statistical modelling to anticipate antimicrobial susceptibilities before laboratory results become available. One approach developed personalized antibiograms from electronic health records spanning thousands of uUTI cases, demonstrating that tailored models can maintain high coverage while narrowing broad-spectrum prescriptions and reducing unnecessary antibiotic use. Another investigation implemented gradient-boosted decision tree models to predict ciprofloxacin resistance and extended-spectrum beta-lactamase production in patients presenting with UTIs in the emergency department; through threshold optimisation and feature selection, the model reduced the rate of ineffective empiric therapy by one-fifth and offers a point-of-care decision support tool. A complementary study employed additive Bayesian network modelling to map cross-resistance patterns of Escherichia coli isolates from urine and other clinical sources, revealing source-dependent magnitudes of co-resistance and highlighting the need to account for sample origin when estimating the likelihood of multidrug resistance.

Antimicrobial Resistance Prediction in Urinary Tract Infections publication trend

The graph below shows the total number of articles in antimicrobial resistance prediction in urinary tract infections across all publications each year (not limited to Nature Index journals).

Technical terms

Antimicrobial resistance: The capacity of microorganisms to survive exposure to drugs designed to kill them or inhibit their growth.

Machine learning: A field of artificial intelligence employing algorithms to identify patterns in data and generate predictive models.

Personalized antibiogram: An individualized prediction of a pathogen’s antibiotic susceptibility profile derived from patient-specific data.

Extended-spectrum beta-lactamase (ESBL): A group of enzymes produced by certain bacteria that confer resistance to beta-lactam antibiotics.

Bayesian network modelling: A probabilistic graphical approach for representing and analysing conditional dependencies among variables.

Interpretability: The degree to which a predictive model’s operations and outputs can be understood by humans.

References

1. Personalized antibiograms for machine learning driven antibiotic selection. Communications Medicine (2022).
2. Machine learning model for predicting ciprofloxacin resistance and presence of ESBL in patients with UTI in the ED. Scientific Reports (2023).
3. Bayesian network modeling of patterns of antibiotic cross-resistance by bacterial sample source. Communications Medicine (2023).
4. Interpretable machine learning-based decision support for prediction of antibiotic resistance for complicated urinary tract infections. npj Antimicrobials and Resistance (2023).
5. Development of Predictive Models to Inform a Novel Risk Categorization Framework for Antibiotic Resistance in Escherichia coli–Caused Uncomplicated Urinary Tract Infection. Clinical Infectious Diseases (2024).
6. Can the application of machine learning to electronic health records guide antibiotic prescribing decisions for suspected urinary tract infection in the Emergency Department?. PLOS Digital Health (2023).