What question did this study set out to answer?

To develop machine learning models that accurately predict urinary tract infections using electronic health record data at the time of urine culture order.

January 14, 2026Open Forum Infectious DiseasesOpen Access

P-780. Machine Learning Models for Early Urinary Tract Infection (UTI) Prediction from Electronic Health Records

View Full Paper

Ask AI

Bookmark

Authors

NMNicholas P MarshallPalo Alto University FAFatemeh AmrollahiPalo Alto University FHFateme Nateghi HaredashtPalo Alto University

Discussion

Member takes

Implication

Predictive models improve UTI detection in adults using EHR data, suggesting better antibiotic stewardship implies enhanced decision-making.

Key Points

To develop machine learning models that accurately predict urinary tract infections using electronic health record data at the time of urine culture order.
Developed machine learning models using structured electronic health record data from adult patients.
Analyzed 300,381 urine cultures from 164,327 patients across two hospitals.
Models trained with XGBoost and assessed using ROC-AUC, NPV, and PPV metrics.
Baseline model achieved ROC-AUC of 0.73 with high NPV (95%) but low specificity (51%).
Model I improved ROC-AUC to 0.81, specificity to 67%, and PPV to 24%.
Model II, including prior antibiotic prescriptions, reached ROC-AUC of 0.89 with specificity of 80% and PPV of 35%.

Cite This Study

Marshall et al. (2026) studied this question.

synapsesocial.com/papers/6966f2e313bf7a6f02c0029d https://doi.org/10.1093/ofid/ofaf695.991

View Full Paper

Ask AI

Bookmark

Implication

Predictive models improve UTI detection in adults using EHR data, suggesting better antibiotic stewardship implies enhanced decision-making.

P-780. Machine Learning Models for Early Urinary Tract Infection (UTI) Prediction from Electronic Health Records

Discussion

Member takes

Implication

Key Points

Cite This Study

P-780. Machine Learning Models for Early Urinary Tract Infection (UTI) Prediction from Electronic Health Records

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider