What question did this study set out to answer?

The aim was to evaluate an AI model's performance in predicting UTI diagnosis across different demographic groups in the emergency department.

May 8, 2026JMIR AIOpen Access

Evaluating the Potential Impact of AI on Urinary Tract Infection Diagnosis in the Emergency Department Across Demographic Groups: Retrospective Cohort Study

View Full Paper

Ask AI

Bookmark

Authors

Retrospective cohort study evaluates AI's diagnostic accuracy for UTI across demographics, indicating potential improvements.

The aim was to evaluate an AI model's performance in predicting UTI diagnosis across different demographic groups in the emergency department.
Multisite retrospective analysis of nonpregnant adults in ED with urinalysis and urine culture from 9 US health system EDs.
Developed an extreme gradient boosting model using data from June 2013 to August 2021, employing 5-fold cross-validation.
Compared AI model performance in diagnosis with that of physicians, using a composite proxy for physician diagnosis from UTI prescriptions.
Out of 149,449 encounters, 22,521 (15.1%) had positive cultures; 20,080 (13.4%) were diagnosed with UTI.
AI model achieved an area under the receiver operating characteristic curve of 0.93 (95% CI 0.93-0.93).
The AI model had lower overdiagnosis and underdiagnosis rates compared to physicians across all demographic groups.

Iscoe et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

May 8, 2026JMIR AIOpen Access

View Full Paper

Ask AI

Bookmark

Authors

Retrospective cohort study evaluates AI's diagnostic accuracy for UTI across demographics, indicating potential improvements.

The aim was to evaluate an AI model's performance in predicting UTI diagnosis across different demographic groups in the emergency department.
Multisite retrospective analysis of nonpregnant adults in ED with urinalysis and urine culture from 9 US health system EDs.
Developed an extreme gradient boosting model using data from June 2013 to August 2021, employing 5-fold cross-validation.
Compared AI model performance in diagnosis with that of physicians, using a composite proxy for physician diagnosis from UTI prescriptions.
Out of 149,449 encounters, 22,521 (15.1%) had positive cultures; 20,080 (13.4%) were diagnosed with UTI.
AI model achieved an area under the receiver operating characteristic curve of 0.93 (95% CI 0.93-0.93).
The AI model had lower overdiagnosis and underdiagnosis rates compared to physicians across all demographic groups.

Iscoe et al. (2026) studied this question.

Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark