Abstract
Objective: To develop and internally validate a rapid, low‐cost model for early identification of urinary tract infection (UTI) in hospitalised adults using routinely available urinalysis parameters. Methods: This single‐centre retrospective case‐control study included 619 inpatients treated between January 2022 and May 2025. UTI cases required a compatible clinical diagnosis and a positive urine culture meeting the prespecified colony‐count threshold; controls did not meet these criteria. Admission urine specimens underwent automated reagent‐strip testing and automated particle analysis. Candidate predictors were age, sex, documented diabetes history and 12 routinely reported urinalysis variables. Least absolute shrinkage and selection operator (LASSO) regression with ten‐fold cross‐validation was applied in the 70% development set; retained predictors were entered into multivariable logistic regression and evaluated in a 30% hold‐out test set. Results: Six urinalysis variables—microscopic white‐blood‐cell grade, leukocyte esterase, urinary blood, nitrite, vitamin C and turbidity—were retained. The nomogram achieved an area under the receiver operating characteristic curve of 0.949 (95% CI: 0.922‐0.976) in the test set and showed close agreement between predicted and observed probabilities (mean absolute error: 0.016). Decision curve analysis suggested greater net benefit than ‘treat all’ or ‘treat none’ strategies across threshold probabilities of 10%–70%. SHAP analyses identified white‐blood‐cell grade and turbidity as the strongest contributors. Conclusion: A model based on six admission urinalysis variables showed strong internal discrimination for identifying UTI in this single‐centre sample. Prospective multicentre validation, local recalibration and workflow evaluation are required before clinical use.
Funding Information
  • Natural Science Foundation of Hunan Province (2026JJ82499)

This publication has 23 references indexed in Scilit: