Background: Whether routine urinary analysis has a net benefit for urinary tract infection (UTI) screening is unclear.
Methods: Using the laboratory information system (LIS), we retrospectively extracted the data of urine culture and routine analysis between January 2017 and April 2017. Receiver operating characteristic (ROC) curve, logistic regression model, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were used to estimate the screening performance of routine urinary analysis. Decision curve analysis (DCA) was used to estimate the net benefit of routine urinary analysis.
Results: A total of 927 specimens with 156 UTIs were included in the present study. The area under ROC curves (AUCs) of white blood cells (WBCs) and bacteria were 0.729 and 0.836, respectively. The logistic regression model incorporating WBCs, bacteria and nitrite together had an AUC of 0.851, which is significantly higher than that of WBCs. NRI and IDI analyses also indicated that WBCs, bacteria and nitrite, when used together, had better a screening performance than each single test alone. DCA revealed that 0.08 net benefit can be obtained for bacteria and the model, while the net benefit of WBCs is limited.
Conclusions: WBCs, bacteria and nitrite, when used together, can significantly improve the efficiency for UTI screening. Bacteria and the model incorporating WBCs, bacteria and nitrite have a net benefit in UTI screening, while the net benefit of WBCs, when used alone, is limited.
Keywords: Urinary tract infection (UTI); decision curve analysis (DCA); routine urine analysis; screening; sensitivity; specificity.
2020 Annals of Translational Medicine. All rights reserved.