Retrospective Evaluation of the Diagnostic Performance and Safety of a Reflex Urine Culture Algorithm in Pediatrics.
Academic Article
Overview
abstract
BACKGROUND: Obtaining urine cultures (UCx) in the absence of symptoms and inflammation can result in misdiagnosis of urinary tract infections (UTIs) and inappropriate use of antibiotics. Reflex UCx is a diagnostic stewardship strategy in which a UCx is performed only if a urinalysis (UA) meets UTI criteria. We retrospectively applied a reflex UCx algorithm to assess its performance and safety in children. METHOD: From 11/1/2020 to 7/23/2023, we studied children age > 24 months to <18 years with paired UA and UCx, i.e., a UTI evaluation, in emergency departments or inpatient units at our 8-hospital network. We assessed the diagnostic performance of pyuria of ≥10 WBC/hpf using a positive UCx as the reference standard; assessed reflex UCx exclusion criteria; and reviewed medical records of children with <10 WBC/hpf and positive UCx for signs and symptoms consistent with clinical UTI. RESULTS: Of 79 960 UA and UCx, 10 226 UTI evaluations were eligible for inclusion. Of the 8.7% (890/10226) positive UCx, 81.5% (725/890) had UA with ≥10 WBC/hpf and 18.5% (165/890) had UA with <10 WBC/hpf. The sensitivity of ≥10 WBC/hpf was 81% and the negative predictive value was 98%. Of 165 UTI evaluations with positive UCx and < 10 WBC/hpf, 26% (43/165) met reflex UCx exclusion criteria, 58 (35%) met neither exclusion criteria nor clinical UTI definition, and 39% (64/165) not meeting exclusion criteria fulfilled the clinical UTI definition, representing 0.6% (64/10226) of all UTI evaluations. CONCLUSIONS: Retrospective application of a reflex UCx algorithm demonstrated high NPV and sensitivity, suggesting the algorithm had low risk for missing clinical UTIs and could have avoided unnecessary UCx.
