102 - PROV-UTICalc (A PROspective Validation of UTICalc for Children Aged 2 to 24 Months Presenting to the Emergency Department with Fever)

Publication Number: 102.6114

Ceilidh Kinlin, University of Ottawa, Ottawa, ON, Canada; Jocelyn Gravel, Université de Montreal, Montréal, PQ, Canada; Nicholas Barrowman, University of Ottawa Faculty of Medicine, Ottawa, ON, Canada; Natasha Wills-Ibarra, CHEO Research Institute, Ottawa, ON, Canada; Ramona L. Cook, CHU Ste Justine, montreal, PQ, Canada; Vid Bijelic, Children's Hospital of Eastern Ontario, Ottawa, ON, Canada; Nader Shaikh, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States; Roger Zemek, University of Ottawa Faculty of Medicine, Ottawa, ON, Canada; Maala Bhatt, Children's Hospital of Eastern Ontario, Ottawa, ON, Canada

Background: Febrile illness is a common presenting complaint to pediatric emergency departments (EDs) with 7% of cases attributed to urinary tract infections (UTIs). Diagnosing UTIs in pre-verbal children is challenging. UTICalc is a clinical decision support tool for children aged 2- 24 months with fever that estimates the pretest and post-test probability of UTI based on clinical and lab characteristics. The clinical model provides decision support for the need for urine testing and has a reported sensitivity of 96% and specificity of 35% for a 2% probability cutoff of UTI. The clinical + laboratory model incorporates urine test results to guide empiric treatment with a reported sensitivity of 95% and specificity of 92% for a 5% probability cutoff of UTI. No external prospective validation of UTICalc has been conducted to date.

Objective: To perform a multicenter, prospective external validation of UTICalc version 3.0.

Design/Methods: We conducted a multicenter prospective observational study in two Canadian tertiary EDs from October 2022 to October 2024, enrolling children aged 2-24 months presenting to the ED with fever (≥ 38.0oC) in the preceding 24 hours. Exclusion criteria included unmeasured fevers, known congenital kidney or urinary tract abnormalities, immunosuppression, antimicrobial use at enrolment, or insurmountable language barriers. Data were collected from parents, clinicians, and chart reviews. UTIs were defined by positive urinalyses (> trace leukocyte esterase OR > 5 WBC/hpf) AND significant culture growth of a uropathogen. Follow-up was conducted to detect missed UTIs. Performance characteristics of the tool were evaluated using previously published probability cutoffs.

Results: Of 2458 children enrolled, 2290 (median age 12 months, IQR 6-17; 47% female) were included in our analysis. Urinalysis was performed in 1062 (46%), cultures in 421 (18%), and 101 (4%) met criteria for positive UTI. The clinical model had a sensitivity of 96% (95%CI:92-99%), specificity of 34% (95%CI:32-36%) and overall area under the receiver operating characteristic curve of 83% (95%CI: 79-87%). The clinical + laboratory model had a sensitivity of 94% (95%CI:89-99%) and specificity of 87% (95%CI:85-90%) and overall area under the receiver operating characteristic curve of 96% (95%CI:94-98%).

Conclusion(s): This large, multi center external prospective validation demonstrated comparable sensitivities and only slightly reduced specificities compared to the original UTICalc derivation. These results reinforce the value of UTICalc as a clinical decision support tool for evaluating young febrile children presenting to the ED.