61 - Duration of Fever as an Indicator of Multisystem Inflammatory Syndrome in the Pediatric Emergency Department

Poster Number: 61
Publication Number: 61.408

Paige Kennedy, Children's National Health System, Arlington, VA, United States; Jaclyn Kline, Children's National Health System, Washington, DC, United States; Greg Bond, Children's National Health System, Annapolis, MD, United States; Kristen A. Breslin, Children's National Hospital, Washington, DC, United States

Background: Since the description of Multisystem Inflammatory Syndrome in Children (MIS-C) during the COVID-19 pandemic, providers have had to distinguish which febrile patients presenting to the pediatric emergency department (PED) are at risk for MIS-C. The CDC, Children’s National Hospital (CNH), and Children’s Hospital of Philadelphia (CHOP) have created algorithms to guide providers. These algorithms vary on duration of fever prompting testing; CDC requires 1 day while CNH and CHOP require 3 days. CNH and CHOP algorithms differ on the recommended tier one screening tests.

Objective: Determine the sensitivity and specificity of the CDC, CNH and CHOP algorithms for predicting MIS-C and compare the cost of laboratory workups.

Design/Methods: Retrospective review of the electronic health record at an academic PED and associated community satellite PED from May-July 2020. Encounters were included if the patient was < 21 years old, the reason for visit included fever or they had a triage temperature ≥38 C, and if reason for visit included abdominal pain or rash. We extracted demographic information, laboratory results, diagnoses, and disposition. Review was then performed to establish which patients would require evaluation under each algorithm and which patients were ultimately diagnosed with MIS-C. Cost per laboratory test was determined from the CNH laboratory test index.

Results: We identified 200 encounters that met inclusion criteria for fever and abdominal pain or rash. 53% met criteria for MIS-C per CDC’s algorithm and 27% per CNH/CHOP algorithms. CDC’s algorithm had a sensitivity of 0.86 (95% CI 0.58, 0.98) and specificity of 0.51 (0.44, 0.59), while CNH/CHOP had a sensitivity of 0.86 (0.58, 0.98) and specificity of 0.78 (0.72, 0.84). Adding the CHOP tier 1 testing decreased sensitivity to 0.80 (0.51, 0.94) and increased specificity to 0.98 (0.94-0.99). When patients presenting in shock were deemed to meet criteria despite only one system involvement, this increased the sensitivity of both to 1 (0.75, 1) and maintained the same specificities. Tier 1 defined by CNH is $3243 per patient and tier 1 testing defined by CHOP is $1092.Conclusion(s): Algorithms requiring 3 days of fever are more specific and equally sensitive at identifying MIS-C as those requiring 1 day of fever in our single-center sample. Including patients with one system involvement that presented in shock increased sensitivity with equal specificity. CHOP’s Tier one algorithm decreased overall cost of laboratory testing and increased specificity, however also marginally decreased sensitivity of identifying MIS-C.
Paige Kennedy CV 2022Kennedy, Paige_Final CV.pdf