Impact of a clinical decision rule on antibiotic prescription for children with suspected lower respiratory tract infections presenting to European emergency departments: a simulation study based on routine data


Hagedoorn, Nienke N, Wagenaar, Josephine HL, Nieboer, Daan, Bath, David, Von Both, Ulrich, Carrol, Enitan D ORCID: 0000-0001-8357-7726, Eleftheriou, Irini, Emonts, Marieke, Van der Flier, Michiel, De Groot, Ronald et al (show 16 more authors) (2021) Impact of a clinical decision rule on antibiotic prescription for children with suspected lower respiratory tract infections presenting to European emergency departments: a simulation study based on routine data. JOURNAL OF ANTIMICROBIAL CHEMOTHERAPY, 76 (5). pp. 1349-1357.

[img] Text
20210107_Manuscript_FKTimplementation_rev_notrackchanges.Accepted.docx - Author Accepted Manuscript
Download (464kB)

Abstract

<h4>Background</h4>Discriminating viral from bacterial lower respiratory tract infections (LRTIs) in children is challenging thus commonly resulting in antibiotic overuse. The Feverkidstool, a validated clinical decision rule including clinical symptoms and C-reactive protein, safely reduced antibiotic use in children at low/intermediate risk for bacterial LRTIs in a multicentre trial at emergency departments (EDs) in the Netherlands.<h4>Objectives</h4>Using routine data from an observational study, we simulated the impact of the Feverkidstool on antibiotic prescriptions compared with observed antibiotic prescriptions in children with suspected LRTIs at 12 EDs in eight European countries.<h4>Methods</h4>We selected febrile children aged 1 month to 5 years with respiratory symptoms and excluded upper respiratory tract infections. Using the Feverkidstool, we calculated individual risks for bacterial LRTI retrospectively. We simulated antibiotic prescription rates under different scenarios: (1) applying effect estimates on antibiotic prescription from the trial; and (2) varying both usage (50%-100%) and compliance (70%-100%) with the Feverkidstool's advice to withhold antibiotics in children at low/intermediate risk for bacterial LRTI (≤10%).<h4>Results</h4>Of 4938 children, 4209 (85.2%) were at low/intermediate risk for bacterial LRTI. Applying effect estimates from the trial, the Feverkidstool reduced antibiotic prescription from 33.5% to 24.1% [pooled risk difference: 9.4% (95% CI: 5.7%-13.1%)]. Simulating 50%-100% usage with 90% compliance resulted in risk differences ranging from 8.3% to 15.8%. Our simulations suggest that antibiotic prescriptions would be reduced in EDs with high baseline antibiotic prescription rates or predominantly (>85%) low/intermediate-risk children.<h4>Conclusions</h4>Implementation of the Feverkidstool could reduce antibiotic prescriptions in children with suspected LRTIs in European EDs.

Item Type: Article
Uncontrolled Keywords: PERFORM consortium, Humans, Respiratory Tract Infections, Anti-Bacterial Agents, Retrospective Studies, Child, Emergency Service, Hospital, Europe, Netherlands, Clinical Decision Rules
Depositing User: Symplectic Admin
Date Deposited: 12 Feb 2021 11:01
Last Modified: 09 Mar 2023 08:10
DOI: 10.1093/jac/dkab023
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3115547
Dimensions
Altmetric
CORE (COnnecting REpositories)