Selecting appropriate empirical antibiotic regimens for paediatric bloodstream infections: application of a Bayesian decision model to local and pooled antimicrobial resistance surveillance data

Abstract

Objectives: The objective of this study was to evaluate the ability of weighted-incidence syndromic combination antibiograms (WISCAs) to inform the selection of empirical antibiotic regimens for suspected paediatric bloodstream infections (BSIs) by comparing WISCAs derived using data from single hospitals and from a multicentre surveillance dataset.

Methods: WISCAs were developed by estimating the coverage of five empirical antibiotic regimens for childhood BSI using a Bayesian decision tree. The study used microbiological data on ∼2000 bloodstream isolates collected over 2 years from 19 European hospitals. We evaluated the ability of a WISCA to show differences in regimen coverage at two exemplar hospitals. For each, a WISCA was first calculated using only their local data; a second WISCA was calculated using pooled data from all 19 hospitals.

Results: The estimated coverage of the five regimens was 72%-86% for Hospital 1 and 79%-94% for Hospital 2, based on their own data. In both cases, the best regimens could not be definitively identified because the differences in coverage were not statistically significant. For Hospital 1, coverage estimates derived using pooled data gave sufficient precision to reveal clinically important differences among regimens, including high coverage provided by a narrow-spectrum antibiotic combination. For Hospital 2, the hospital and pooled data showed signs of heterogeneity and the use of pooled data was judged not to be appropriate.

Conclusions: The Bayesian WISCA provides a useful approach to pooling information from different sources to guide empirical therapy and could increase confidence in the selection of narrow-spectrum regimens.