Use this URL to cite or link to this record in EThOS:
Title: Record linkage study to discriminate between community-acquired, healthcare-associated and hospital-acquired bloodstream infection in children in England
Author: Henderson, K. L.
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Current national guidance for empiric antibiotic therapy of children with bloodstream infections (BSIs) indicates that treatment options are dependent on whether a BSI is community-acquired (CA) or hospital-acquired (HA). Recent changes in healthcare delivery have seen an emergence of patients from the community with healthcare-associated (HCA) BSIs who require different antibiotics compared to 'classic' CA BSIs; research in this field has predominantly concentrated on adults. This thesis aims to predict the relative proportions of CA, HCA and HA BSIs in children aged 1 month to 5 years by identifying demographic and clinical characteristics associated with these three groups using two datasets. This thesis introduces the structure of the microbiology laboratory surveillance (LabBase2) and clinical (Hospital Episode Statistics) datasets and describes the probabilistic methodology used to link them, encompassing the match-weight calculation, threshold selection and evaluation of the final linked dataset. Three statistical approaches were used to predict CA, HCA and HA BSI using indicators of child susceptibility to infection, hospital exposure and timing of BSI. The combination of CA and HCA BSI accounted for 74% or more of all BSIs, but differences in the ranges of individual CA and HCA BSI estimates from each of the three statistical approaches were too wide (difference of 48%) to discern a consistent proportion of either CA or HCA BSI; conversely, the estimated proportion of HA BSI was consistently predicted (26% or less of all BSI). The cross-tabulation of child characteristics and invasive pathogens by predicted CA, HCA and HA BSI group identified only a few low prevalence risk factors (more than two chronic conditions, indwelling devices, hospital discharge in the month prior to the BSI) that had a high predictive value for HCA as opposed to CA BSI. Pathogens historically associated with causing HA BSI, particularly Gram-negative pathogens that are harder to treat with antibiotics, were frequently isolated on or before the day of hospital admission. This thesis demonstrated the wide range of pathogens isolated before and on admission to hospital, reflecting the growing mix of children acquiring CA and HCA BSIs in England in the community setting. A few, low-frequency clinical characteristics were predictive of CA, HCA or HA BSI, however, additional data from other sources (e.g. outpatient, primary care), may help to increase the accuracy of the prediction models. The mix of pathogens isolated from children is likely to become more heterogeneous as healthcare provision outside of hospital increases, suggesting that the concept of using clinical characteristics to identify the source of BSI is becoming less relevant. This highlights the importance of developing rapid indicators, such as early bedside testing, to inform appropriate treatment strategies.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available