Mathematical models of Clostridium diffcile transmission
Date
2018
Authors
McLure, Angus
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Clostridium difficile infections (CDIs) are some of the most
common hospital-acquired infections and the most common cause of
antibiotic-associated diarrhoea. CDIs lead to great loss of life,
severe health outcomes, and incur very high financial costs
through treatment, extended hospital stays, and readmissions.
Despite extensive research and many resources committed to the
prevention and treatment CDIs in hospitalised patients, hospitals
continue to be hotspots for this disease. Meanwhile, there is an
emerging awareness of the burden this disease places on the
broader community including patients who have not recently been
hospitalised. In the community approximately 5% of adults and a
higher proportion of infants are asymptomatically colonised.
Colonisation is also common in livestock and the pathogen has
been isolated from meat and vegetables. However, the various
sources of transmission in the community and the consequences for
infections within and beyond hospitals are not well understood.
This thesis develops and employs mathematical models of C.
difficile transmission to explore three themes: improving models
to capture the complex epidemiology of C. difficile, populations
that sustain C. difficile transmission, and the classifi cation
of CDIs as hospital or community-acquired. Addressing the fi rst
theme, I argue that the essential epidemiology of C. difficile is
captured by modelling the interactions of three key factors:
pathogen, immunity, and gut flora. I argue that modelling
transmission in an integrated model of adults and infants across
hospitals and communities provides insights that hospital-only
and adult-only models cannot. By incorporating seasonality into
these models, I argue that seasonal variation of antibiotic
prescription rates is more likely to be the main driver of CDI
seasonality than seasonal transmission.
In the second theme, I argue that most hospitals -- though
hotspots for transmission -- are not disease sustaining
populations. Instead, transmission outside hospitals maintains
the disease in the hospital and community. I argue that reducing
transmission in the hospital cannot eliminate the disease in the
broader population, but that reducing transmission from adults or
infants in the community could interrupt transmission in the
human population. Similarly, I argue that C. difficile in the
community may be driven by transmission from animal reservoirs if
as few as 3.5-26.0% of human infections are acquired from animal
or food sources. In the final theme, I argue that an illusion of hospital-driven
disease is in part perpetuated by surveillance defi nitions that
systematically misclassify many community-acquired cases as
hospital-acquired. The incubation period for C. difficile
infections often exceeds the two-day or three-day cut-offs
commonly used to classify patients recently admitted to hospital.
I argue that many patients who acquire the pathogen prior to
admission develop symptoms after the cut-off and are therefore
incorrectly classifi
ed as having acquired the infection during
their hospital stay. Furthermore, I argue that time since
hospital discharge is a poor indicator of whether a CDI is
hospital or community-acquired.
Description
Keywords
mathematical models, Clostridium difficile, hospital-acquired infections, zoonoses, asymptomatic carriage
Citation
Collections
Source
Type
Thesis (PhD)
Book Title
Entity type
Access Statement
License Rights
Restricted until
Downloads
File
Description