Please use this identifier to cite or link to this item:
|Title:||Epidemiological investigations into the 2007 outbreak of equine influenza in Australia|
|Authors:||Firestone, Simon Matthew|
|Publisher:||University of Sydney.|
Faculty of Veterinary Science
|Abstract:||Equine influenza is a highly contagious and widespread viral respiratory disease of horses and other equid species, characterised by fever and a harsh dry cough. Prior to August 2007, Australia was one of only three countries to have remained free of equine influenza. An incursion of equine influenza virus H3N8 in that month resulted in a four-month outbreak during which approximately 69,000 horses were infected on an estimated 9599 premises across two States. Most of the geographic spread occurred within the first 10 days and was associated with the movement of infected horses prior to the implementation of movement controls. The outbreak was contained through a series of interventions that ultimately led to the eradication of equine influenza from Australia. During and immediately after the outbreak, intensive epidemiological investigations, laboratory and retrospective analytical studies were conducted culminating in a series of detailed reports and publications, and the collation of a highly detailed outbreak dataset. Further research into the factors that contributed to the spread of the outbreak and the effectiveness of measures implemented to control and contain it was considered important. The aim of this thesis was therefore to investigate the factors that contributed to the spread of the 2007 equine influenza outbreak in Australia and to develop statistical methods and tools useful for informing the surveillance and control of future emergency animal disease events. A case-control study was conducted to investigate premises-level risk factors, specifically whether compliance with advised biosecurity measures prevented the spread of equine influenza onto horse premises. Horse owners and managers on 200 properties across highly affected areas of New South Wales were interviewed. The proximity of premises to the nearest infected premises was the factor most strongly associated with case status. Case premises were more likely than control premises to be within 5 km and beyond 10km of an infected premises. Having a footbath in place on the premises before any horses were infected was associated with a nearly four-fold reduction in odds of infection (odds ratio = 0.27; 95% confidence interval: iv 0.09, 0.83). This protective association may have reflected overall premises biosecurity standards related to the fomite transmission of equine influenza: there was high correlation amongst several, generally protective, variables representing personal ‗barrier hygiene‘ biosecurity measures (hand-washing, changing clothes and shoes, and having a footbath in place). The movement of infected horses and local disease diffusion were known to be important mechanisms of spread early in this outbreak. A network analysis was conducted to investigate the relative contribution of each mechanism. The relationship between infected and susceptible horse premises (contact through animal movements and spatial proximity) was described by constructing a mixed transmission network. During the first 10 days of the 2007 equine influenza outbreak in Australia, horses on 197 premises were infected. A new likelihood-based approach was developed and it was estimated that 28.3% of early disease spread (prior to the implementation of horse movement restrictions) was through the movement of infected horses (95% CI: 25.6, 31.0%). Most local spread was estimated to have occurred within 5 km of infected premises. Based on a direct estimate of the shape of the spatial transmission kernel, the incidence beyond 15 km was very low. The median distance that infected horses were moved was 123 km (range 4–579 km). In an extension of the network analysis, novel methods were developed to delineate spatial clusters of infected premises and describe the sequence of cluster formation and the widespread dispersal experienced during the first 30 days of the outbreak. Premises identified as infected by the movement of infected horses were found to be critical to the seeding of infection in spatial clusters. Combined analysis of spatial and contact network data demonstrated that early in this outbreak local spread emanated outwards from the small number of infected premises in the contact network, up to a distance of around 15 km. A purely spatial method of modelling epidemic spread (kriging) was imprecise in describing the pattern of spread during this early phase of the outbreak (explaining only 13% of the variation in estimated date of onset of v infected premises), because early dissemination was dominated by network-based spread. Prior to this thesis, there was an abundance of anecdotal information regarding the role of meteorological factors and other environmental determinants in the spread of the 2007 equine influenza outbreak in Australia. A survival analysis was therefore conducted to empirically estimate the association between meteorological variables (wind, air temperature, relative humidity and rainfall) and time-to-infection in the largest cluster of the outbreak, in northwest Sydney. The equine influenza outbreak dataset was structured to enable generalised Cox regression modelling of the association between time-varying covariates representing premiseslevel meteorological conditions. The cumulative incidence in the northwest Sydney cluster was estimated to be 53.0% (95% CI: 51.4, 54.7%). Local spatial spread of equine influenza was found to be associated with relative humidity, air temperature and wind velocity. Meteorological conditions 3–5 days prior were strongly associated with hazard of influenza infection. Strong winds (>30 km hour-1) from the direction of nearby infected premises were associated with influenza infection, as was low relative humidity (<60%). A nonlinear relationship was observed with air temperature: the lowest hazard was on days when maximum daily air temperature was between 20–25 °C. Drawing on the findings of the above studies, a spatially-explicit stochastic epidemic model of equine influenza transmission was developed to investigate the underlying disease process, estimate the effectiveness of several control measures applied during the 2007 outbreak and to provide a dynamic modelling framework for rapid assessment of future equine influenza outbreaks in Australia. A reversible jump Markov chain Monte Carlo algorithm was used to estimate Bayesian posterior distributions of key epidemiological parameters based on data from two highly affected regions. A large amount of regional heterogeneity was observed in the underlying epidemic process, the estimated rate of decay of transmission by distance from infected premises, the intra-premises transmission rate and the effect of premises area. Model outputs were highly cross-correlated both temporally and spatially with data observed during vi the 2007 outbreak, and with outputs of a previous model. Pseudo-validation of the model against data, not used in its development, demonstrated of how it may be applied to develop rapid assessments of future outbreaks affecting horse populations in comparable regions to those studied. The study results documented in this thesis have elucidated the key factors underlying the spread of the 2007 equine influenza outbreak in Australia, and presented new methods of describing such rapidly spreading epidemics. The movement of infected horses, meteorological variables (air temperature, humidity and wind speed), on-farm biosecurity measures and intrinsic features of horse premises (proximity to other infected premises, numbers of horses held and premises area) were all important variables that influenced the spread of infection onto horse premises. These insights allow development of better policy and control programs in the event of a future equine influenza virus incursion.|
|Description:||Doctor of Philosophy(PhD)|
|Rights and Permissions:||The author retains copyright of this thesis.|
|Type of Work:||PhD Doctorate|
|Appears in Collections:||Sydney Digital Theses (Open Access)|
This work is protected by Copyright. All rights reserved. Access to this work is provided for the purposes of personal research and study. Except where permitted under the Copyright Act 1968, this work must not be copied or communicated to others without the express permission of the copyright owner. Use the persistent URI in this record to enable others to access this work.
|Firestone_S_Thesis_2012.pdf||7.7 MB||Adobe PDF||View/Open|
Items in Sydney eScholarship Repository are protected by copyright, with all rights reserved, unless otherwise indicated.