Main content

Time-varying Individual-level Infectious Disease Models

Show simple item record

dc.contributor.advisor Deardon, Rob
dc.contributor.author Zhang, Lin
dc.date.accessioned 2014-01-07T23:38:27Z
dc.date.available 2014-01-07T23:38:27Z
dc.date.copyright 2013-12
dc.date.created 2013-12-20
dc.date.issued 2014-01-07
dc.identifier.uri http://hdl.handle.net/10214/7778
dc.description.abstract Individual-level models (ILMs) of infectious disease spread are a system of statistical models which can be used to model infectious disease transmission through a population in discrete-time. These models allow researchers to incorporate risk factors at the individual level; thus they are suited for modeling epidemics spatially. Individuals, here, may refer to people, animals, or plants, or aggregated units such as animals on a farm or students in a school. ILMs are usually fitted to data within a Bayesian statistical framework using Markov chain Monte Carlo (MCMC) methods. Ideally, covariate data and the infection status of individuals over time would be used to obtain parameter estimates for the ILMs. However, owing to various practical reasons, there are often situations in which the collection of infectious disease data at the individual level is infeasible. Instead, infectious disease data is collected at a regional level (e.g. a level which actually consists of spatially aggregated sets of individual units), such as health units or census regions. Therefore, it is reasonable to assume that the infectivity of such aggregated units varies as the status of infectiousness (i.e. the number/proportion of infectious individuals) within the aggregated unit changes. In the thesis, ILMs are extended to allow for time-varying susceptibility, infectivity and contact functions. A series of time-varying infectivity ILMs (TVI-ILMs) are then developed for the problem of modeling disease spread at the regional level. A method of carrying out model comparison and assessment based on the use of probability scoring rules is also developed and explored. Finally, the TVI-ILMs are extended to allow for infectivity curves that are dependent on regional-level covariates. Models and methods are tested on a combination of simulated epidemic data, and data from the 2009 H1N1 influenza pandemic collected in Southern Ontario. en_US
dc.description.sponsorship the Natural Sciences and Engineering Research Council (NSERC) of Canada Discovery Grants Program, the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA)/University of Guelph Partnership and OMAFRA-HQP Program, the Canada Foundation for Innovation (CFI) grant, "Centre for Public Health and Zoonoses (CPHAZ)", the GEOIDE. en_US
dc.language.iso en en_US
dc.subject spatial-temporal epidemic en_US
dc.subject individual-level models en_US
dc.subject data aggregation en_US
dc.subject time-varying infectivity en_US
dc.subject Markov chain Monte Carlo en_US
dc.subject Bayesian inference en_US
dc.title Time-varying Individual-level Infectious Disease Models en_US
dc.type Book chapter en_US
dc.degree.programme Mathematics and Statistics en_US
dc.degree.name Doctor of Philosophy en_US
dc.degree.department Department of Mathematics and Statistics en_US
dc.rights.license All items in the Atrium are protected by copyright with all rights reserved unless otherwise indicated.


Files in this item

Files Size Format View Description
Zhang_Lin_201401_PhD.pdf 2.633Mb PDF View/Open Thesis

This item appears in the following Collection(s)

Show simple item record