| dc.description.abstract | The dengue virus (DENV) remains a considerable global public health concern. The interactions between the virus, its mosquito vectors and the human host are complex and only partially understood. Dependencies of vector ecology on environmental attributes, such as temperature and rainfall, together with host population density, introduce strong spatiotemporal heterogeneities, resulting in irregular epidemic outbreaks and asynchronous oscillations in serotype prevalence. Human movements across different spatial scales have also been implicated as important drivers of dengue epidemiology across space and time, and further create the conditions for the geographic expansion of dengue into new habitats. Previously proposed transmission models often relied on strong, unrealistic assumptions regarding key epidemiological and ecological interactions to elucidate the effects of these spatio-temporal heterogeneities on the emergence, spread and persistence of dengue. Furthermore, the computational limitations of individual based models have hindered the development of more detailed descriptions of the influence of vector ecology, environment and human mobility on dengue epidemiology. In order to address these shortcomings, the main aim of this thesis was to rigorously quantify the effects of ecological drivers on dengue epidemiology within a robust and computational efficient framework. The individual based model presented included an explicit spatial structure, vector and human movement, spatio-temporal heterogeneity in population densities, and climate effects. The flexibility of the framework allowed robust assessment of the implications of classical modelling assumptions on the basic reproduction number, R₀, demonstrating that traditional approaches grossly inflate R₀ estimates. The model's more realistic meta-population formulation was then exploited to elucidate the effects of ecological heterogeneities on dengue incidence which showed that sufficient levels of community connectivity are required for the spread and persistence of dengue virus. By fitting the individual based model to empirical data, the influence of climate and on dengue was quantified, revealing the strong benefits that cross-sectional serological data could bring to more precisely inferring ecological drivers of arboviral epidemiology. Overall, the findings presented here demonstrate the wide epidemiological landscape which ecological drivers induce, forewarning against the strong implications of generalising interpretations from one particular setting across wider spatial contexts. These findings will prove invaluable for the assessment of vector-borne control strategies, such as mosquito elimination or vaccination deployment programs. | en_GB |
