Cost benefit analysis of vaccination in biosecurity : spatial network approach

Assessing the economic impacts of large-scale livestock vaccination programs is necessary for decision-makers to enhance the efficiency and sustainability of these programs in developing countries. However, it is not easy to achieve accurate assessments in practice. While computing vaccination cost is rather simple, measuring its benefit is not so straightforward. The main reason for this challenge is that researchers cannot ascertain how livestock epidemics may spread if no vaccination program is in place. Given the fact that the cost benefit analysis (CBA) approaches, which have been used in the current literature, have trouble predicting such information, this thesis introduces a new CBA approach: the spatial network model-based CBA. Compared to non-spatial CBA approaches, it can incorporate the stochasticity of the epidemic spread process, spatial and dynamic features of livestock epidemics when estimating the outbreak size in the baseline scenario. As a result, this approach may distinguish the impacts of vaccination from those of other factors and better predict the outbreak size. Additionally, by using this approach researchers can tailor their spread analysis level to suit the real-life data paucity condition in developing countries. While the novel approach has a number of advantages when being used to evaluate the economic benefit of livestock vaccination in developing countries, the spatial network model used to predict baseline outbreak size has a complex parameterisation process. To make the approach feasible, the thesis discusses some important practical issues that should be considered when performing this model in practice. It discusses two potential estimation biases due to isolation (the failure of filtering non-susceptible premises in the sample) and unobserved sources (the failure of capturing unobserved infection sources). Additionally, some technical hints, namely selecting functional form for the transmission kernel and data scaling for improving computation efficiency, are provided. Two case studies are presented to exemplify how the new CBA approach can be applied to analyse the benefits of vaccination in developing countries. Using actual outbreak data of H5N1 Highly Pathogenic Avian Influenza (HPAI) and Foot and Mouth Disease (FMD), geographical locations, livestock production details, and climatic conditions of communes, this thesis examines the costs and benefits of large-scale vaccination programs in Vietnam. It concludes that both the FMD vaccination program in the South Central and the H5N1 HPAI vaccination program in the Mekong River Delta are cost-effective. The estimations also confirm the biasedness caused by isolation and unobserved sources. Some important findings relating to the epidemiology of these two diseases such as the impacts of sunshine time on the FMD spread, are also found. Overall, this thesis offers a new CBA approach that potentially provides reliable economic analyses of livestock control measures in developing countries. We believe that applying this approach may provide policymakers with better assessments when considering solutions to improve the efficacy of vaccination programs. In the future, the spatial network model can be integrated with advanced economic analyses such as the optimisation model or the computable general equilibrium (CGE) model to be an important decision-supporting tool for policymakers in developing countries