Final Activity Report Summary - EPIMODEL (Applied stochastic modelling in veterinary epidemiology)
A new model for the airborne spread of the virus of foot-and-mouth disease was developed, based on the so-called Lagrangian particle model. The model improved the former Gaussian model, since it was additionally applicable on hilly regions. By using the model, it was possible to identify risk areas for different animal species such as sheep, cattle and swine around an infected farm. The model took into account for the calculations weather data, atmospheric conditions and topography of landscape. The model could also be applied to other diseases with just some parameters, e.g. virus emission of infected and sensitivity of susceptible animals, having to be accordingly updated.
A new stochastic model for vector-borne infectious diseases was developed on the basis of a former deterministic process model. Data from the recent Usutu virus outbreak in Austria were used to develop the process model. The Usutu virus was an arbovirus transmitted by mosquitoes which caused disease in blackbirds. The virus was first detected in Austria in 2001 and a major outbreak occurred in 2003. The main advantage of the new model was that it enabled statistical inference, including parameter estimation and assessment of uncertainty of predictions, and it allowed for enhanced sensitivity analyses. We applied the hierarchical Bayesian approach to embed the nine compartment SEIR model into a stochastic environment. Statistical analysis of the new model was made by Markov chain Monte Carlo, which was a computer intensive simulation method. We analysed the Austrian Usutu virus data that were obtained through monitoring performed between 2001 and 2005. We explored the structure of interrelationships among parameters and then calculated Bayesian estimates for model parameters and predictions. The same model would be applied to explain the dynamics of the West Nile virus in the United States of America and to predict recent bluetonge outbreaks in Central Europe.
A new stochastic model for vector-borne infectious diseases was developed on the basis of a former deterministic process model. Data from the recent Usutu virus outbreak in Austria were used to develop the process model. The Usutu virus was an arbovirus transmitted by mosquitoes which caused disease in blackbirds. The virus was first detected in Austria in 2001 and a major outbreak occurred in 2003. The main advantage of the new model was that it enabled statistical inference, including parameter estimation and assessment of uncertainty of predictions, and it allowed for enhanced sensitivity analyses. We applied the hierarchical Bayesian approach to embed the nine compartment SEIR model into a stochastic environment. Statistical analysis of the new model was made by Markov chain Monte Carlo, which was a computer intensive simulation method. We analysed the Austrian Usutu virus data that were obtained through monitoring performed between 2001 and 2005. We explored the structure of interrelationships among parameters and then calculated Bayesian estimates for model parameters and predictions. The same model would be applied to explain the dynamics of the West Nile virus in the United States of America and to predict recent bluetonge outbreaks in Central Europe.