Simulation modelling to support national policy making in the control of bovine herpesvirus 1

Bovine herpesvirus 1 (BHV1) is the causative agent of infectious bovine rhinotracheitis (IBR), a respiratory disease in cattle. Increased international legislation, together with a high prevalence of BHV1 infected cattle in The Netherlands, put pressure on Dutch livestock industry to eradicate BHV1. The main objective of this thesis was the development and application of simulation models to support policy makers in various phases of the decision-making process with respect to a national BHV1 eradication programme in The Netherlands. To meet this objective, three simulation models were developed. First, a state-transition model was developed to evaluate the epidemiological and economic consequences of various control strategies for endemic BHV1 in The Netherlands. Based on international developments and results of this model, a compulsory vaccination programme for BHV1 was implemented in The Netherlands in May 1998. According to model outcome, this programme was expected to reduce the prevalence of infected dairy cattle to 5% in about 5 years, with expected direct costs approximately EUR 100 million, a pay-back period of about 8 years and less than 1% outbreaks per year on certified BHV1-free herds. A second model, classified as spatial, dynamic and stochastic, was developed to evaluate control strategies for outbreaks in a BHV1 free country. Results showed that farm type with first introduction of BHV1 had a considerable impact on the number of secondarily infected farms and total costs. To support policy makers during the eradication programme, the epidemic model was adapted for an endemic situation, resulting in the third model called 'InterIBR-endemic'. This model closely interacted with a BHV1 monitoring programme. As part of internal validation, various experiments with this model were performed to improve understanding of model behaviour. To support sensitivity analysis, the techniques of experimental design and metamodelling were applied to help set priorities for further epidemiological research. The uncertainty of the yearly reactivation rate of latently infected animals affected the costs by EUR 43 million, which is about 40% of the expected total costs. Also, survival analysis was applied to quantify the association between farm characteristics and the risk of certificate loss during simulation. Early 1999, compulsory vaccination in The Netherlands was postponed due to contamination of live marker vaccine with bovine virus diarrhoea virus.