Integration of empirical network data and agent-based modelling to examine the risk of equine influenza infection in equine athletes in Ontario, Canada

Horses are frequently transported, creating opportunities for the spread of pathogens. Disease transmission models for equine infectious diseases face limitations on their generalizability due to challenges in describing equine movement and the structure of their contact networks beyond simplistic assumptions. This study aimed to combine a stochastic, agent-based, SEIR model for equine influenza disease dynamics with an observed Ontario, Canada equine contact network structure to quantify the potential magnitude of equine influenza outbreaks in Ontario competition horses under different conditions. Different interventions were modelled to help provide insight into the impacts of biosecurity practices to mitigate population risk. Eight scenarios with different levels of vaccination (42.5–95 %) and horse contact rates (normal distributions with means of 2 and 5) were simulated within the competition network. Outcomes of interest for each scenario included attack rate, number of infected home facilities, number of infected competitions, and outbreak duration. For each scenario, 200 stochastic iterations were performed. The results demonstrate that decreasing contact between horses was more effective at reducing key outcome indicators (attack rate, number of home facilities with infected horses, number of competitions with infected horses) compared to any change in vaccination coverage among the non-competitor horse population. This model integrating disease dynamics of equine influenza and a parameterization of an Ontario competition network outlines the importance of the role of contact-related behaviours when discussing biosecurity risk mitigation measures for populations of Ontario equine athletes.