Speaker
Description
In Spring 2025, a foot-and-mouth disease (FMD) outbreak occurred along the Hungary-Slovakia border region, affecting 11 farms before being contained. While the small number of cases is encouraging from a control perspective, it poses a major statistical challenge: standard methods for inferring transmission parameters and evaluating control measures are not suitable for such limited data. Nevertheless, timely and quantitative estimates are crucial for guiding contingency planning and resource allocation.
To address this challenge, we developed a simulation-based inference framework built on a farm-level branching-process model. Our approach is to generate a large ensemble of simulated outbreaks under a range of plausible parameters and retain those that closely match the observed epidemic. These “look-alike” epidemics allow us to construct posterior distributions for key epidemiological quantities. Applied to the Hungary-Slovakia data, the posterior concentrates the pre-control farm-to-farm effective reproduction number around 3-4 (median ≈ 3.6) and the post-control effective reproduction number below one (median ≈ 0.8), with a generation-interval median of about 12 days.
Our method is transparent, biologically grounded, and computationally efficient, making it suitable for real-time decision support during small outbreaks. It was used in real time to inform the food safety authority of Hungary responsible for animal health and agriculture production as well.
