Controlling the Spread of Livestock Diseases with the Help of Stochastic Models

Bayesian tools for fitting stochastic epidemic models to complex data sets describing genetic information on pathogens and locations of infections were developed by extending the established approach of data-augmented Markov Chain Monte Carlo. These new methods were adopted by scientists engaged in controlling diseases of farm livestock who extended them by including farm-level covariates and contact information. A user-friendly computer package BORIS (Bayesian Outbreak Reconstruction Inference and Simulation) was developed and applied in the control of real-world epidemics. BORIS’s capacity to reconstruct sources (and timings) of infections and to accommodate unobserved infections has proved particularly valuable. Since July 2018, BORIS has improved understanding of the epidemic within the New Zealand Ministry for Primary Industries (MPI) eradication programme for Mycoplasma bovis, a bacterial disease affecting cattle, estimated to cost NZD 886,000,000. Specifically, BORIS has been used to infer times and sources for infections to build confidence in the approach to surveillance and control and potentially to identify gaps in surveillance. This eradication programme has been effective with only 5 premises out of more than 60,000 having active disease in November 2022. BORIS has also enhanced understanding of Foot and Mouth Disease (FMD) in Japan, demonstrating its high transmissibility from farms holding predominantly pigs.