The Development of an Individual-Based Model to simulate the Progress of a Foot-and-Mouth Disease Epidemic

Abstract

Disease control strategies including quarantine or vaccination have been used to minimize damages incurred by the foot-and-mouth disease (FMD). Disease spread simulation models can be used for planning such strategies. Still, few models have been developed for detailed simulations of the FMD dispersal, e.g., within a farm and between farms, which would provide information for assessment of a given control strategy. The objectives of this study was to develop the FMD-IBM, an individual-based model (IBM) in order to simulate the intra- and inter-farm dispersal of FMD at a regional scale. The FMD-IBM model was based on the Reed-Frost equation which determines the probability of infection events between hosts as well as between farms. The parameters affecting the within-farm disease spread process were fitted using the calibration set data generated by the Davis Animal Disease Simulation (DADS) model as little observation data have been open to the public. A range of parameter values were evaluated to determine the set of parameters that had the highest degree of agreement for the occurrence dates of events between the FMD-IBM and the DADS model in a intra disease dispersal. Parameterization of regional-level disease spread was also performed using the epidemic case data provided by the World Organisation for Animal Health (OIE). Simulations of the disease dispersal events using the FMD-IBM were performed for two FMD outbreaks including the Andong epidemic in 2010 and Chungcheongnam-do epidemic in 2016, which represent highly severe and slight outbreak of the FMD. The FMD-IBM provided detailed information on disease dispersal including data specific to a farm, e.g., the date of infection. Such information was used to determine the speed of disease dispersal at each farm. It was found that the rate of the FMD spread ranged from 0.3 to 1.3 km per day at most farms. The parameter values that represent the infectability of a farm were lower in the 2016 epidemic than the 2010 epidemic. Those results suggested that the FMD-IBM could allow preparation of farm specific data for planning and decision-making on control measures against the FMD. Still, the FMD-IBM model had limited ability to predict long dispersal events, which may be improved by adding a new algorithm to take trajectory data into account.