Mathematical modeling and social network analysis applications in foot-and-mouth disease transmission and livestock movements in U.S. production types

Abstract

The U.S. has been FMD-free since 1929. The U.S. has a large beef industry with over 45% of cattle on-feed concentrated in feedlots with a one-time head capacity ≥32,000 cattle. The country has a complex production system in which there is a continuous flow of livestock. An incursion of FMD could be devastating, so an understanding of the dynamics of a hypothetical outbreak in these large operations is needed. Also, an understanding of movement patterns to identify areas at-risk that can be targeted during disease response is needed. Mathematical modeling is the only tool available to study epidemics of infectious diseases such as FMD while Social Network Analysis (SNA) is an approach that helps to understand movement patterns. Parameterization of mathematical models is challenging due to the variability of the FMDv and the lack of specific data to U.S. beef populations. We developed an FMD expert survey to collect key parameter values of FMD natural history and transmissibility in beef U.S. feedlots. Data synthetized, used in combination with experimental and outbreak investigation data, will help to parameterize FMD-transmission models to evaluate implications of epidemics in U.S. beef feedlots. We developed a meta-population model to study FMDv transmission and evaluate interventions strategies within U.S. beef feedlots. We found that the projected outbreak duration was shorter for those feedlots with over 12,000 cattle population that operated with one hospital-pen compared to those feedlots that operated with two hospital pens. Restriction of movements of cattle from home pens to hospital pens within the feedlots was found to prolong the projected outbreak duration but did not interrupt FMDv transmission in feedlots modeled. Partial depopulation interventions were not found to be highly efficient in controlling FMDv transmission or required depopulation of a large proportion of cattle in feedlots modeled. We used social network analysis to describe inter-state movements of beef cattle, dairy cattle, swine, and small ruminants, and identify trade-communities within the contiguous U.S. for each livestock type network. We found that outputs generated resemble the nature of the beef feedlot industry (cow-calf to feedlot) while areas with large animal counts in the swine and dairy cattle networks were found to have high degree centrality. We also found between 1 to 2 largest communities in the beef cattle, dairy cattle, and swine networks which accounted for up to 65% of arcs in each network. The outputs of these networks could be useful to parameterize network models to assess disease transmission such as FMD at a national scale and evaluate the application of intervention strategies.