Quantifying the dynamics of pig movements improves targeted disease surveillance and control plans

Abstract

SummaryTracking animal movements over time can fundamentally determine the success of disease control interventions throughout targeting farms that are tightly connected. In commercial pig production, animals are transported between farms based on growth stages, thus it generates time-varying contact networks that will influence the dynamics of disease spread. Still, risk-based surveillance strategies are mostly based on a static network. In this study, we reconstructed the static and temporal pig networks of one Brazilian state from 2017 to 2018, comprising 351,519 movements and 48 million transported pigs. The static networks failed to capture time-respecting movement pathways. Therefore, we propose a time-dependent network susceptible-infected (SI) model to simulate the temporal spread of an epidemic over the pig network globally through the temporal movement of animals among farms, and locally with a stochastic compartmental model in each farm, configured to calculate the minimum number of target farms needed to achieve effective disease control. In addition, we propagated disease on the pig temporal network to calculate the cumulative contacts as a proxy of epidemic sizes and evaluated the impact of network-based disease control strategies. The results show that targeting the first 1,000 farms ranked by degree would be sufficient and feasible to diminish disease spread considerably. Our finding also suggested that assuming a worst-case scenario in which every movement transmit disease, pursuing farms by degree would limit the transmission to up to 29 farms over the two years period, which is lower than the number of infected farms for random surveillance, with epidemic sizes of 2,593 farms. The top 1,000 farms could benefit from enhanced biosecurity plans and improved surveillance, which constitute important next steps in strategizing targeted disease control interventions. Overall, the proposed modeling framework provides a parsimonious solution for targeted disease surveillance when temporal movement data is available.