Manipulation of contact network structure and the impact on foot-and-mouth disease transmission

Abstract

The movements of livestock between premises and markets can be characterised as a dynamic network where the structure of the network itself can critically impact the transmission dynamics of many infectious diseases. As evidenced by the 2001 foot-and-mouth disease (FMD) epidemic in the UK, this can involve transmission over large geographical distances and can result in major economic loss. One consequence of the FMD epidemic was the introduction of mandatory livestock movement restrictions: a 13-day standstill in Scotland for cattle and sheep after moving livestock onto a farm (allowing many exemptions) and a 6-day standstill for cattle and sheep in England and Wales (with minor exemptions, e.g. direct movements to slaughter). Such standstills are known to be effective but commercial considerations result in pressures to relax them. When contemplating legislative changes such as a change in length of movement restrictions we need to consider the consequent effect these could have on the emergent properties of the system, i.e. the network structure itself. In this study, we investigate how disease dynamics change when the local contact structure of the recorded livestock movement network in Scotland is altered through rewiring movements between premises. The network rewiring used here changes the structure of the recorded trade network through a combination of altered movement restrictions and redirection of movements between holdings and markets to avoid nonsensical activity (e.g. movements to markets on days when they are inactive) while conserving other characteristics (e.g. movement date as closely as possible and market sales of the correct animal production type). Rewiring results in networks with higher clustering coefficients and lower network density. The impact of rewiring on a hypothetical foot-and-mouth disease outbreak in Scotland was assessed by stochastic simulation, considering scenarios with and without exemptions to the standstill rules. As expected, rewiring leads to a decrease in outbreak size and - if standstill exemptions are prohibited - higher probability of smaller outbreaks. Without exemptions, a shorter movement standstill is almost as effective as a longer standstill period, indicating that a simpler biosecurity system would offer minimal additional risk for FMD. These results suggest that explicitly manipulating the contact network structure in a sensible way has the potential to significantly impact disease control.

The Use of Sheep Movement Data to Inform Design and Interpretation of Slaughterhouse-Based Surveillance Activities

“…The spatial distribution will also be different from June. While the Census data are the best denominator for the pig sector (5,11) and cattle movement data are now used to provide the numbers of cattle for the June Census, for sheep the SGI is more usually used (13,30). In this study, the gross analysis of within-country non-slaughter movements demonstrated this spatially dynamic nature of the Scottish sheep population and provided context for discussion with regard to the fifth challenge (direct move only).…”

“…For the most part, published literature of analyses of animal movement data has focused on network analysis and risk-based surveillance, predominantly in the cattle and pig sectors [e.g., (6)(7)(8)(9)(10)(11)]. Sheep sector movements have been included where they contribute to the spread of specified disease(s) under investigation (12,13) or where a larger scale, multi-sector, perspective is taken (14). There are a few publications of sheep population demographics (15,16) and network analyses (17,18).…”

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

“…Consequently, during this period, we observed an increase in the number of nodes and 386 connections; therefore, a given node would require more steps to reach another farm, within the 387 network which could alone reduce the speed of disease transmission (Vidondo and Voelkl, 388 2018 However, these followed the assumption that movement aggregation would not overestimate or 397 introduce edge biases (Vidondo and Voelkl, 2018). It has been suggested that to some extent, 398 static networks provide useful information about disease related risks (Mohr et al, 2018). pig transport data demonstrated that only a small fraction of 0.34% of the farms was connected 421 within one month (Sterchi et al, 2019).…”

Bayesian optimisation of restriction zones for bluetongue control

“…Thulke et al 20 also found that movement restrictions were more impactful in their model than pre-emptive culling in limiting the spread of classical swine fever in pigs. Mohr et al 21 recently looked at the impact of altering the contact network structure of livestock in Scotland on simulated foot-and-mouth disease dynamics. Optimal control measures have been studied extensively for plant diseases as well, even more so than for animal diseases.…”