Using national movement databases to help inform responses to swine disease outbreaks in Scotland: the impact of uncertainty around incursion time.

Thibaud Porphyre,Lisa A Boden,George J Gunn,Mark E J Woolhouse,Carla Correia-Gomes,Harriet K Auty

doi:10.1038/srep20258

Abstract

Modelling is an important component of contingency planning and control of disease outbreaks. Dynamic network models are considered more useful than static models because they capture important dynamic patterns of farm behaviour as evidenced through animal movements. This study evaluates the usefulness of a dynamic network model of swine fever to predict pre-detection spread via movements of pigs, when there may be considerable uncertainty surrounding the time of incursion of infection. It explores the utility and limitations of animal movement data to inform such models and as such, provides some insight into the impact of improving traceability through real-time animal movement reporting and the use of electronic animal movement databases. The study concludes that the type of premises and uncertainty of the time of disease incursion will affect model accuracy and highlights the need for improvements in these areas.

Highlights

The epidemics of bovine spongiform encephalopathy in Europe[1] and of foot-and-mouth disease in the UK2 showed the importance of using mathematical models of disease transmission in providing key information to design contingency planning for animal disease outbreaks
In order to achieve this objective, we have focused on diseases of pigs (e.g. swine fevers such as classical swine fever (CSF) or African swine fever (ASF) viruses) which have non-specific clinical signs as well as a high potential to be transmitted through animal movements[17,18]
We explored the usefulness and limitations of using pig movement data to inform models when attempting to respond to an infectious disease incursion

Summary

Introduction

The epidemics of bovine spongiform encephalopathy in Europe[1] and of foot-and-mouth disease in the UK2 showed the importance of using mathematical models of disease transmission in providing key information to design contingency planning for animal disease outbreaks. The emphasis on using dynamic network models for contingency planning, but not during an outbreak, may be due to an assumption that they are less useful for making predictions of disease spread or identifying high risk farms in scenarios in which disease incursion has already occurred[6,16] This assumption may be based on two prior beliefs: (i) that data quality may be compromised by time-lags in data recording; and (ii) that the date of infection, which is critical to appropriate data selection, may be difficult to ascertain with any certainty. This may affect model predictions (and the uncertainty around them) of the patterns of disease spread

Objectives

Methods

Results

Conclusion