-
More

EPIC/APHA meeting on Livestock Data and Demographics

Theo Pepler

Having accepted an invitation to attend the annual modelling symposium organised by the Animal and Plant Health Agency (APHA), in the first week of February I found myself on a train to their offices in Weybridge. The symposium is aimed towards increasing collaboration, communication and understanding amongst colleagues doing simulation and statistical modelling work in animal and plant health—both within and outwith APHA—and to aid in model development and maintenance.

 

The focus of this year’s EPIC APHA meeting was to exchange knowledge about how data on livestock are used for inference about livestock populations and demographics.

For the past couple of years, we have taken the opportunity of having EPIC / APHA colleagues gathered together to hold an additional meeting so that matters of interest to both groups can be discussed. This year the EPIC /APHA meeting preceded the modelling symposium and I will discuss it in a separate blog.

The focus of this year’s EPIC APHA meeting was to exchange knowledge about how data on livestock are used for inference about livestock populations and demographics. This also includes further analyses such as simulation modelling of infectious disease spread. The exchange of ideas allows us to ensure that the work of EPIC and APHA is complementary, and additionally provides an opportunity for individuals in each organisation to meet the relevant contacts in the other group.

The EPIC/APHA discussion meeting started off with presentations from some of the EPIC members in attendance. Harriet Auty gave a brief introduction to EPIC, mainly for the benefit of newer members in the APHA team. She discussed the structure of our research consortium, and our mission in providing evidence-based and policy relevant information on exotic animal disease outbreaks to Scottish Government.

Aaron Reeves talking about the importance of high quality, accurate data in the research work of EPIC. 

Aaron Reeves then talked about his role as a member of the EPIC Data Team, curating livestock demographic data from various sources and augmenting these data sets to improve their value in the work of other EPIC researchers. This crucial work has allowed EPIC to gain access to various types of data relevant to our mission, including data on farm demographics, livestock movements, diagnostic test results, environmental and wildlife risk factors, and disease surveillance. Aaron emphasised the importance of overseeing the appropriate and secure use of these data, and noted that data quality and accuracy are essential for EPIC’s research.

In her talk on the use of sheep movement data to inform design and interpretation of slaughterhouse based surveillance activities, Julie Stirling noted the importance of including cross-border moves—from and to the rest of the UK. If we look only at Scottish data, it is clear that we will have a significant blind spot when it comes to disease surveillance.

Sibylle Mohr and Kajetan Stanski then gave a combined presentation of their involvement in the UK-wide Foot and Mouth Disease exercise in May 2018. It was interesting to consider the lessons learnt by EPIC during this exercise, and to be reminded of the ways in which the exercise helped EPIC to improve our preparedness for real disease outbreaks.

 

 

Stephen Catterall demonstrating how partial data on animal movement histories impacts EPIC’s ability to produce accurate infectious disease risk maps.

Rounding off the presentations from EPIC’s side, Stephen Catterall talked about his research on the improvement of disease contact tracing using individual animal data. Noting that individual sheep IDs are needed for proper contact tracing, he showed how partial observation of the individual animal movement data introduces uncertainty about sheep movement histories. This happens when some sheep are not scanned when moved from one location to another, and it greatly impacts upon our ability to produce disease risk maps for Scotland.

After a sandwich lunch, there was opportunity for our APHA colleagues to present some of their work. Within APHA there are different Livestock Demographic Data Groups (LDDGs), each focused on a specific livestock species. Alessandro Foddai, epidemiologist in the cattle LDDG, started off by discussing different cattle population indicators gathered by APHA. These include indicators such as herd sizes, age at first calving, number of importing herds, number of animals imported (per importing herd), and on and off movements per herd. He pointed out knowledge of these indicators is important in answering many epidemiological questions. For example, if we want to demonstrate that the occurrence of a disease in the cattle population is lower than a specified level, we need to know how many cattle (and herds) there are in the population.

Katerina Chaintarli, veterinary epidemiologist in the sheep and goat LDDG, then talked about the different sheep population indicators. Changes in these indicators—flock size, number of APHA field visits, fallen stock, movements, imports, etc.—might point to changes in the industry, and changes in the exposure risk for disease, amongst other things.

...it is certainly a creative and pragmatic approach of addressing the question they were faced with.

In the final presentation from APHA’s side, Adam Brouwer talked about their work in estimating the number of poultry small holdings. By combining very limited poultry-specific data—counts of smallholdings in protection zones from avian flu outbreaks—with census and land cover data, they have developed a method to estimate the number of poultry smallholdings across other areas of Great Britain. Although Adam was very honest about the limitations of their method, it is certainly a creative and pragmatic approach of addressing the question they were faced with.

What struck me again in this meeting, in particular from the nature of the presentations, is how related and complementary the work of APHA and EPIC is. The LDDG members are focused on getting the relevant data, and getting it right. If you base any research work on incorrect denominator data, you will inevitably end up with the wrong answers. The EPIC presentations, on the other hand, reflected our focus on the study of infectious disease systems, making inference about how disease outbreaks will affect the Scottish livestock industry and how these effects can be mitigated through government policy. Almost all of EPIC’s work is dependent on high quality, accurate data, obtained from either APHA or other data providers.

Theo Pepler blog part II: APHA Modeling Symposium

Views expressed are those of the guest blogger and do not necessarily reflect those of the EPIC project or funders.

EPIC Blog Homepage

Website design by Innovation Digital Limited