Royal (Dick) School of Veterinary Studies / The Roslin Institute
The aim of this project is to design improved and preferably optimised travelling schedules for veterinarians testing farms for bovine Tuberculosis (bTB) in the low risk areas of England. This project requires a student with good computational and programming skills, and an appropriate background in a quantitative subject (e.g. mathematics, computing, physics or engineering) with an interest in data science and/or data-driven network analysis.
BTB is estimated to cost GB nearly £100m per annum with the average cost of a single farm outbreak estimated to be £34K (https://consult.defra.gov.uk/bovine-tb/bovine-tb-cattle-controls-post-movement-testing/supporting_documents/
). Most costs are associated with high risk areas (HRAs), however costs in low risk areas (LRAs) remain substantial. While cattle-to-cattle transmission is almost entirely responsible for new LRA incidents, the ongoing outbreak of bTB in cattle and badgers in East Cumbria (https://www.bbc.co.uk/news/uk-england-cumbria-40812602
) highlights the risk of LRAs becoming HRAs.
If a geographical area where cattle are reared goes too long without testing, this exacerbates the risk of local disease spread, including to badgers. Thus regular cattle herd testing, consistently applied across all testing areas, is useful for identifying when a disease outbreak has the potential to become established in wildlife. However, to do this requires greater travel distances between farms and therefore has the potential to both strain resources and incur greater costs. As the numbers of both cattle herds and veterinary practices in many LRAs are declining, these constraints are becoming more severe, in extremis resulting in a critical failure of the testing system.
The aim of this project is to develop epidemiologically based definitions of testing area units and, using these unit definitions, to identify methods to improve and preferably optimize veterinary testing patterns, in order to reduce the strain on logistics and the costs given the identified constraints for minimum consecutive visit times to those units.
In the project, the student will first collate and interpret data from UK Farm Vets on testing logistics and cost to estimate reductions in effort cost to Defra and the service providers. Then, based on existing models of M. bovis transmission in cattle and badgers [1, 2], the student will simulate the introduction and potential spread of bovine Tuberculosis in LRAs for areas that have previously had no persistent infection in wildlife, and use the simulation outputs to identify the required criteria for repeat testing.
The student would then identify improved (and preferably optimal) solutions based on those criteria that would minimise journey times and distances, thereby recommending an alternative system that would substantially alleviate the stresses on test delivery.
The student will learn fundamental topics in applications of data science, such as epidemiological modelling, network modelling and data management. The project will involve developing data driven methods for optimisation and network science in relation to real world problems. The project will be jointly supervised across the University of Edinburgh’s Roslin Institute and Vet School (Prof. Rowland Kao), the School of Informatics (Dr. Rik Sarkar) and the University of Stirling Dept of Mathematics and Computing Science (Dr. A. O’Hare).
All candidates should have or expect to have a minimum of an appropriate upper 2nd class degree. To qualify for full funding students must be UK or EU citizens who have been resident in the UK for 3 years prior to commencement.