Foot and Mouth Disease: Modelling Exercise
What was the Exercise about?
The Exercise was based on a simulated Foot and Mouth Disease outbreak in Great Britain. Participants were able to test modelling systems and practise the processes of presenting and discussing model predictions to assist decisions. During the FMD Modelling Exercise (15th – 22nd September), FMD models from three groups were used: Imperial College, Warwick University and Risk Solutions. The data for the FMD outbreak scenario was generated by Massey University’s InterSpread Plus FMD Model.
What was the purpose of the Exercise?
The Exercise was to test FMD models and strengthen communication links with external FMD modellers. Disease control policy was not being tested or reviewed as part of the Exercise. The outcomes and decisions made during the Exercise were based on the scenario used and were for Exercise purposes only.
What are the outcomes from the Exercise?
The evaluation of the Exercise is still underway and will not be complete for some weeks.
What is a model?
A model is a representation of reality, developed to use computers to predict possible outcomes of various scenarios. It can also be helpful to communicate these predictions to “non-experts” in a visual way. However, their predictions result from the knowledge and assumptions of their developers and the information they are supplied with. Hence, their limitations must be fully understood to set their predictions in context.
Why were three models being used?
No model is 100% accurate at predicting possible outcomes and different models have different strengths. By using several, it was possible to compare their results and use them in conjunction with each other.
What information do the models need to be given to work?
All the models require locations for individual farms, with basic farm information, especially the numbers of pigs, sheep and cattle. They also need information about animal movements, the state of a disease outbreak (such as the dates on which farms became infected) and information about possible control measures. Thus all the models require substantial inputs.
How do the models differ?
Risk Solutions’ model allows the most detailed analysis of resource limits and the costs of control, being the only model that identifies specific resource units, such as staff and vaccine doses.
The Imperial College and Warwick models focus more on analysing the progress of an epidemic, estimating important properties of the spread of FMD. They identify similarities and differences between current and past epidemics and can analyse changes over time and differences between places.
When should the models be used?
Between epidemics, the models can be used to identify the main issues for controlling future epidemics. The Imperial College and Warwick University models are particularly suitable for analysing past epidemics and investigating regional vulnerabilities, given assumptions about how disease spreads and how it will be controlled. Other models may be more suitable for comparing alternative control strategies and Risk Solutions’ model is especially suitable for cost benefit analyses. The Massey University and Risk Solutions’ models can also be used for training and developing training scenarios.
During an epidemic models can be used to guide action and monitor the effectiveness of control measures. The strength of Risk Solutions’ model may be in costing alternative options, whereas the other models may be valuable for detecting epidemiological differences from previous epidemics and deciding what the greatest risks are.
How early in an outbreak can the models be used?
All the models can be used from very early in an outbreak. However, the interpretation of the early results and interpretation of the epidemiological picture will be based on the results of previous scenario modelling and the detailed analysis of animal movement data. Initially they would have to use parameters set based on previous information and knowledge. A number of cases would be required before any parameters could be reset based on observations of the current outbreak. Risk Solutions’ model depends on the other models for resetting parameters. It is estimated that about 50 cases are required to start estimating values for the rate of disease spread, while 100-200 might be required to start changing the models of infection spread.
During an outbreak, resources are often an issue. Can models represent resources?
Risk Solutions’ model includes explicit, costed resource units, such as manpower and vaccination doses. It also includes estimation of the resource required to check non-infected reports and direct contacts, as well as other issues such as contamination by contact with infected animals or equipment.
The Warwick model simply restricts culling and vaccination through limits on the numbers of man-hours available nationally each day, as well as limits on the total supply of vaccine.
The Imperial models do not model resources. They can be restricted by either imposing time delays on culling and vaccination or setting daily limits.
Can anyone use the models?
The models differ in how far they have been developed for third party
use. All of the models require trained, expert users, but the models at
Imperial College and Warwick are solely intended to be research tools
for internal use, whereas Risk Solutions’ model has a highly developed
graphical user interface.
Page last modified:
14 January, 2008