AI to predict the incursion risk of exotic diseases

Risk assessment addressing the incursion risk of exotic animal diseases is usually based on mechanistic models, that describe the underlying mechanisms contributing to the probability of disease introduction for individual introduction routes. The risk of disease incursion is, however, not only understood from connections between infected and free territories, but also includes risk factors related to, e.g., the natural environment, the climate, and human behavior. This is especially true when considering zoonotic and vector-borne diseases.

The main objective of this project is to explore the application of artificial intelligence (AI) in assessing disease incursion risks. The project will indicate which machine learning technique is most suited for the assessment of disease incursion risks and which metrics are most appropriate to validate the results obtained.

The use of AI to predict disease incursion risks will be piloted using the global databases included in a mechanistic model, the Rapid Risk Assessment Tool (RRAT). These databases provide data on disease occurrence worldwide and international trade and travel for a 3-year period. The number of disease incursions in this 3-year period is limited, as incursions of exotic animal diseases are rare events. The output data needed to construct the AI model is therefore limited and a synthetic output dataset of risk scores for selected diseases and countries will be generated using the RRAT.

It is expected that the results of this project will contribute to further development and application of AI in risk assessment for exotic disease incursion. The use of AI will make it possible to account for multiple risk factors (drivers) contributing to the risk of exotic disease incursion, resulting in a more comprehensive estimate of the risk.