Andrew McCarren, Associate Professor, Dublin City University

The spread of Bovine Tuberculosis (bTB) among cattle herds has been a topic of considerable interest to Government, Farming and the science communities both nationally and internationally. In 2021, an estimated €105 million was spent on the TB eradication program in direct costs alone in Ireland; herd incidence in 2021 was 4.3%. In cattle, the primary bronchopneumonic infection may remain localized or progress slowly for considerable periods of time, eventually leading to generalized lesions. Therefore, tuberculin positive cattle are regarded as “open” cases of TB, potentially capable of transmitting infection to other animals and humans. The beef industry in Ireland supports in excess of 70,000 beef farmers and generates an ex-farm output value of €2.5bn. Therefore, the importance of managing the spread of TB in the cattle herd has both a societal and economic impact to the State.

The use of Machine Learning has only recently received attention in the fight against TB eradication and the techniques currently applied have not considered either the imbalanced nature of the outcome class or the temporal nature of the data. Our team has spent a considerable amount of time in the past examining imbalance, anomaly detection and temporal  prediction techniques in many studies in health, sports science, engineering and agriculture. Now we aim to apply these learnings in conjunction with the Department of Agriculture, Food and the Marine to examine their potential in predicting TB outbreaks at herd level.