Development of a syndromic surveillance system for Irish dairy cattle using milk recording data.

In the last decade and a half, emerging vector-borne diseases have become a substantial threat to cattle across Europe. To mitigate the impact of the emergence of new diseases, outbreaks must be detected early. However, the clinical signs associated with many diseases may be nonspecific. Furthermore, there is often a delay in the development of new diagnostic tests for novel pathogens which limits the ability to detect emerging disease in the initial stages. Syndromic Surveillance has been proposed as an additional surveillance method that could augment traditional methods by detecting aberrations in non-specific disease indicators. The aim of this study was to develop a syndromic surveillance system for Irish dairy herds based on routinely collected milk recording and meteorological data. We sought to determine whether the system would have detected the 2012 Schmallenberg virus (SBV) incursion into Ireland earlier than conventional surveillance methods. Using 7,743,138 milk recordings from 730,724 cows in 7037 herds between 2007 and 2012, linear mixed-effects models were developed to predict milk yield and alarms generated with temporally clustered deviations from predicted values. Additionally, hotspot spatial analyses were conducted at corresponding time points. Using a range of thresholds, our model generated alarms throughout September 2012, between 4 and 6 weeks prior to the first laboratory confirmation of SBV in Ireland. This system for monitoring milk yield represents both a potentially useful tool for early detection of disease, and a valuable foundation for developing similar tools using other metrics.