D4Dairy-project – How digitalization and data integration pave the way to dairy health improvement

Abstract

Lameness in dairy cattle still remains a widespread animal welfare and economic issue. Early detection and prediction of lameness events may help to reduce negative effects for farms. However, integrating lameness assessment into the daily work routine is not feasible for many farms and may be prone to subjective bias for use in genetics. External assessment, carried out by veterinarians or claw trimmers, is often not done regularly and thus lameness may not be detected before it becomes clinical. Lameness is caused by a variety of factors including housing, feeding and management and may be associated with changes in milk performance or behavior. Technical advances and growing digitalization in various areas of dairy farming increased the amount and quality of data availability on farm supporting data-driven decision making. These advances happened incrementally with the consequence that data from different management areas is available in different formats or software programs and thus mainly used to make informed decisions in the respective area. Integrating these different data sources may enable more informed decisions by considering information from other management areas, which would be missed otherwise.

Thus, one aim of the D4Dairy project is the integration of various data to enhance early lameness detection and decision-support on individual farms. Our approach is based on the idea that lame cows show changes in performance and behavior, even before they show clinical signs of lameness. By integrating already existing data from national performance recordings, farm records, veterinary records, claw trimmings and different kinds of milking and sensor systems measuring e.g. activity, temperature or rumination (smaXtec, SCR by Allflex) a decision-support tool for early lameness detection and prediction should be developed. Subsequently, the outcomes of this approach should be used for the definition of auxiliary traits for claw health to be included into the breeding value estimation. Aside from the integration of different farm data this also requires the integration of data across farms, considering different systems, such as husbandry (e.g. free stall vs. pasture), milking (e.g. automatic milking system vs. conventional milking parlor with fixed milking times) or sensor systems (e.g. activity and temperature vs. activity and rumination monitoring).

In our contribution we want to show the steps towards data integration, which challenges we encountered and how we handled them. Furthermore, we present how we aim to use the outcomes for including claw health into the current breeding program and breeding value estimation.