Unraveling mastitis in dairy cows: from multi-dimensional severity to quarter-level milk
Mastitis is one of the most persistent and costly challenges in the dairy industry. It reduces milk yield, compromises animal welfare, and imposes a substantial economic burden on dairy farms. While clinical mastitis (CM) is commonly evaluated based on observable symptoms, the disease is considerably more complex than what is apparent at the surface level. Equally important – but far less understood – is the recovery of milk yield following infection, particularly at the level of individual udder quarters. This thesis aims to investigate mastitis from two complementary perspectives: the multidimensional severity of the disease and quarter-level milk yield recovery following mastitis.
The first part of this thesis aims to assess the severity of CM across multiple dimensions. Traditionally, CM severity is evaluated based on clinical signs observed at the milk, udder or systemic levels; however, these measures capture only a single dimension. To provide a more comprehensive assessment, we developed a four-dimensional framework for evaluating severity in the production, somatic cell count (SCC), clinical, and pathogen dimensions. By examining the relationships among these dimensions, we found that although they are positively correlated, each captures a distinct aspect of the disease. The study also revealed that farm-level factors have a significant influence on severity across dimensions. Together, these findings underscore the importance of adopting a multidimensional perspective when evaluating mastitis and lay the foundation for more accurate and informative severity scoring systems.
The second part of this thesis aims to explore how CM severity affects quarter-level milk yield recovery. The percentage recovery was defined to quantify milk yield recovery first. Next, the recovery dynamics were compared between inflamed and uninflamed quarters. Finally, we investigated how CM severity in the four dimensions affects quarter-level milk yield recovery. The results indicate that in less severe cases, inflamed and uninflamed quarters exhibited similar recovery trajectories, whereas in prolonged cases, inflamed quarters recovered slower and worst. Severity in different dimensions affected inflamed and uninflamed quarters differently. Specifically, higher severity in each dimension was associated with worse recovery in inflamed quarters. In contrast, for uninflamed quarters, higher severity was associated with worse recovery during the first week, followed by improved recovery over the subsequent three weeks. Moreover, severity played different roles in shaping short-term and long-term recovery for inflamed and uninflamed quarters.
To derive more general conclusions about milk yield recovery following mastitis, the third part of this thesis extended cases to presumptive mastitis cases (PMC), identified using a combination of quarter-level milk yield, SCC, and electrical conductivity. We also examined the effects of farm, parity, and lactation stage (LS) on quarter-level milk yield recovery. The results showed that PMC followed similar recovery patterns with CM cases. Additionally, recovery during the first three days was primarily influenced by cow-level variation, whereas farm, parity, and LS became influential from day four onward. Overall, these findings confirm the existence of consistent recovery patterns following mastitis and suggest that milk yield recovery is not determined only by whether a quarter was infected but is also by initial severity and farm- or cow-level factors.
Taken together, this research offers a more comprehensive understanding of how mastitis affects dairy cows and how they recover afterward. By evaluating mastitis severity across multiple dimensions and exploring its association with quarter-level recovery, this work provides valuable insights to support selective treatment plans and more effective mastitis management strategies. Ultimately, the findings contribute to improved animal welfare, better milk production, and a more sustainable and resilient dairy sector.
Yifan Song recently completed her PhD in Precision Livestock Farming at KU Leuven in Belgium, where her research focused on data-driven analysis of mastitis in dairy cows. She holds a Master’s degree in statistics from Shandong University and a bachelor’s degree in mathematics from Lanzhou University in China. After graduating, she worked as a data analyst in industry, which strengthened her interest in exploring broader applications of data and advanced analytics. Motivated to deepen her expertise, she moved to Belgium in 2021 to begin her doctoral studies at the University of Leuven, combining a strong quantitative background with applied research in animal health and precision livestock farming.
