Clinical mastitis detection by on-line measurements of milk yield, electrical conductivity and milking duration in commercial dairy farms

The increasing spread of on-line measurement tools for milk parameters increases the need to investigate the factors that could influence the accuracy of these measurements for mastitis detection in field conditions. This paper reports the assessment of the dynamics of milking yield, electrical conductivity and milking duration in dairy cows before clinical mastitis outcome, and the influence of some of the most common sources of variation that could influence diagnosis accuracy. Three herds were enrolled in this study, all clinical cases observed by the milker in a period of 3 months were sampled and analysed. Overall 165 clinical mastitis cases and 1249 control cases were considered. The statistical analysis showed that herd, age and their interaction with time have a significant influence on the variance of yield, conductivity and milking duration. Time of milking showed to have the highest level of influence. Morning milking showed to be the best time to assess changes in the different on-line measurements, even when the milking intervals were equal.