Asma Bousselmi, Eden Poulin, Édith Charbonneau, Sébastien Buczinski, Éric R Paquet
This research communication describes the evaluation of automated milk feeder (AMF) data for early detection of bovine respiratory disease (BRD) in dairy calves under commercial conditions. BRD is a major health challenge in the dairy industry, contributing to increased morbidity and mortality. Early identification could reduce its impact and improve herd health management. Five commercial dairy farms in Quebec, Canada, representing diverse calf management practices, were enrolled. Daily AMF data - including milk consumption (mL/day), drinking speed (mL/min) and numbers of rewarded and unrewarded visits - were collected over 1 year. The dataset comprised 90,088 daily records from 1,995 Holstein calves. Calves first accessed AMF between 7 and 14 days of age, depending on farm practices. Disease records were obtained from producers, and calves were classified as ill at their first treatment for BRD. The proportion of calves treated for BRD ranged from 20% on Farm 2 to 24% on Farm 3. Healthy and BRD calves were paired by month and duration of AMF use (±2 days). Linear mixed models were fitted for each farm with health status, days relative to treatment, month, and their interactions as fixed effects, and calf as a random effect. Heterogeneity between farms was evident in the results, and multiple farms showed a significant difference between healthy calves and those with BRD for each of the studied parameters, but with different days associated with the significant result. The earliest significant differences were observed at 0 days for drinking speed, -5 days for the number of unrewarded visits, -3 days for both consumption and the number of unrewarded visits, -4 days for the number of unrewarded visits and -5 days for both consumption and drinking speed on Farms 1-5, respectively. AMFs show promise for early BRD detection, but detection timing and the usefulness of specific measurements vary across farms.