Observational study reveals early milking and behavioral shifts in dairy cows, suggesting routine farm sensor data can detect lameness before visible symptoms appear.
Key Points
To identify deviations in milking characteristics and behavior before and after clinical lameness detection to evaluate their potential as early-warning indicators.
Monitored daily milking and pedometer data from 67 Holstein cows (parity 2.62 ± 1.57) with farmer-detected lameness across a 60-day window (±30 days around detection) on a dairy farm in Northern Italy.
Tracked milk yield, milking duration, flow dynamics, lying time, lying bouts, and step counts.
Modeled longitudinal trends using generalized additive mixed models (GAMMs) adjusting for parity and lactation stage.
Daily milk yield declined 12 to 16 days prior to lameness detection, with losses reaching up to 16% in multiparous cows alongside reduced milk duration and slower flow rates.
Lying time increased by 8% to 11% before detection, while activity patterns diverged by parity (decreased steps in primiparous cows and increased steps in multiparous cows).
Milking and behavioral parameters reverted toward pre-lameness baselines following therapeutic hoof trimming.