Comparison of oestrus detection methods in dairy cattle

Sixty-seven Holstein-Friesian cows, from 20 days postpartum, were recruited into the study and fitted with both a pedometer (SAE Afikim) and a Heatime neck collar (SCR Engineers) and allocated a heat mount detector (either scratchcard [Dairymac] or KaMaR [KaMaR]) or left with none, relying only on farm staff observation. Common production stressors and other factors were assessed to determine their impact on the ability of each method to accurately detect oestrus and to investigate effects on the frequency of false-positive detections. Only 74 per cent of all potential oestrus periods (episodes of low progesterone) were identified by combining information from all methods. There was no difference between the methods in terms of sensitivity for detecting 'true oestrus events' (approximately 60 per cent), with the exception of scratchcards, which were less efficient (36 per cent). Pedometers and KaMaRs had higher numbers of false-positive identifications. No production stressors had any consequence on false-positives. The positive predictive values for neck collars or observation by farm staff were higher than those of other methods, and combining these two methods yielded the best results. Neck collars did not detect any of the nine oestrus events occurring in three cows with a body condition score (BCS) of less than 2, and the efficiency of correctly identifying oestrus was also reduced by high milk yield (odds ratio [OR]=0.34). Pedometer efficiency was reduced by lameness, low BCS or high milk yield (OR=0.42, 0.15 or 0.30, respectively).