Comparison of a camera - software system and typical farm management for detecting oestrus in dairy cattle at pasture

To compare the sensitivity, specificity, predictive values and accuracy of detection of oestrus using a novel oestrus detection strip (ODS) and a camera-software device (CSD) with typical farm management practices of visual observation and use of tail paint in dairy cattle at pasture.

Dairy cows (n = 480) in a seasonal-calving herd managed at pasture under typical commercial conditions in New Zealand were stratified by age, body condition score and days in milk, then randomly allocated to one of two groups prior to the planned start of mating (PSM). Tail paint was applied to all cows and oestrus detected by visual observation of oestrous behaviour and removal of paint, by farm staff. One group (n = 240) was fitted with ODS and also monitored for signs of oestrus using a CSD, while the Control group (n = 240) was monitored using tail paint and visual observations only. Cows detected in oestrus were artificially inseminated (AI), and pregnancy status determined using rectal palpation and ultrasonography, 51-52 days after the end of a 55-day A period. Results of pregnancy diagnosis were used to confirm the occurrence of oestrus, and the sensitivity, specificity, predictive value and accuracy of detection of oestrus compared between oestrus detection methods.

The sensitivity and accuracy of oestrus detection in the Control group, using visual observation and tail paint, were low. Compared with the Control group, detection of oestrus using the ODS and CSD resulted in greater sensitivity (85% vs 78%; p = 0.006), specificity (99.6% vs 98.0%; p < 0.001), positive predictive value (PPV; 88% vs 51%; p < 0.001) and overall accuracy (99.0% vs 98.0%; p < 0.001). Negative predictive value (NPV) did not differ significantly between groups (99.4% vs 99.3%; p = 0.28). Pregnancy rate to first service was higher in the CSD group than in the Control group (72% vs 39%; p < 0.05). Use of the CSD significantly increased the cumulative proportion of cows pregnant to AI over the breeding period (p = 0.044).

The ODS and CSD was satisfactory for detection of oestrus in seasonal calving dairy herds grazing on pasture and could improve the sensitivity and accuracy of detection of oestrus in herds where these are low.