Assessment of automated activity monitors for managing reproduction in dairy cows

Abstract

Automated activity monitoring (AAM) systems are among the tools available for managing reproduction on dairy cattle farms. The objective of this thesis was to explore the perception, use, and performance of different reproduction management practices, especially AAM systems. In a national survey, questions were asked on general and reproduction management, reproductive performance and opinions and perceptions of respondents toward reproduction on Canadian dairy farms. Farms had, on average, 77 lactating cows housed in tiestall (61%), freestall (37%), or bedded pack (2%) barns. The main reproduction management practice (means employed for > 50% of inseminations) used on farms were visual estrus detection (51%), timed artificial insemination (AI; 21%), AAM (10%), other management practice (2%), and a combination of management practices (16%). The average pregnancy rate (PR) was 17.6%, and did not differ among herds using different reproduction management practices. Improving performance was the main factor influencing decisions concerning reproduction for 80% of the respondents, and they perceived the performance of the tools and technologies they use (synchronization programs and AAM system) to be better than the performance of visual estrus detection. In a randomized controlled trial, cows were assigned either to be inseminated following a timed AI protocol (Double Ovsynch), or to be inseminated based on estrus detection by AAM. Overall, the odds of pregnancy at first insemination and by 88 DIM were not significantly different between the Double Ovsynch group and the AAM group, but there were significant interactions between treatment and herd for both pregnancy outcomes, indicating differences in the relative performance of Double Ovsynch and AAM between herds. Finally, a second randomized controlled trial assessed the association between the length of the luteal phase (normal or short), and the duration and intensity of estrus as detected by AAM. There was no significant effect of treatment when models were adjusted for confounders. To conclude, there is a large variability in performance among herds using AAM systems for managing reproduction on dairy farms and other herd characteristics and management strategies are likely causing this inconsistency.