A CASE-CONTROL STUDY TO IDENTIFY FARM FACTORS AFFECTING FERTILITY OF DAIRY HERDS - MULTIVARIATE DESCRIPTION OF FACTORS

Objectives To determine whether a multivariate model could be develope d to describe farm factors affecting fertility of dairy herd and to ex amine the factors that determined better reproductive performance on d airy farms in New South Wales. Procedure We examined the results of a survey of breeding practices, mating policies, recording systems, timi ng of mating, heat detection, treatment of metritis and manager demogr aphics of dairy farms using Principal components analysis, logistic re gression and the sign test. Results Eight principal components for hea t detection computed using Principal component analyses were included in the multivariate model to reflect primary and secondary heal detect ion signs used in the breeding decisions of managers. A final model id entified three variables: optimum number of days to first service, num ber of people detecting heats, and age of manager which significantly influenced the risk of being in the high or low reproductive performan ce groups. A failure to breed at an appropriate time after calving and having more people responsible for heat detection was associated with herds with a prolonged intercalving interval. While the model provide d a good fit for the data, it was not highly discriminatory. Examinati on of 83 management practices affecting fertility using the sign test (P = 0.01) indicated that managers of herds with good reproductive per formance employed better management and breeding practices more often than managers of herds with poor reproductive performance. Conclusions Mating cows sooner after calving and using fewer people to detect oes trus should reduce calving to conception intervals. Veterinarians inve stigating herd reproductive performance should consider the potential for a number of subtle failures of management to significantly impair fertility rather than limit investigations to a few factors.