Serum urea nitrogen and conception rate: the usefulness of test information

Total of 627 AI in 332 Holstein cows in nine herds were used to examine the relationship between serum urea N and conception rate. Cows were assigned randomly to one of three isocaloric diets varying in protein degradability and content. The AI occurred from 50 to 150 DIM. Mean serum urea N for each cow from 50 to 150 DIM was used to examine conception rate and serum urea N. The clinical interpretation of serum urea N on conception rate is evaluated using Bayes theorem from two approaches (dichotomization vs. continuous). Test information resulting from dichotomization of serum urea N into high and low categories (maximizing the average of test sensitivity and specificity) is compared with likelihood ratio approaches allowing a continuous measure. Likelihood ratio test indicates that conception rate decreases with serum urea N of > 14.9 mg/dl, but dichotomized test suggests that the decrease does not occur until serum urea N is > 20 mg/dl.