Predicting cumulative risk of bovine respiratory disease complex (BRDC) using feedlot arrival data and daily morbidity and mortality counts.
Academic Article
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
Although bovine respiratory disease complex (BRDC) is common in post-weaning cattle, BRDC prediction models are seldom analyzed. The objectives of this study were to assess the ability to predict cumulative cohort-level BRDC morbidity using on-arrival risk factors and to evaluate whether or not adding BRDC risk classification and daily BRDC morbidity and mortality data to the models enhanced their predictive ability. Retrospective cohort-level and individual animal health data were used to create mixed negative binomial regression (MNBR) models for predicting cumulative risk of BRDC morbidity. Logistic regression models were used to illustrate that the percentage of correctly (within |5%| of actual) classified cohorts increased across days, but the effect of day was modified by arrival weight, arrival month, and feedlot. Cattle arriving in April had the highest (77%) number of lots correctly classified at arrival and cattle arriving in December had the lowest (28%). Classification accuracy at arrival varied according to initial weight, ranging from 17% (< 182 kg) to 91% (> 409 kg). Predictive accuracy of the models improved from 64% at arrival to 74% at 8 days on feed (DOF) when risk code was known compared to 56% accuracy at arrival and 69% at 8 DOF when risk classification was not known. The results of this study demonstrate how the predictive ability of models can be improved by utilizing more refined data on the prior history of cohorts, thus making these models more useful to operators of commercial feedlots.
