043 Evaluation of Statistical Process Control Procedures to Monitor Feeding Behavior Patterns and Detect Onset of Bovine Respiratory Disease in Growing Bulls

Abstract

The objective of this study was to evaluate the effectiveness and accuracy of monitoring feeding behavior patterns using cumulative summation (CUSUM) procedures to predict the onset of bovine respiratory disease (BRD). Growing bulls (N = 231) consigned from independent producers for the purpose of evaluating growth efficiency during a 70-d trial were used in this study. Within a 10-d period, 30 bulls were treated for BRD based on observed clinical symptoms and elevated rectal temperature (> 39.5 C); all remaining bulls (n = 201) were deemed to be healthy. The clinically ill and healthy bulls were used to evaluate the sensitivity and specificity of CUSUM models, respectively. Frequency and duration of bunk visit (BV) events, head down (HD) duration and DMI were continuously monitored during the trial. All data were standardized prior to generating CUSUM charts in a daily accumulative manner. CUSUM charts were constructed with PROC CUSUM (SAS 9.4). Accuracies (average of sensitivity and specificity) of these single-trait CUSUM models for detection of BRD were 81.6, 80.3, 82.8 and 80.2%, respectively, and average day of detection prior (P < 0.05) to observed symptoms of BRD were 4.4, 3.5, 4.5 and 2.6 d, respectively. In addition, factors derived from a principal component analysis (PCA) of all 4 traits were monitored using similar CUSUM procedures. This 4-trait PCA-based model was more accurate (84.7 %) than any of the 4 single traits and signaled (P < 0.05) 3.7 d prior to observed symptoms of BRD. When DMI was removed from the full PCA model, accuracy of prediction (84.3% vs. 84.7%) and signal day prior to observed symptoms of BRD (3.3 vs. 3.7 d) were minimally altered, demonstrating the validity of monitoring feeding behavior traits for early BRD detection. These results demonstrate that the use of PCA-derived factors in CUSUM charts was more accurate compared to single-trait CUSUM charts. Moreover, due to the multivariate aspect of PCA, the use of PCA-based CUSUM charts to monitor feeding behavior patterns should be more robust in applications for preclinical detection of BRD. Results from this study demonstrate the potential value of using CUSUM procedures to monitor electronic feeding-behavior systems to enable more accurate preclinical detection of BRD in feedlot cattle.