Methods to improve accuracy of preclinical detection of bovine respiratory disease (BRD) are needed to reduce the economic impact of this disease, improve animal welfare, and promote more judicious use of antimicrobials in beef cattle. The objectives of this study were to retrospectively characterize time-series deviations in DMI and feeding behavior patterns preceding the display of observed clinical symptoms associated with BRD and to identify those feeding behavior traits that would be most predictive of BRD. The study was conducted with 231 seed stock bulls (391 ± 55 kg initial BW) representing 5 breeds that were housed in a facility equipped with GrowSafe feed bunks at a commercial bull-test facility. All bulls were vaccinated against standard viral and bacterial pathogens before and on arrival at the facility. Daily DMI and feeding behavior traits (frequency and duration of bunk visit events, head-down duration, variance of nonfeeding intervals, and time to approach feed bunk following feed-truck delivery) were measured for 70 d with a GrowSafe system. During a 10-d period from Day 28 to 37 of the trial, 30 bulls were administered antimicrobial therapy for clinical symptoms of BRD (rectal temperatures > 39.5°C). All remaining bulls ( = 201) were administered metaphylactic therapy on Day 38 of the trial in response to an acute decrease in feed intake. A retrospective analysis was conducted using a 2-slope broken-line regression model to identify inflection points in DMI and feeding behavior traits relative to onset of illness. The bulls were separated into 2 cohort groups based on observed clinical illness ( = 30) or those metaphylactically treated ( = 201), with the 2-slope broken-line regression model applied separately to each cohort. The model-detected inflection points for DMI were 6.8 and 3.8 d before observed clinical illness and metaphylactic treatment, respectively, and the reductions in DMI from detected inflection points to the day of observed clinical illness and day of metaphylactic treatment were 39.3 and 49.8%, respectively. Furthermore, the model-detected inflection points for individual feeding behavior traits ranged from 1.3 to 14.2 d before observed clinical illness and from 3.8 to 12.6 d before metaphylactic treatment. Results from this study demonstrate the potential value of electronic behavior-monitoring systems to improve the sensitivity and specificity of preclinical detection of BRD in feedlot cattle.

the value of using CUSUM procedures to monitor feeding behavior patterns to more accurately detect BRD prior to clinical symptoms in feedlot cattle.

No abstract