PLoS One. 2024 Apr 25;19(4):e0302172. doi: 10.1371/journal.pone.0302172. eCollection 2024.
ABSTRACT
Clinical illness (CI) scoring using visual observation is the most widely applied method of detecting respiratory disease in cattle but has limited effectiveness in practice. In contrast, body-mounted sensor technology effectively facilitates disease detection. To evaluate whether a combination of movement behavior and CI scoring is effective for disease detection, cattle were vaccinated to induce a temporary inflammatory immune response. Cattle were evaluated before and after vaccination to identify the CI variables that are most indicative of sick cattle. Respiratory rate (H2 = 43.08, P < 0.0001), nasal discharge (H2 = 8.35, P = 0.015), and ocular discharge (H2 = 16.38, P = 0.0003) increased after vaccination, and rumen fill decreased (H2 = 20.10, P < 0.0001). Locomotor activity was measured via leg-mounted sensors for the four days preceding and seven days following vaccination. A statistical model that included temperature, steps, lying time, respiratory rate, rumen fill, head position, and excess saliva was developed to distinguish between scores from before and after vaccination with a sensitivity of 0.898 and specificity of 0.915. Several clinical illness signs were difficult to measure in practice. Binoculars were required for scoring respiratory rate and eye-related metrics, and cattle had to be fitted with colored collars for individual identification. Scoring each animal took up to three minutes in a small research pen; therefore, technologies that can automate both behavior monitoring and identification of clinical illness signs are key to improving capacity for BRD detection and treatment.
PMID:38662753 | DOI:10.1371/journal.pone.0302172
