Sheep lameness detection from individual hoof load

Abstract

Lameness is a leading cause of morbidity in sheep but routine inspection of sheep for lameness is labor intensive. The objectives of the current study were to (i) build and test a custom hoof weigh crate (HWC) to measure the individual hoof load of sheep, (ii) quantify the relationship between hoof health status and the load a sheep distributes to each hoof, and (iii) evaluate the ability of the HWC to differentiate healthy from infected hooves. The overall footprint of the HWC was 950 mm × 450 mm, wherein two strain-gauge load cells were placed underneath each of four individual hoof platforms. An experiment was conducted over nine non-consecutive days between July and October 2017. On each experimental day, a total of 20 ewes (consisting of lame and healthy ewes) were placed on the HWC for five-minutes each. Each sheep hoof was visually assessed for lesions by three independent operators and a hoof lesion score assigned (scale from 0- healthy to 4-severe footrot). In addition to individual hoof load, the load placed on each hoof was divided by the sum of the load of the respective contralateral pair (front or back hooves), and multiplied by 100 to express the contralateral load percentage. A linear mixed model was invoked for each of the two load parameters as the dependent variable while hoof lesion score, contralateral pair, and their interaction were included as fixed effects. Each hoof was classified into a hoof lesion score category based on its load parameter values, and the numbers of correct and incorrect classifications were quantified. Healthy front hooves naturally carried more weight (60% of total weight) in comparison to healthy back hooves (40% of total weight), but when front or back hooves were infected to the same extent, they carried the same load. Results from the linear mixed model showed a small mean difference (4.5 kg) in hoof load between healthy front hooves and those with a mild infection (i.e., score = 1), but there was no hoof load difference between healthy back hooves and those with a mild infection (P < 0.05). The lowest proportion of misclassified hooves (sensitivity of 100% and specificity of 95%) was observed when the contralateral load percentage was used to differentiate between healthy hooves and those with extensive inter-digital dermatitis. Results herein indicate that the HWC could be used to automatically detect extensive infection in sheep hooves.