Technical note: The development of a reliable 5-point gait scoring system for use in dairy goats

Abstract

Numerical rating scales are frequently used in gait scoring systems as indicators of lameness in dairy animals. The gait scoring systems commonly used in dairy goats are based on 4-point scales that focus on detecting and judging the severity of a definite limp. An uneven gait, such as a shortened stride or not "tracking up," is arguably the precursor to the development of a limp; thus, identifying such changes in gait could provide opportunity for early treatment. The objectives of this study were (1) to develop a 5-point gait scoring system that included an "uneven gait" category and compare the distribution of gait scores generated using this system to scores generated using a 4-point system, and (2) to determine whether this system could be reliably used. Forty-eight Saanen cross 2- and 3-yr-old lactating does were enrolled from a commercial dairy goat farm. Two observers carried out weekly live gait scoring sessions for 7 wk using the developed 5-point scoring system. The first 2 wk were used as training sessions (training sessions 1-2), with the subsequent 5 wk completed as gait assessments (assessments 1-5). In addition to training session 1 being lived scored, the goats were also video-recorded. This allowed observer 1 to re-score the session 4 times: twice using the developed 5-point system and twice using the previously used 4-point system. Comparisons of score distributions could then be made. Using the 4-point system, 81% of the goats were assigned score 1 (normal gait). Using the 5-point system, only 36% of the goats were assigned score 1 (normal gait), with 50% assigned score 2 (uneven gait). High levels of intra-observer reliability were achieved by observer 1 using both gait scoring systems [weighted kappa (κw) = 1.00: 4-point, κw = 0.96: 5-point]. At training session 1 (wk 1), inter-observer reliability was only moderate (κw = 0.54), but this was improved during the subsequent training session 2 (κw = 0.89). Inter-observer reliability was high among assessments 1 to 5 (κw = 0.90-1.00). During the training sessions, sensitivity for gait scores 1 and 2 was 77 and 65% (training session 1) and 89 and 94% (training session 2), respectively. Sensitivity was high among assessments 1 to 5 (score 1: 83-100%, score 2: 97-100%). This highlights the likely reason why existing gait scoring systems for dairy goats do not include an "uneven gait" category, as distinguishing it from a normal gait was challenging without training. In conclusion, with training, a 5-point gait scoring system could be reliably used. The 5-point system was found to be more sensitive than the 4-point system, allowing for a potential precursor to lameness to be identified. Further work is needed to determine whether the score can be reliably used in an on-farm setting.