Agreement between subjective gait assessment and markerless video gait-analysis in endurance horses.

Abstract

BACKGROUND: Subjective evaluation of gait by official endurance veterinarians (OEVs) is used to determine 'fitness-to-compete' in horses participating in endurance competitions. Objective gait analysis systems could aid in quick and verifiable judgements. OBJECTIVES: To assess the agreement between objective analysis of head and pelvis vertical movement asymmetry performed with a markerless artificial intelligence motion tracking system (AI-MTS) and subjective lameness assessment performed by an accredited FEI OEV to judge horse gaits. STUDY DESIGN: Cross-sectional. METHODS: During three endurance competitions, 110 horses were enrolled. The OEV performed 188 gait examinations, which were simultaneously recorded with a smartphone. The vertical motion asymmetry of the head and pelvis was later analysed from the videos through the AI-MTS application. The gaits were scored as 'no asymmetry', 'mild asymmetry' or 'severe asymmetry'. The agreement was evaluated using Fleiss' multi-rater kappa statistic (&#x3ba;). RESULTS: The overall agreement between the two methods was fair (k&#x2009;=&#x2009;0.26, p&#x2009;<&#x2009;0.001). Within the three gait asymmetry categories, substantial agreement was obtained for the 'severe' (k&#x2009;=&#x2009;0.75, p&#x2009;<&#x2009;0.001) category, fair agreement was detected for the 'no asymmetry' category (k&#x2009;=&#x2009;0.25, p&#x2009;<&#x2009;0.001), and no agreement was identified for the 'mild' category (k&#x2009;=&#x2009;0.13, p&#x2009;=&#x2009;0.08). MAIN LIMITATIONS: Comparison between AI-MTS and a single OEV; absence of a tripod during video recording; and video recording from a different point of view than the OEVs. CONCLUSIONS: Mild asymmetry was the most challenging gait category to identify. Substantial agreement between the subjective lameness evaluation by OEV and AI-MTS assessment was observed for the 'severe' category. AI-MTS may be a helpful tool to assist OEVs in decision-making during endurance competitions.