Repeatability of subjective evaluation of lameness in horses.

REASONS FOR PERFORMING STUDY: Previous studies have suggested that agreement between equine veterinarians subjectively evaluating lameness in horses is low. These studies were limited to small numbers of horses, evaluating movement on the treadmill or to evaluating previously-recorded videotape.
OBJECTIVES: To estimate agreement between equine practitioners performing lameness evaluations in horses in the live, over ground setting.
METHODS: 131 mature horses were evaluated for lameness by 2-5 clinicians (mean 3.2) with a weighted-average of 18.7 years of experience. Clinicians graded each limb using the AAEP lameness scale by first watching the horse trot in a straight line only and then after full lameness evaluation. Agreement was estimated by calculation of Fleiss' (kappa). Evaluators agreed if they picked the same limb as lame or not lame regardless of the severity of perceived lameness.
RESULTS: After only evaluating the horse trot in a straight line clinicians agreed whether a limb was lame or not 76.6% of the time (kappa= 0.44). After full lameness evaluation clinicians agreed whether a limb was lame or not 72.9% of the time (kappa= 0.45). Agreement on forelimb lameness was slightly higher than on hindlimb lameness. When the mean AAEP lameness score was >1.5 clinicians agreed whether or not a limb was lame 93.1% of the time (kappa= 0.86), but when the mean score was < or = 1.5 they agreed 61.9% (kappa= 0.23) of the time. When given the task of picking whether or not the horse was lame and picking the worst limb after full lameness evaluation, clinicians agreed 51.6% (kappa= 0.37) of the time.
CONCLUSIONS: For horses with mild lameness subjective evaluation of lameness is not very reliable. POTENTIAL RELEVANCE: A search for and the development of more objective and reliable methods of lameness evaluation is justified and should be encouraged and supported.