Reliability of equine visual lameness classification as a function of expertise, lameness severity and rater confidence.
Authors: Starke Sandra Dorothee, Oosterlinck Maarten
Journal: The Veterinary record
Summary
# Editorial Summary: Visual Lameness Assessment Reliability Starke and Oosterlinck examined a persistent clinical challenge using three-dimensional horse animations with induced movement asymmetries ranging from 0–60 per cent, rated by equine professionals with varying levels of experience at a veterinary seminar. Surprisingly, neither years of clinical experience nor caseload exposure significantly improved the ability to identify lame or sound horses, with high hindlimb lameness sensitivity particularly poor—only 28 per cent accuracy in correctly classifying sound animals as sound, compared with 72 per cent for forelimb assessment. Rater confidence bore no relationship to diagnostic accuracy, and participants frequently misidentified which limb was affected in hindlimb cases or failed to detect subtle forelimb lameness entirely, with some assessment metrics performing at or below chance level. These findings suggest visual gait evaluation alone cannot reliably distinguish between sound and mildly lame horses regardless of assessor expertise, which has significant implications for clinical practice and the design of lameness protocols. The marked asymmetry between forelimb and hindlimb assessment accuracy highlights a specific knowledge gap that may warrant targeted training approaches alongside objective measurement tools such as force plate analysis or inertial measurement units, particularly when subtle lameness detection is clinically important.
Read the full abstract on PubMed
Practical Takeaways
- •Visual assessment alone cannot reliably detect mild lameness or hindlimb lameness in horses, regardless of your experience level—use objective diagnostic tools when lameness is suspected
- •High confidence in your visual assessment does not correlate with accuracy; remain skeptical of subtle lameness diagnoses made purely by observation
- •Hindlimb lameness is particularly difficult to detect visually and is frequently overdiagnosed (false positives); seek additional diagnostic confirmation before treatment decisions
Key Findings
- •Years of experience and case load exposure had no significant effect on correct assessment of lame or sound horses, except for forelimb lameness at 60% asymmetry (P=0.014)
- •Forelimb lameness was correctly classified as sound 72% of the time on average, while hindlimb lameness was only correctly classified as sound 28% of the time
- •Rater confidence was not associated with diagnostic performance
- •Visual gait assessment is unreliable at differentiating between sound and mildly lame horses regardless of assessor expertise or background