Agreement between two inertial sensor gait analysis systems for lameness examinations in horses
Authors: T. Pfau, H. Boultbee, H. Davis, A. Walker, M. Rhodin
Journal: Equine Veterinary Education
Summary
# Editorial Summary: Inertial Sensor Systems in Equine Gait Analysis Two inertial sensor systems commonly used for quantifying lameness through trunk movement symmetry—one measuring dorsoventral movement in the local reference frame and the other measuring global vertical movement—were tested for interchangeability when examining 13 horses equipped with both devices simultaneously. Agreement between the systems was closer for pelvic movement (limits of agreement 2.5–5.3 mm across in-hand and lunge work) than for head movement (6.4–9.7 mm), with pelvic measurements showing substantially tighter clustering. The widths of limits of agreement for head movement align closely with the currently accepted 6 mm threshold used clinically for lameness detection, whilst pelvic measurements fall near or marginally below the 3 mm threshold, meaning that measurement differences between the two systems of this magnitude reflect instrumental variation rather than actual changes in the horse's locomotion. For practitioners undertaking multi-centre lameness studies or comparing a horse's results across different facilities, these findings indicate that switching between systems should not be interpreted as clinical deterioration or improvement, though individual facilities should ideally maintain consistency with a single system to minimise this source of measurement variance.
Read the full abstract on the publisher's site
Practical Takeaways
- •If using inertial sensor gait analysis in multi-centre studies, expect systematic differences between the two common systems; differences up to ~10 mm in head movement or ~5 mm in pelvic movement may reflect equipment variation rather than true lameness changes
- •Pelvic measurements are more reproducible between systems than head measurements, so prioritise pelvic symmetry data when comparing horses assessed on different equipment
- •Avoid switching between systems mid-investigation on individual horses, as this introduces measurement noise that could obscure genuine lameness progression
Key Findings
- •Limits of agreement between two inertial sensor systems ranged from 6.4–9.7 mm for head movement and 2.5–5.3 mm for pelvic movement depending on assessment context
- •Pelvic movement measurements showed narrower limits of agreement than head movement between the two systems
- •Measurement differences falling within reported thresholds (6 mm head, 3 mm pelvis) should not be interpreted as changes in lameness when switching between systems
- •The two systems are not perfectly interchangeable but differences are of comparable magnitude to current clinical lameness thresholds