Investigating Reliability of Qualified Saddle Fitters and Coaches When Observing Saddle Fit on the Horse During Ridden Exercise

Introduction: Saddles play a crucial role in horse-rider interaction, impacting balance, comfort, and performance. Ill-fitting saddles can compromise performance, alter kinematics and induce back problems and asymmetry in both the horse and rider (Greve and Dyson, 2013). Regular assessments by a qualified saddle fitter are recommended. Despite perception of professionalism and qualification status, subjective challenges within the saddle fitting industry persist. Additionally, coaches’ ability to recognise saddle issues is pivotal for rider guidance. This study addresses gaps in understanding agreement in dynamic saddle fit assessments among professionals, aiming to enhance industry standards and credibility. This study aimed to quantify agreement within three group variables: 1) among Society of Master Saddlers qualified saddle fitters (QSF), 2) among United Kingdom Coaching Club/British Horse Society qualified coaches and 3) between the two groups when observing saddle fit on the horse during ridden exercise, using Likert-scale observation sheets.
Materials and Methods: Eight QSFs and four coaches observed twenty-nine horse-rider-pairs conducting a ridden exercise test, evaluating dynamic saddle fit criteria. Recruitment was conducted using social media and snowball sampling, enlisting riders ≥18 years old, pain-free and not undergoing musculoskeletal treatment, with horses engaged in regular activity and the pair demonstrating confidence in various environments. Participation required saddle fitting by a qualified saddle fitter within six months. Observers required proof of industry-qualification. Data collection replicated a dressage competition day, featuring a rider-prescribed warm-up and standardised exercise test. Likert-scale responses for observation sheets related to suggested dynamic saddle fit criteria were used (Figure 1). Inter-observer reliability and agreement were calculated using Fleiss Kappa, percentage observed agreement, and correlation tests.
Results: Agreement varied from poor to fair and was dependent on the criterion evaluated and the group assessed. Poor agreement was found for saddle length among coaches (k=-0.134) and between QSF and coaches (k=-0.041). Slight agreement was found across all group variables for stirrup level (1) k=0.207, 2) k=0.164, 3) k=0.158), and also for overall saddle fit (1) k=0.146, 2) k=0.170, 3) k=0.078). Slight agreement was found for saddle stability among QSF (k=0.129) and between the two groups (k=0.154), also for saddle balance (1) k=0.126, 3) k=0.152). Slight agreement was found for saddle length among QSF (k=0.049). Fair agreement was found among coaches for saddle stability (k=0.371) and saddle balance (k=0.306).
Discussion and Conclusion: Desirable agreement levels ideally range from moderate to perfect (Landis and Koch, 1977), yet none of the evaluated group variables achieved this. Varying levels of agreement were observed among QSF and coaches in assessing saddle fit dynamically, with notably lower agreement on saddle length which is a factor that significantly alters equine spinal kinematics. Factors such as observation angle, communication limitations, and differing expertise purportedly influence agreement levels (Guire et al., 2017). This highlights the need for standardised guidelines and objective measures to validate saddle fit. Collaborative efforts between QSF and coaches are crucial for holistic horse-rider care. Standardised protocols considering visibility, communication, and evaluation criteria are essential to promote consistency and trust in the industry, ultimately enhancing horse welfare and performance.