Abstract
The primary objective of this study was to investigate whether rider experience influences the assessment and grading of lameness in horses based on under-saddle gait analysis. Thirteen adult sports horses in active training were included in the study. After a baseline lameness and neurologic examination by the principal investigators, horses were videotaped while being ridden by an experienced and a less experienced rider. A 3-minute video was made for each horse and rider and 26 videos were randomly ordered and compiled on a DVD. Veterinarians with different levels of experience in evaluating lameness and veterinary students viewed the DVD and assigned a lameness score to each horse/rider combination. In a model accounting for the expertise of the evaluator, there was no difference in overall lameness scores between experienced and less experienced riders. This result was consistent for both sound and unsound horses. The overall lameness scores reported by specialists and students, however, differed significantly. The lameness score reported by the study participants while the horse was ridden was significantly associated with the subjective baseline lameness assessment reported by the principal investigators for the same limb when the horse was not under saddle. Additional work is necessary to determine whether riders with even lower skill levels would further alter the balance and motion pattern of the horse and have more influence on subjective grading of lameness.
Résumé
L’objectif principal de la présente étude était d’examiner si l’expérience du cavalier influence l’évaluation et la gradation de la boiterie chez des chevaux basées sur l’analyse de la démarche sous-selle. Treize chevaux sportifs adultes en entraînement actif ont été inclus dans l’étude. Suite à un examen neurologique et de boiterie de base par les chercheurs principaux, les chevaux ont été filmés alors qu’ils étaient montés par un cavalier d’expérience et un cavalier moins expérimenté. Une vidéo de 3 minutes a été réalisée pour chaque combinaison cheval/cavalier et les 26 vidéos ont été compilées de manière aléatoire sur un DVD. Des vétérinaires ayant différents niveaux d’expérience à évaluer les boiteries et des étudiants vétérinaires ont visualisé le DVD et donné un pointage à chaque combinaison cheval/cavalier. Dans un modèle tenant compte de l’expérience de l’évaluateur, il n’y avait aucune différence dans les pointages globaux entre un cavalier expérimenté et un moins expérimenté. Ce résultat était constant autant pour les chevaux solides que fragiles. Toutefois, les pointages globaux de boiterie notés par les spécialistes et les étudiants étaient significativement différents. Le pointage de boiterie rapporté par les participants à l’étude alors que le cheval était conduit était significativement associé avec l’évaluation subjective de base rapportée par les chercheurs principaux pour le même membre lorsque le cheval n’était pas sous-selle. Des études additionnelles sont nécessaires afin de déterminer si des cavaliers avec encore moins d’expérience influenceraient encore plus l’équilibre et le patron de mouvement du cheval et aurait plus d’influence sur la gradation subjective de la boiterie.
(Traduit par Docteur Serge Messier)
Introduction
Visual gait assessment and lameness grading are critical to the diagnosis and ongoing evaluation of musculoskeletal diseases in horses (1–4). Previous studies have shown that the American Association of Equine Practitioners (AAEP) lameness grading system has a low degree of interobserver agreement (1,2,5,6). There are also modern quantitative and more objective techniques for gait assessment such as kinematic gait analysis. While there should be less variability between observers with objective kinematic analysis of gait and movement characteristics using more precise measurements, these methods can be impractical in a practice setting and should be used in conjunction with clinical findings (7,8).
Evaluation under saddle is an important component of identification, diagnosis, and ongoing assessment in the initial lameness examination and during treatment (3). Certain orthopedic clinical signs and musculoskeletal conditions become apparent only when the horse is ridden (3,9). In addition, lameness is often evaluated and monitored for improvement based on performance (3,4). Kinematic parameters, including gait characteristics, are highly influenced by the instructions given by the rider (behavioral effects) as well as body positioning of the rider (inertial effects) (7,9,10,11). The effect of the rider on the horse’s motion pattern could interfere with the ability of veterinarians to correctly attribute a disturbed motion pattern to either an intrinsic process in the horse, such as musculoskeletal abnormalities, or to external influences, such as a lack of synchrony between horse and rider (9). However, the degree to which the rider influences subjective evaluations of lameness under saddle, either by disguising or enhancing gait abnormalities, has not been well-documented. We hypothesized that lameness grading under saddle will be altered because of rider experience.
The primary objective of this study was to determine whether the rider’s experience influences the gait and lameness score of sound and unsound horses. The second objective was to determine whether or not gait abnormalities are scored differently by veterinary specialists, equine practitioners, and veterinary students. The final objective was to compare the results of the traditional baseline lameness examination by the principal investigators (PIs) to the lameness score reported by the students and participating veterinarians for the same limb when the horse was under saddle.
Materials and methods
Study design
Thirty-two sports horses in active training were initially screened and evaluated for abnormal gait by the principal investigators (F.M. and F.A.). Only sound horses and horses with abnormal gait due to mild lameness [American Association of Equine Practitioners (AAEP) lameness score 1 or 2] during the baseline examination were included in the study. Horses with an AAEP lameness score of ≥ 3 during the baseline examination were excluded from the study due to concerns about purposely riding and making an obviously lame horse perform a series of potentially painful exercises. Horses were also excluded from the study if neurologic examination revealed deficits. All horses were privately owned, more than 4 y old, and active in dressage, show jumping, hunter jumping, or western pleasure. Consent was obtained from the owners. The study was approved by the Animal Research Ethics Board at the University of Saskatchewan and followed the guidelines of the Canadian Council on Animal Care.
Nine trainers and owners agreed to participate and selected horses in active training. Horses were included in the “lame” group if at the time of the baseline examination: 1) they had been diagnosed or had been treated for lameness by a veterinarian (other than the authors) within the previous 3 mo; and 2) the owner reported intermittent lameness due to this condition within the previous 30 d. In addition, in order to be included in the “lame” group, the horses had to have no neurologic deficits, a 1 or 2 out of 5 AAEP lameness score, and a positive response to hoof tester or a positive flexion test that were consistent with the previous diagnosis when the principal investigators carried out lameness and neurologic examinations. Horses were included in the “sound” group if their owner or trainer reported no history of lameness in the previous 30 d and there were no abnormal findings during the baseline lameness or neurologic examinations for this study. Horses were therefore classified as sound if they had no abnormalities on palpation, flexion tests, or hoof testing, and no lameness was noted by the PIs during hand walking or trotting or when evaluated under saddle (AAEP lameness score 0).
Out of the 32 horses initially evaluated, only 13 horses met the inclusion criteria and were included in the study: 6 horses in the “lame” group and 7 in the “sound” group.
Baseline lameness and neurologic examinations
Baseline lameness and neurologic examinations were conducted by 2 equine practitioners (F.M. and F.A.). Neurologic examination evaluated the horse’s mentation and behavior, posture, signs of muscle mass asymmetry or atrophy, and gait. Horses were then assessed for signs of cranial nerve deficits, ataxia, proprioceptive deficits, paresis, spasticity, or dysmetria.
Lameness examination by the PIs at the study baseline included a distant examination, palpation and manipulation of the limbs and back, hoof testing, and upper and lower limb flexion tests. The horses were examined at a distance to detect any abnormality in conformation, posture, or symmetry. Each horse’s gait was evaluated at a walk and a trot in a 25-m straight line on a flat hard surface, while being led with a loose lead shank to allow free motion of the head and neck. Palpation was conducted in the following order: the forelimbs, the neck, the back, and pelvic regions, and then the hind limbs. The limbs were evaluated in a systematic fashion, starting at the hoof and proceeding proximally, including assessing bony structures, soft tissues, and joints for evidence of swelling, localized heat, deformities, effusion, degree of passive motion, and pain. Joints and tendinous structures were palpated with the horses bearing weight on the limb as well as with the limb lifted off the ground. Hoof testers were applied to the frog and the sole from heel to toe on both sides of the hoof, assessing elicited pain. Limb flexions were conducted for 1 min and horses were immediately evaluated at a trot in hand in a straight line moving away from the examiner. Any exacerbation of lameness or lameness not previouly seen and observed over at least 3 strides was classified as a positive response.
Finally, the PIs conducted a baseline assessment of the horse’s gait under saddle to detect and grade any gait abnormality, alteration of the cranial or caudal phase of the stride, or presence of an abnormal limb flight. The summary baseline gait analysis was recorded based on the previously mentioned examination from a distance, hands-on examination, and under-saddle gait evaluation. An overall baseline lameness score (AAEP lameness score) was assigned for each horse after all gait and locomotor patterns were assessed. Findings of the lameness and neurologic examinations were recorded on a standardized form. Baseline data from each limb of each horse used in subsequent analyses included whether lameness was observed at a walk and trot in hand, positive or negative flexion test, and positive or negative hoof tester result.
Rider selection
All horses in each group were ridden by 1 inexperienced and 1 experienced rider. Riders designated as the lower level experience group (“inexperienced riders”) had ridden horses for a maximum of 1 y, spent a similar number of hours riding (maximum of 45 min riding time per day and no more than 3 riding days per week), and had completed a comparable number of formal lessons (maximum of 20). Riders that participated as part of the experienced group (“experienced riders”) were also selected to be as similar as possible in terms of riding ability and number of years riding (minimum of 5 y) and had succesfully competed (placed 1st, 2nd, or 3rd at a sports horse show) at a high level within their sport during the previous 5 y. The selected riders were also of similar body weights. Five experienced riders, 2 of whom were professional riders, and 4 inexperienced riders met the inclusion criteria and agreed to participate in the study.
Videotaping horses under saddle
Immediately after the baseline lameness and neurologic examinations were completed, the 13 horses that met the inclusion criteria for the “lame” or “sound” groups were videotaped while being ridden by 1 experienced and 1 inexperienced rider through a standardized series of locomotor patterns. The locomotor patterns and riding exercises were designed to enable the horses to be evaluated at a walk, trot, and canter as well as to incorporate lateral flexion (Figure 1). Both the experienced and inexperienced riders were selected from the riding facility where the horses were stabled. The order in which the experienced and inexperienced riders rode the horses was systematically randomized. Each horse was ridden by the 2 riders with only the minimum time necessary to allow 1 rider to dismount and the second rider to mount the horse. The riders were blinded to the history and baseline lameness examination results for all horses.
Figure 1.
Riding patterns.
All horses were videotaped in the ring at their riding facility from the same angle and at the same distance from the video camera. Eleven horses were ridden in 1 of 4 different rings with sand footing and 2 horses were ridden in 1 of 2 different rings with grass footing. The arenas were all similar in size, with a minimum of 20 × 40 m (approximately 65.6 × 131 ft).
Video editing and lameness evaluation by study participants
The videos were edited to exclude the rider from the frame (Figure 2). A 3-minute video was made for each horse and rider combination and all footages (total 26) were randomly ordered and compiled on a DVD. Each footage was assessed by a group of veterinarians with varying expertise in sports medicine and lameness analysis [4 American College of Veterinary Surgeons (ACVS) Diplomates, 3 ACVS senior surgery residents, and 5 equine practitioners], as well as 5 third-year veterinary students with a strong sports-horse background. They were blinded to the selection criteria and baseline examination results for the horses and were asked to identify any gait abnormality and assign a lameness score to individual limbs at each specific gait and an overall lameness score based on all gaits and locomotor patterns.
Figure 2.
Still picture from one of the video footages. The videos were edited such that the rider’s body was blanked out. The rider’s aids and commands given to the horse were not seen and therefore the riding skills of the riders were not obvious to the study participants.
Since the AAEP scoring system grades lameness at both the walk and the trot combined and is therefore not suitable for assessing individual limbs at specific gaits (walk only, trot only and canter only), a 3-grade lameness score was used for that purpose. The 3-grade lameness score was used to grade the individual limbs at each specific gait (walk, sitting trot, rising trot, and canter). This lameness score was defined as: score 0 — lameness not perceptible (sound); grade 1 — lameness difficult to observe; and grade 2 — lameness consistently observable.
The study participants used a standard 5-grade AAEP lameness scoring system to assign an overall lameness score for each limb combining all gaits. Grades 0, 1, and 2 of the 3-grade lameness score system used to assess individual limbs at single specific gaits corresponded to the same grades (0, 1, and 2) of the traditional AAEP lameness score system used to assess individual limbs combining all gaits. A standard form was used to record the video evaluation lameness scores for each horse and rider combination recorded in the footage.
Statistical analysis
All data were entered, recoded, and summarized in a commercial spreadsheet program (Microsoft Excel; Microsoft, Redmond, Washington, USA). The lameness scores measured in this study should be considered as ordinal data. Ordinal logistic regression models, however, would not converge after considering the important fixed and random effects.
Therefore to address the first 2 objectives, linear mixed models (PROC MIXED; SAS, Version 9.2, SAS Institute, Cary, North Carolina, USA) were used to evaluate the association of the following variables with overall lameness score for each limb (n = 1768 observations) as reported by the study participants after viewing the video footage for each horse and rider combination. The variables examined as fixed effects in the model included horse soundness as assessed by the PIs, rider experience, the interaction between rider experience and soundness, and evaluator experience. Limb position, reported as front compared to hind limbs, was also evaluated as a potential modifier for the effect of rider experience. The model accounted for the lack of independence in the data associated with multiple measurements using random intercepts for limb (n = 4 per horse-rider combination in each video footage), unique horse and rider combination captured in the video footage (n = 13 horses × 2 riders), and evaluator (n = 17). The results of the models were reported as the difference associated with category of rider experience and then the difference associated with category of observer experience, as well as the least squared mean lameness scores with 95% confidence intervals (CIs) for each category (95% CI).
Variance partition coefficients (VPCs) were estimated to determine the contributions of the random effects to the final model using the following formulas: 1) VPClimb = σ2limb/(σ2limb + σ2horse*rider combination + σ2evaluator); 2) VPChorse*rider combination = σ2horse*rider combination/(σ2limb + σ2horse*rider combination + σ2evaluator); and 3) VPCevaluator = σ2evaluator/(σ2limb + σ2horse*rider combination + σ2evaluator).
To evaluate the third objective, linear mixed models for continuous outcomes (PROC MIXED, SAS, Version 9.2, SAS Institute) were again used to evaluate whether baseline assessments by the PIs (numeric value) were associated with the overall lameness score for each limb reported by study participants based on assessment of the video footages (n = 1768 observations). The following baseline results from the principal investigors were considered as fixed effects in a series of models for overall limb lameness score: 1) the numeric value of lameness score; 2) whether or not the PIs reported a lameness score of 2 at a walk and trot in hand (1 or 0); 3) positive flexion test for the limb (1 or 0); and 4) positive hoof tester for the limb (1 or 0). The limb position (left fore, left hind, right fore, right hind) and expertise of the assessor were also considered as potential confounders in all models. This model also included random intercepts for limb (n = 4 per horse-rider combination in each video footage), unique horse and rider combination captured in the video footage (n = 13 horses × 2 riders), and evaluator (n = 17).
In the second step for the third objective, generalized linear mixed models (PROC GLIMMIX, Version 9.2; SAS Institute) with a logit link function and binomial distribution for dichotomous outcomes (whether or not the participant score was ≥ 2) were evaluated using the same fixed and random effects. The results of the mixed models were reported as regression coefficients with 95% CI and the results of the generalized linear mixed models were reported as odds ratios (OR) with 95% CI.
For all models with continuous outcomes, the residuals and random effects (best linear unbiased predictors) were graphed against significant predictors to assess the homogeneity of the variance. The residuals and random effects at each level were also examined graphically to determine whether the residuals were adequately normally distributed. Finally, for both continous and dichotomous outcomes, the residuals and random effects were graphed against the predicted values to determine whether there were any extreme outliers or influential results.
Results
Selected horses
Horses from various breeds and activities were included in the study. All horses were in active training and stabled at private riding facilities around Saskatoon, Saskatchewan, Canada and Buenos Aires, Argentina. Five horses (38%) were Warmbloods, 3 (23%) were American Paints, 2 (15%) were Quarter horses, 2 (15%) were Thoroughbreds, and 1 (8%) was Arabian. Of the 13 horses participating in the study, 6 (46%) were competing in show jumping at an intermediate level, 6 (46%) were used for English and Western pleasure riding, and 1 (8%) was a top performance endurance horse. All horses were trained at the same facility where they were stabled, examined, and videotaped. Nine horses (69%) were female and 4 (31%) were geldings. Horses varied in age from 4 to 18 y [mean: 7.4; standard deviation (SD): 4.0; median: 7].
Summary of lameness examination results
The PIs assigned 6 of the 13 horses with an overall AAEP baseline lameness score of 1 or 2, in addition to ≥ 1 positive result on flexion testing or hoof testing (Table I). At a trot, 1 out of 13 horses was lame in the left front limb (LF) and 2 were lame in the left hind limb (LH). One of 13 horses tested positive in a flexion test of the LF, 1 tested positive in a flexion test of the right front limb (RF), 4 tested positive in a flexion test of the LH, and 5 tested positive in a flexion test in the right hind limb (RH). When assessed with hoof testers, 1 horse tested positive in the LF and 2 horses tested positive in the RF.
Table I.
Baseline lameness examination results obtained by the principal investigators (PI). A lameness score was assigned to each horse while hand-walking and hand-trotting in a 25-m straight line, in addition to an overall lameness score following the AAEP guidelines. The “lame” group was comprised of horses with a lameness history, an overall AAEP score of 1 or 2, as well as either a positive upper or lower limb flexion test or a positive hoof test response (horses ID # 6, 10, 14, 18, 19, and 25)
| Horse ID | Lameness score: walk and trot in a straight line | Flexion tests | Hoof test | Lame under saddle | Overall AAEP lameness score and limb assigned by PI | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
||||||||||
| LF | RF | LH | RH | LF | RF | LH | RH | LF | RF | |||
| 2 | 0 | 0 | 0 | 0 | − | − | − | − | − | − | no | 0 |
|
| ||||||||||||
| 6 | 0 | 0 | 0 | 0 | − | − | − | − | + | + | yes | 1 LH |
|
| ||||||||||||
| 9 | 0 | 0 | 0 | 0 | − | − | − | + | − | − | no | 0 |
|
| ||||||||||||
| 10 | 0 | 0 | 0 | 0 | − | − | + | + | − | + | yes | 1 LH |
|
| ||||||||||||
| 11 | 0 | 0 | 0 | 0 | − | − | − | − | − | − | no | 0 |
| 12 | 0 | 0 | 0 | 0 | − | − | − | − | − | − | no | 0 |
| 13 | 0 | 0 | 0 | 0 | − | − | − | − | − | − | no | 0 |
|
| ||||||||||||
| 14 | 0 | 0 | 2 | 0 | − | − | + | − | − | − | yes | 2 LH |
| 18 | 0 | 0 | 0 | 0 | − | − | + | + | − | − | yes | 1 LH |
| 19 | 2 | 0 | 0 | 0 | + | + | − | − | − | − | yes | 2 LF |
|
| ||||||||||||
| 21 | 0 | 0 | 0 | 0 | − | − | − | − | − | − | no | 0 |
|
| ||||||||||||
| 25 | 0 | 0 | 2 | 0 | − | − | + | + | − | − | yes | 2 LH |
|
| ||||||||||||
| 26 | 0 | 0 | 0 | 0 | − | − | − | + | − | − | no | 0 |
LF — left front limb; RF — right front limb; LH — left hind limb; RH — right hind limb; AAEP — American Association of Equine Practitioners.
Based on the overall AAEP scores from the evaluation of video footage of each horse, the results among the study participants were much more variable. Maximum scores assigned to any limb among all study participants ranged from 0 to 2 out of 5 for 1 horse, 0 to 3 for 6 horses, 0 to 4 for 3 horses, 1 to 3 for 2 horses, and 2 to 4 for 2 horses.
Rider experience and reported lameness score from video footage
In a model (n = 1768) accounting for the expertise of the evaluator, the effect of rider experience on the overall lameness score for each limb did not vary whether or not the horse was considered sound, i.e., there was no interaction between rider experience and the horse’s classification as sound or lame on entry to the study (P = 0.72). There was also no interaction between rider experience and whether the lameness was observed in the front or hind limb (P = 0.40).
In a model accounting only for the expertise of the evaluator, there was no difference in overall lameness scores between experienced and inexperienced riders (difference 0.017, 95% CI: −0.065 to 0.097, P = 0.70). The mean lameness score was 0.42 (95% CI: 0.28 to 0.55) for the “inexperienced rider group” and 0.40 (95% CI: 0.26 to 0.54) for the “experienced rider group.”
Evaluator experience and reported lameness scores from video footage
There were no significant differences in the overall lameness scores reported by practitioners and specialists (difference −0.05, 95% CI: −0.24 to 0.14, P = 0.60) or those reported by practitioners and students (difference 0.15, 95% CI: −0.05 to 0.35, P = 0.14) (n = 1768). However, there was a significant difference in the overall lameness scores reported by specialists and students (difference 0.20, 95% CI: 0.01 to 0.39, P = 0.04). The mean lameness scores were 0.44 (95% CI: 0.26 to 0.62) for the “practitioners group,” 0.49 (95% CI: 0.33 to 0.65) for the “specialists group,” and 0.29 (95% CI: 0.11 to 0.47) for the “students group.” The VPC was 0.03 for limb, 0.002 for horse and rider combination, and 0.22 for evaluator. In the final model, the proportion of variance not explained by the random effects was 0.74.
Association between PI baseline evaluation and participant evaluation of video footage
Most measures of baseline lameness reported by the PIs were associated with the lameness score for the same limb reported by the study participants while the horse was being ridden, after accounting for the limb evaluated and the experience of the evaluator (n = 1768). The baseline lameness score for the limb reported by the PIs was associated with the participant lameness score based on the video footage. The mean participant lameness score increased by 0.58 units (95% CI: 0.39 to 0.78, P = 0.0001) for every 1 unit increase reported by the PIs. Similarly, whether or not the PIs assigned the limb a lameness score of 2 was also associated with the participants’ score for the limb. The mean lameness score assigned by study participants increased by 1.17 score points (95% CI: 0.78 to 1.56, P = 0.0001) when the PI baseline lameness score was 2 compared to when it was ≥ 2. Finally, if the PIs considered the limb to be positive on flexion test, the mean lameness score of study participants increased by 0.33 score points (95% CI: 0.08 to 0.57, P = 0.01). The results of a hoof tester examination were not associated with the lameness scores for the limb reported by the participants (P = 0.88).
We also considered whether or not the baseline lameness assessment by the PIs was associated with an increased likelihood that the study participants would report a lameness score of ≥ 2 for the same limb when viewing the video footage of the horse under saddle (n = 1768). Limbs classified by the PIs with a baseline lameness score of 2 (OR: 14.5, 95% CI: 8.2 to 25.4, P = 0.0001) were more likely to be assigned a lameness score of ≥ 2 by study participants observing the animals being ridden after accounting for limb and observer experience. Similarly, limbs reported as positive on the flexion test were also more likely to be given a score of ≥ 2 by study participants (OR: 4.0, 95% CI: 2.6 to 6.1, P = 0.0001). There was no significant association, however, between the hoof tester results and the likelihood that study participants would describe the limb as having a score of ≥ 2 (OR: 3.1, 95% CI: 0.8 to 12.4, P = 0.11).
Discussion
Although other studies have demonstrated the importance of the interaction between the rider and the horse and its effect on the horse’s gait and motion (9,10,12–14), the objective of this study was to determine to what extent riders with different levels of experience can influence subjective evaluations of under-saddle lameness, by either disguising or enhancing gait abnormalities. Rider experience was not associated with lameness scores in either sound or unsound horses in this analysis. This differs from previous reports that suggested subtle differences in the balance of the rider’s weight could mask or exacerbate lameness and that poor riding can potentially lead to a false-positive diagnosis of lameness (3). The absence of a rider experience effect in this study could have been the result of low study power due to a limited number of horses (lame and sound) and rider combinations or it could be a real finding. The relatively tight confidence intervals for the difference between experienced and less experienced riders, however, suggest that the power of the study was adequate for the range of rider experience and the extent and types of lameness included in this study.
The finding in this study that reported lameness did not differ with rider experience could also be due to smaller than expected variation in the ability of our experienced and less experienced riders. Without riding errors such as substantial unbalanced distribution of weight or inappropriate shift of center of gravity, our inexperienced riders would not have altered the horse’s balance and motion pattern in such a way as to influence the lameness grading. Less skilled riders could be less balanced, which could alter the gait of the horses enough to influence the lameness grading. While the inexperienced riders had only very basic riding skills, in general the PIs subjectively perceived them to be off balance for only very short periods of time, while overall their weight distribution and dynamic balance were adequate.
The expertise of the professionals evaluating lameness is also important to consider when using subjective scores to identify subtle changes in the horse’s balance and motion pattern. There were no significant differences in the overall lameness scores reported by practitioners and specialists or those reported by practitioners and students. The overall lameness scores reported by specialists and those reported by students, however, differed significantly. The specialist group consisted of 4 ACVS Diplomates with many years of experience in evaluating lame horses and 3 ACVS residents with less experience. The practitioners group included equine veterinarians from the equine field service unit, 1 with more than 10 y and the rest with 2 to 4 y of experience in equine practice. The similarity in overall lameness scores reported by the specialist and practitioner groups could reflect the fact that the ACVS senior residents that were included in the specialists group and the veterinarians with 2 to 4 y experience from the equine field service unit that were included in the practitioners group had comparable lameness identification skills. Furthermore, the students group included third year veterinary students with a strong sports-horse and riding background who could be expected to have observation skills approaching some of the less experienced practitioners. Only the differences in the overall lameness scores reported by specialists and students were statistically significant. It is also possible that lameness scores between evaluators would have differed more if horses with a broader range of lameness severity, such as AAEP grade 3, had been included in the study. The very small variance partition coefficients associated with limb and horse and rider combinations, however, suggest that differences between horses did not account for a substantial part of the unexplained variation in the final model for observer lameness score.
Previous studies have shown that there is a low degree of interobserver agreement on subjective gait analysis and lameness grading (2,5,6). In our study, the subjective lameness assessments by the PIs were significantly associated with the lameness score perceived by the study participants for each limb while the horse was being ridden. Limbs classified by the PIs as having a lameness score of 2 while the horses were not being ridden were more likely to be assigned a lameness score of ≥ 2 by study participants observing the horses being ridden. Similarly, limbs reported as positive on the flexion test were also more likely to be given a score of 2 or more by study participants. Although there was no association between the hoof tester results and the likelihood that participants would describe the limb as having a score of ≥ 2 in the present study, the very small number of positive hoof tester results severely limited our ability to address this question.
The intrinsic nature of subjectivity when assessing lameness with the naked eye, as well as multiple factors affecting the horse’s gait and motion, i.e., rider’s skills, balance, and weight, behavioral effects of the rider, instructions given by the rider, as well as type, cause, location, and severity of lameness, makes it challenging to conduct this type of research. Although in our study we accounted for many of these factors, such as the expertise of the evaluator, effect of rider experience, limb evaluated, and whether the horse was classified as lame or sound, many other influences on the horse’s gait and motion pattern could not be measured and considered in the analysis. Investigating a larger population of lame horses with a wider range of lameness severity and location within and among limbs is warranted. It would also be valuable to evaluate the effect of horse use and the type of lameness and to explore differences among individual riders and the type of arena footing more specifically. Furthermore, assessing a larger number of horses would also make it possible to determine whether the rider’s ability influences subjective grading of lameness depending on the diagonal on which the rider sits in rising trot and the presence of lameness in the front versus the hind limb.
In conclusion, rider experience did not change subjective grading of lameness in the present study when videotaped footage of horses under saddle was assessed. In a model accounting for the expertise of the evaluator, however, there was a significant difference in the overall lameness scores reported by ACVS Diplomates and residents and scores from senior veterinary students. Additional research studies are necessary in order to determine whether riders with even lower skill levels would further alter the balance and motion pattern of the horse and have a more apparent influence on the subjective grading of lameness.
Acknowledgments
This study was funded by the Western College of Veterinary Medicine (WCVM) Research Trust Fund at the University of Saskatchewan. The authors thank Peter Downing and Wayne Giesbrecht of Media Access and Production (eMAP) at the University of Saskatchewan for assistance in editing and compiling the video and Diego J. Roiz of Haras Villanueva for his collaboration with the equitation technical support.
References
- 1.Guide to Veterinary Services for Horse Shows. 7th ed. Lexington: American Association of Equine Practitioners; 1999. [Google Scholar]
- 2.Stashak TS. Examination for lameness. In: Stashak TS, editor. Adams’ Lameness in Horses. 5th ed. Philadelphia, Pennsylvania: Lippincott Williams and Wilkins; 2002. pp. 113–183. [Google Scholar]
- 3.Ross MW. Movement. In: Ross MW, Dyson SJ, editors. Diagnosis and Management of Lameness in the Horse. St. Louis, Missouri: WB Saunders; 2003. pp. 72–79. [Google Scholar]
- 4.Keegan KG. Evidence-based lameness detection and quantification. Vet Clin North Am Equine Pract. 2007;23:403–423. doi: 10.1016/j.cveq.2007.04.008. [DOI] [PubMed] [Google Scholar]
- 5.Keegan KG, Wilson DA, Wilson DJ, et al. Evaluation of mild lameness in horses trotting on a treadmill by clinicians and interns or residents and correlation of their assessments with kinematic gait analysis. Am J Vet Res. 1998;59:1370–1377. [PubMed] [Google Scholar]
- 6.Hewston M, Christley RM, Hunt ID, Voute LC. Investigations of the reliability of observational gait analysis for the assessment of lameness in horses. Vet Rec. 2006;158:852–857. doi: 10.1136/vr.158.25.852. [DOI] [PubMed] [Google Scholar]
- 7.Ishihara A, Reed SM, Rajala-Schultz PJ, Robertson JT, Bertone AL. Use of kinetic gait analysis for detection, quantification, and differentiation of hind limb lameness and spinal ataxia in horses. J Am Vet Med Assoc. 2009;234:644–651. doi: 10.2460/javma.234.5.644. [DOI] [PubMed] [Google Scholar]
- 8.Weishaupt MA, Wiestner T, Hogg HP, Jordan P, Auer JA, Barrey E. Assessment of gait irregularities in the horse: Eye vs. gait analysis. Eq Vet J Suppl. 2001;33:135–140. doi: 10.1111/j.2042-3306.2001.tb05376.x. [DOI] [PubMed] [Google Scholar]
- 9.Peham C, Licka T, Schobesberger H, Meschan E. Influence of the rider on the variability of the equine gait. Hum Mov Sci. 2004;23:663–671. doi: 10.1016/j.humov.2004.10.006. [DOI] [PubMed] [Google Scholar]
- 10.Powers P, Harrison A. Effects of the rider on the linear kinematics of jumping horses. Sports Biomech. 2002;1:135–146. doi: 10.1080/14763140208522792. [DOI] [PubMed] [Google Scholar]
- 11.Ishihara A, Bertone AL, Rajala-Schultz PJ. Association between subjective lameness grade and kinetic gait parameters in horses with experimentally induced forelimb lameness. Am J Vet Res. 2005;66:1805–1815. doi: 10.2460/ajvr.2005.66.1805. [DOI] [PubMed] [Google Scholar]
- 12.Symes D, Ellis R. A preliminary study into rider asymmetry within equitation. Vet J. 2009;181:34–37. doi: 10.1016/j.tvjl.2009.03.016. [DOI] [PubMed] [Google Scholar]
- 13.Blokhuis M, Aronsson A, Hartmann E, Van Reenen C, Keeling L. Assessing the rider’s seat and horse’s behavior: Difficulties and perspectives. J App Anim Welf Sci. 2008;11:191–203. doi: 10.1080/10888700802100876. [DOI] [PubMed] [Google Scholar]
- 14.Schöllhorn W, Peham C, Licka T, Scheidl M. A pattern recognition approach for the quantification of horse and rider interactions. Equine Vet J Suppl. 2006;36:400–405. doi: 10.1111/j.2042-3306.2006.tb05576.x. [DOI] [PubMed] [Google Scholar]