Obesity is an issue in many species and is a leading factor in the development of other diseases (Raffan, 2013; Kipperman and German, 2018). The Blue Cross reported an increase in equine obesity within recent years, rising from 7.8% in 2013, to 16.9% in 2014 and 23.2% in 2015 (Murray et al, 2015). Simultaneous with this increase, a rise in related diseases, including laminitis, can also be observed, suggesting obesity could be a factor in the surge of select diseases (Murray et al, 2015). This study aims to identify which condition assessment provides the most consistent results and the effect owner education has on their ability to condition score. The use of alternative condition assessments to weigh bridges is important in equines because of the limitations faced by owners. Unlike smaller companion animals, horses require large, specialised scales, which can be expensive and hard to frequently access (Bushell and Murray, 2016; Witherow, 2019). The temperament of horses also provides further limitations and many horses may find the appearance of scales scary, subsequently not standing on scales to get an accurate weight (Witherow, 2019).
Method
The study design was based on the observation and collection of data concerning the assessment of a horse's body condition. Ethical approval was gained from the University Ethics Committee, before any data were collected. A total of seven assessors and seven horses participated in the study. The assessors and horses had to meet the inclusion criteria detailed below:
- Assessors:
- A minimum of 6 months' equine experience
- Confident when working around horses
- Basic knowledge of correct etiquette around horses
- Basic relationship with the horses used in the project already established
- Horses:
- No known underlying health conditions
- Not of an extremely nervous disposition.
Data collection
For this study a smaller group of seven horses and seven assessors collected the relevant data. The assessors varied in experience and this mirrors the typical ownership of horses. Four Irish sports horses and three cobs were used in the study, enabling evaluation of how breed might affect the results. Before measurements were taken, standardisation measures were put in place, involving the main researcher reading through the method for each condition assessment technique. Additionally, the main researcher also demonstrated each condition assessment technique, reducing the possibility of an assessor incorrectly carrying out a condition assessment. Further to this, when the data were being collected, the horses were tied up, using a quick release knot, on a flat, level surface, without compromising assessor or horse safety, and to ensure that the data were collected under the same conditions.
Five different methods described in the literature that are used for assessment of body condition, were used for the study, and are described below:
Body condition scoring
For this study, the body condition scoring scale (BCS) used was adapted from Henneke et al (1983), as seen in Figure 1. This BCS scale is a scientific method of assessing a horse's condition and scores from one to nine, rather than one to five (Rendle et al, 2018a). A one to nine scale was selected because of the added depth obtained from a one to nine scale (Roche et al, 2004). In addition to this, Weijters et al (2010) investigated the response rates of people in surveys using scales, finding when people did not feel confident in a question they would opt for the middle of the given scale. Subsequently, while still having a middle number the larger scale reduces the chances of assessors choosing the middle number and encouraging them to look closer at the condition of each horse.
Cresty neck scoring
Cresty neck scoring is a condition score given to horses based on the presence or amount of fat on the top line of a horse's neck, also known as a crest (Carter et al, 2009). When using this form of assessment, it should be considered that some breeds of horses are at a higher predisposition to having a crest, therefore unless breed is considered, this scoring system can lead to bias in results (Dugdale et al, 2011; Giles et al, 2015; Rendle et al, 2018a). Additionally, fat deposits on the neck are known to develop and deplete much slower than localised areas of fat in the rest of the body (Dugdale et al, 2011; Giles et al, 2015). Despite this, cresty neck scoring was still included in the study, to provide part of a more in-depth view of each area of the horse, which a whole-body body condition score may not provide. The breed and sex of each horse was recorded and factored into the cresty neck result analysis, however no stallions or draft horses which are prone to cresty necks, were included within this study.
Heart girth measurement
Heart girth measurements are taken on expiration, with the tape passed directly behind the point of the elbow and just caudal to the end of the withers. The tape should be on a slight diagonal, similar to the method of using a weigh tape. An example of this can be seen in Figure 3. It is a useful method for identifying localised areas of fat deposits and is commonly used in conjunction with belly girth and rump width measurements.
Belly girth measurement
Belly girth measurements, seen in Figure 4, require the tape to be placed on the mid-point region of the horse's back and passed around the widest part of the horse's abdomen, the results can then be read on expiration. This form of condition assessment was included in the study as it is very useful for indicating generalised fat loss, and is therefore an important form of condition assessment when used to assess the short-term effectiveness of weight management programmes. Commonly, it can be paired with heart girth and rump width measurements, to assess changes in a horse's condition over time.
Rump width measurement
Rump width measurements are taken with a tape positioned from the left point of hip to the right point of hip, giving a measurement across the top of a horse's rump, as seen in Figure 5. Similar to heart and belly girth measurements, it can provide an insight into localised areas of fat and is most commonly used in conjunction with heart and belly girth measurements to assess the effectiveness of weight management programmes. However, the breed, sex and muscle tone of the horse being assessed can alter the results significantly, therefore this must be factored into results. Although this is the case, this form of condition assessment was still used in the study to give an insight into localised areas of fat and identify if assessors can accurately take these measurements.
Data analysis
The data collected for this study were related, ordinal data, as a result non-parametric tests were carried out to analyse the data collected. IBM SPSS was used to analyse the data to test for significance.
Results
Body condition scoring
Table 1 shows the frequency of body condition scores for each horse. Horse 4 had the highest variance; two assessors perceived an underconditioned BCS, while one other assessor observed an over conditioned BCS. To identify whether the differences were significant, a Friedman Anova test was used, revealing a p value of 0.012. The p value below the critical significance level, indicates that there is a significant difference between the BCS awarded to each horse. This suggests that each assessor perceived considerably different scores from one another. As a result of this, it is likely BCS are very subjective and vary considerably between horse owners, therefore they are an unreliable form of condition assessment when used in practice.
Table 1. Comparison of body condition scores for main study data
| Horse ID | Mean body condition score given by each assessor | |||||||
|---|---|---|---|---|---|---|---|---|
| Assessor 1 | Assessor 2 | Assessor 3 | Assessor 4 | Assessor 5 | Assessor 6 | Assessor 7 | Standard deviation | |
| Horse 1 | 4 | 4 | 3 | 4 | 4 | 4 | 4 | 0.377 |
| Horse 2 | 6 | 7 | 6 | 6 | 7 | 7 | 6 | 0.534 |
| Horse 3 | 6 | 6 | 6 | 7 | 6 | 7 | 6 | 0.487 |
| Horse 4 | 4 | 5 | 5 | 6 | 5 | 5 | 4 | 0.690 |
| Horse 5 | 7 | 7 | 7 | 8 | 8 | 7 | 7 | 0.487 |
| Horse 6 | 4 | 5 | 5 | 5 | 5 | 5 | 5 | 0.377 |
| Horse 7 | 6 | 6 | 6 | 7 | 6 | 6 | 6 | 0.377 |
| Standard deviation | 1.25 | 1.11 | 1.27 | 1.34 | 1.34 | 1.21 | 1.13 | |
Furthermore, Table 1 also enables comparison of the assessors scores across all horses. The scores given by assessor 1, 3, 4, 5 and 6 had notably different scores for each horse. However, it also indicates the scores given by assessors 2 and 7 were similar across all horses. A Friedman Anova test revealed there was not a significant statistical difference between the BCS that assessors 2 and 7 gave to each horse. Consequently, this suggests the two assessors were unable to differentiate between the conditions of each horse, potentially because of their lack of experience as these two assessors both had under 5 years with horses, or as a result of the differences between the breeds and genders of the horses. Thus, in practice the experience levels of owners should be considered and potentially more work put into owner education.
Cresty neck scoring
Table 2 shows the differences in the cresty neck scores. Scores given to horse 1 were almost unanimous, this is supported by a low standard deviation of 0.3. Dissimilarly, the other horses had higher standard deviation results, because of less consistent scores. Horse 5 had the greatest standard deviation, suggesting this horse had the greatest difference between its highest and lowest scores. As the cresty neck scores were carried out using a one to five scale, the differences between each score on the scale is much more significant than on a one to nine scale, consequently it is likely owner education had an impact in the scores given. Horse 5 was an over conditioned, native breed, which are known for being hard to assess, therefore this could also be a reason why the scores varied.
Table 2. Comparison of cresty neck scores for main study data
| Horse ID | Cresty neck scores given by each assessor | |||||||
|---|---|---|---|---|---|---|---|---|
| Assessor 1 | Assessor 2 | Assessor 3 | Assessor 4 | Assessor 5 | Assessor 6 | Assessor 7 | Standard deviation | |
| Horse 1 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 0.377 |
| Horse 2 | 3 | 2 | 2 | 3 | 4 | 3 | 3 | 0.690 |
| Horse 3 | 3 | 2 | 3 | 3 | 3 | 3 | 4 | 0.577 |
| Horse 4 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 0 |
| Horse 5 | 2 | 2 | 2 | 4 | 4 | 2 | 3 | 0.951 |
| Horse 6 | 1 | 2 | 1 | 1 | 2 | 1 | 2 | 0.534 |
| Horse 7 | 3 | 3 | 3 | 4 | 2 | 2 | 4 | 0.816 |
| Standard deviation | 1.21 | 0.69 | 0.89 | 1.39 | 1.27 | 0.89 | 1.27 | |
Additionally, Table 2 also enables comparison of the assessors' cresty neck scores across the horses. It can be observed that assessors 1, 4, 5 and 7 had a large standard deviation result, suggesting there is a considerable difference between their scores, and they therefore utilised the whole scale. Conversely, Assessors 2, 3 and 6 had lower standard deviation results, suggesting their scores were similar across all the horses, potentially because of the assessors not utilising the whole scale as a result of being unfamiliar with cresty neck scores. A Friedman Anova test was carried out, the test revealed a p value of 0.065. This p value is slightly higher than the critical significance level, and as a result, it can be perceived there is no significant difference across the scores each assessor gave to each of the horses. Suggesting, although the scores varied between assessors, the differences were not statistically significant and would not have had a huge impact on the condition assessments of the horses.
Belly girth
Table 3 shows differences in the belly girth measurements. It can be seen that horse 1 has a large gap between measurements. Whereas, all others only have slight variations. A Friedman Anova test of difference was carried out on the measurements of each of the horses. Giving a p value of 0.17, this value is significantly higher than the critical significance level. Consequently, there is not a significant difference between the belly girth measurements from each assessor on any of the horses. This indicates that although there were slight differences in the measurements by each assessor, the differences would not have affected the condition assessments significantly. As a result of this, it could be interpreted that belly girth measurements are a more consistent form of condition assessment, less affected by owner education than BCS and cresty neck scores. However, to form a full body condition assessment, belly girth measurements would need to be used in conjunction with heart girth and rump width measurements.
Table 3. Comparison of belly girth measurements
| Horse ID | Belly girth measurements From each assessor (kg) | ||||||
|---|---|---|---|---|---|---|---|
| Assessor 1 | Assessor 2 | Assessor 3 | Assessor 4 | Assessor 5 | Assessor 6 | Assessor 7 | |
| Horse 1 | 562 | 585 | 577 | 582 | 582 | 577 | 577 |
| Horse 2 | 673 | 673 | 681 | 681 | 673 | 681 | 673 |
| Horse 3 | 784 | 793 | 784 | 784 | 784 | 793 | 784 |
| Horse 4 | 539 | 532 | 532 | 532 | 539 | 539 | 532 |
| Horse 5 | 723 | 681 | 723 | 681 | 723 | 723 | 723 |
| Horse 6 | 593 | 601 | 601 | 593 | 593 | 601 | 593 |
| Horse 7 | 681 | 681 | 689 | 673 | 689 | 681 | 681 |
Rump width
Table 4 shows the rump width measurements of each horse. When comparing the measurements taken from each individual horse, little variation is observed. Looking at Table 4, it could be perceived that there is little difference in the width of the horses' rumps, despite various breeds being assessed. In order to identify if there was any statistical significance between the measurements, a Friedman Anova was carried out with a resulting p value of 0.531, which is significantly higher than the critical significance level. As a result, there is no significant difference between the rump width measurements from each assessor on any of the horses. Suggesting, owners find rump width measurements easier to learn and use than BCS or cresty neck scores. Moving forward in practice, rump width, heart girth and belly girth could be used to encourage a more consistent, less subjective form of condition assessments used by owners.
Table 4. Comparison of rump width scores for main study data
| Horse ID | Rump width measurements from each assessor (inches) | ||||||
|---|---|---|---|---|---|---|---|
| Assessor 1 | Assessor 2 | Assessor 3 | Assessor 4 | Assessor 5 | Assessor 6 | Assessor 7 | |
| Horse 1 | 25 | 24 | 24 | 25 | 24 | 24 | 25 |
| Horse 2 | 21 | 21 | 22 | 22 | 22 | 21 | 22 |
| Horse 3 | 22 | 22 | 22 | 21 | 22 | 21 | 22 |
| Horse 4 | 20 | 19 | 20 | 20 | 20 | 20 | 20 |
| Horse 5 | 22 | 21 | 22 | 22 | 22 | 22 | 22 |
| Horse 6 | 24 | 24 | 24 | 24 | 24 | 23 | 24 |
| Horse 7 | 21 | 21 | 20 | 20 | 21 | 21 | 21 |
Discussion
The study revealed significant variation in the BCS and cresty neck scores of all the horses assessed. Looking at the results, BCS and cresty neck scoring did not provide consistent, similar results between all the assessors as would be expected from a good condition scoring assessment. This suggests that alternative condition scoring methods need to be formulated and implemented in practice, if the ability of owners to assess their horse's condition is going to improve.
Evaluation of the cresty neck scores and neck region of the whole-body BCS, shows the neck region of a horse is one of the areas the assessors in this study struggled with the most. We already know the neck of a horse can be a difficult area to assess because of anatomical variations between breeds and genders of horses, as well as the potential for underlying conditions like insulin dysregulation to affect the neck (Frank et al, 2006; Silva et al, 2016; Sánchez et al, 2017). The results from this study further support this and suggest that although good for identifying localised areas of fat deposits, cresty neck scores is not an accurate form of condition assessment.
Further to this, the BCS awarded to each horse from the assessors showed substantial differences. From this we can see the education and previous experience of the assessors had an impact on their ability to accurately BCS. Tables 1 and 2 highlight the standard deviation of each assessor's BCS across all the horses, in the hope of identifying if the assessors were able to BCS accurately. In all tables, assessor 4 had the highest standard deviation results, meaning their scores had the most variation, suggesting they observed the greatest difference between each horse. Whereas, in both tables, assessor 2 has the lowest standard deviation results, suggesting they noted little difference between each horse. Assessor 2 consistently scored using a five, six or seven, indicating they leaned towards the higher end of the BCS scale, conflicting current literature. Potter et al (2016) investigated owners' perceptions of BCS in comparison to researchers' assessments, revealing owners' perceived body conditions to be significantly lower than researchers (Potter et al, 2016). It is likely that a lack of education or experience prevented assessor 2 from utilising the full BCS scale, instead leading them to favour the middle of the scale. Weijters et al (2010) investigated the response rates of people in surveys using scales, finding when people did not feel confident in a question they would opt for the middle of the given scale, which could explain the results from assessor 2.
In this study, the belly girth measurements also showed slight variation. In all of the results each horse has a mean difference of 10 kg between the highest and lowest scores. Excluding horse 1 and 7, all results appear to vary between two measurements. This could have been because of the design of the tape used; the numbers printed on the tape increased in intervals of ten, difficulties in interpreting the numbers could explain the differences in the results. The National Equine Health Survey, 2015 found 64% of horse owners use a weigh tape to monitor their horse's condition, compared with only 51% in 2014 (Murray et al, 2015). The rise in the use of weigh tapes justifies further research into accurate measurements.
The results from the rump width measurements show little variation as the largest gap was only one inch. A difference of 1 inch could be because of an error when reading the tape measure, especially in the larger horses as there was no stool available because of health and safety concerns. The consistent results suggest rump width was the most accurate form of assessment, yet rump widths provide little information into the whole-body condition of a horse. Instead a rump width should be compared and contrasted against the scores from BCS and cresty neck score to identify any localised adiposity (Witherow, 2019). Furthermore, the rump width of a horse is likely to vary with workload as a result of muscle development, in addition to the breed of a horse (Kashiwamura et al, 2001; Gómez et al, 2012). Heavy breeds of horse will have a much wider rump width than a finer breed (Kashiwamura et al, 2001; Gómez et al, 2012).
The data collected from this study suggest that heart girth, belly girth and rump width measurements all provided a more consistent and accurate form of condition assessment in comparison to BCS and cresty neck scores. Feedback from the assessors throughout the study also suggested these three forms of condition assessment were the easiest to learn and use. While these three measurements cannot be used individually to form an insight into a horse's overall whole-body condition, through combining them it is possible to gain an insight into a horse's overall condition as well as any localised fat deposits.
Limitations of the study
This study had some limitations because of the number of participants used. Previous, similar studies used much larger sample sizes of 200 to 300 participants, yet these numbers of participants would not have been possible because of the nature of this study, even though an increase in participants would help to increase reliability of the results. Further studies carried out could benefit from an increased number of participants, in order to promote more reliable results and give a better view of the UK's equine obesity issue, rather than just one area of the UK as in this study.
Another limitation was the use of condition assessment scores. Ideally a heart girth along with body length would have been taken, this would have enabled a calculation to work out an estimated weight (heart girth x heart girth x body length / 330). The bodyweight calculation could have been compared again the BCS and belly girth results, providing results that address the main research question better. Studies carried out in the future could use this study as a basis to build on, utilising the additional condition assessment techniques.
Conclusion
This study was able to identify a statistical significance between the BCS and cresty neck scores of each horse. Subsequently, it could be assumed the BCS and cresty neck scores perceived by each assessor were not consistent, therefore belly girth and rump width measurements provide a more consistent form of condition assessment. Nevertheless, the results also highlighted a lack of statistical significance between the scores assessors gave to each of the horses. Furthermore, belly girth and rump width measurements only provide an insight into particular areas of a horse's body with high or low quantities of adipose tissue. Consequently, further studies are required to fully elucidate whether belly girth, heart girth and rump width measurements are able to replace BCS as a commonly used condition assessment technique.
KEY POINTS
- Obesity is now a common issue in leisure horses in the UK.
- Body condition scoring is the most common method used to assess whether a horse is in optimal condition.
- There is significant variation among owners when measuring the body condition score of their horses, despite using a recommended body condition scoring system.
- Other methods such as rump width measurements and belly and heart girth measurements should be used alongside body condition scoring to provide a more accurate and objective measurement of body condition.
- Further research is needed with a larger sample size.