Equine Discomfort Ethogram

Simple Summary Pain and discomfort behavior in horses tends to be especially subtle, and not readily or widely appreciated even by equine professionals, including many long-time horse keepers, trainers, and even by veterinarians, veterinary technicians, and care staff. Based on decades of evaluating the behavior of normal and physically uncomfortable horses in a referral hospital, as well as research context, we describe and illustrate a catalog of behaviors (ethogram) associated with equine physical discomfort. Our objective is to promote an unambiguous universal understanding of equine discomfort behaviors associated with various body systems and anatomic sources. Abstract In recent years, there has been a growing interest in and need for a comprehensive ethogram of discomfort behavior of horses, particularly for use in recognizing physical discomfort in domestically managed horses. A clear understanding of the physical discomfort behavior of horses among caretakers, trainers, and professional health care personnel is important to animal welfare and caretaker safety. This is particularly relevant to pain management for hospitalized equine patients. Various pain scale rubrics have been published, typically incorporating only a few classically cited pain behaviors that, in many cases, are specific to a particular body system, anatomic location, or disease condition. A consistent challenge in using these rubrics in practice, and especially in research, is difficulty interpreting behaviors listed in various rubrics. The objective of this equine discomfort ethogram is to describe a relatively comprehensive catalog of behaviors associated with discomfort of various degrees and sources, with the goal of improving understanding and clarity of communication regarding equine discomfort and pain. An inventory of discomfort-related behaviors observed in horses has been compiled over 35 years of equine behavior research and clinical consulting to medical and surgical services at the University of Pennsylvania School of Veterinary Medicine’s equine hospital. This research and clinical work included systematic evaluation of thousands of hours of video-recordings, including many hundreds of normal, healthy horses, as well as hospitalized patients with various complaints and/or known medical, neurologic, or orthopedic conditions. Each of 73 ethogram entries is named, defined, and accompanied by a line drawing illustration. Links to online video recorded examples are provided, illustrating each behavior in one or more hospitalized equine patients. This ethogram, unambiguously describing equine discomfort behaviors, should advance welfare of horses by improving recognition of physical discomfort, whether for pain management of hospitalized horses or in routine husbandry.


Introduction
The recognition of physical discomfort in horses is important to ensuring adequate welfare, both for general husbandry and for veterinary care procedures. Traditional veterinary evaluation of pain in equine patients has relied heavily on objective physical measures (e.g., heart rate, respiratory rate), and to a lesser extent, endocrine measures (e.g., circulating levels of cortisol, ß-endorphins, catecholamines, or pro-inflammatory mediators),

Original Observations
The information presented in this ethogram was primarily collected by the senior author (SM) over 35 years of equine behavior research and clinical consulting at the University of Pennsylvania School of Veterinary Medicine New Bolton Center. This work included systematic evaluation of thousands of individual horses, each with a minimum of 24 h of continuous video recording. These included normal, healthy horses in research studies or in breeding programs, as well as hospitalized patients with various presenting complaints and/or known medical, neurologic, or orthopedic conditions. Patient behavior evaluations were performed primarily to assist hospital clinicians or referring veterinarians with their diagnostic process. Typically, this involved advising whether a change in behavior or performance appeared to more likely to reflect a psychological/behavioral problem and/or current physical discomfort [11]. In most cases, a source of physical discomfort was eventually diagnosed, providing feedback that enhanced the evaluator's body of knowledge. During this work, an inventory of discomfort-related behaviors associated with various body systems and anatomic sources in horses was compiled. This work also often included evaluation of behavior before and after administration of analgesia. In these cases, if a behavior was suspected to reflect pain, and if it diminished with the administration of analgesia, the behavior could more confidently be judged to reflect pain. At times, clinicians also requested the author's (SM) professional opinion concerning the level of discomfort or quality of life in patients with known painful conditions. These consults provided further information concerning types of discomfort behavior associated with various body systems and anatomic sources.
The resulting list of equine discomfort behaviors in this ethogram includes a total of 64 specific discomfort behaviors grouped into eight categories: posture and weightbearing; limb and body movements; head, neck, mouth, and lip movements; attention to area; ear and tail movements; overall demeanor; altered eating or drinking; and vocalizations/audible sounds. Some behaviors have slight variations in form (e.g., stretching), resulting in a total of 73 entries (see . Each entry includes a name we believe to be commonly used in English, a word definition, a line drawing depiction, and (with one exception, sipping water) a link to a supplemental video file depicting one or more examples of equine patients displaying the behavior. Many, but not all, of these behaviors, have been mentioned in the literature that we reviewed. There were no additional behaviors described in the literature that we did not record in our clinical observations. In addition, Table S1 summarizes behaviors commonly associated with various anatomic sources of discomfort.

Comments
We propose that this ethogram provides a relatively complete catalog of behaviors that can be used as a reference to recognize discomfort in horses, both for general husbandry and for clinical veterinary assessments. Hopefully, it will also prove useful for research and for future pain scale development or refinement.
Recognition of discomfort in a prey species is particularly challenging. Horses have evolved to show little evidence of discomfort or disability in the presence of predators, including humans. This obviously can confound discomfort assessment [36]. In our clinical review of 24-h continuous video, the effect of this phenomenon has been conspicuous (11). In their pain assessments of horses following arthroscopic surgery, Price et al. [30] commented that patients in their study showed a reduced incidence of certain pain behaviors Animals 2021, 11, 580 4 of 21 (e.g., restlessness, weight shifting) when observed directly, even when caretakers viewed from a distance, compared to remotely by video. Recently, we quantitatively evaluated this important effect in 20 hospitalized equine orthopedic patients [7]. In those patients, ongoing discomfort behaviors diminished or stopped altogether during caretaker visits for pain assessment, and resumed once the caretakers had departed. Those patients displayed a mean reduction of over 75% in the number of discomfort behaviors per minute when caretakers were present, and 30% of those patients stopped performing discomfort behaviors altogether during the caretaker visit. Due to this effect, composite pain scoring systems that require a period of direct observation likely underestimate discomfort behaviors. We, therefore, posit that, when assessing discomfort in horses, it is important to observe remotely. Regardless of how well-trained a caretaker may be in behavior observation, if discomfort behavior is interrupted by their presence, information regarding the horse's condition is lost.
Brief periods of direct observation may also lead to misinterpretations of comfort. For example, stalled horses typically go through cycles of foraging and resting, often always standing in one particular area of the stall when resting. If not observing longer periods of continuous video, it may appear that a horse rarely moves from one area, when they are actually going through normal resting/foraging cycles, but always returning to the same area for standing rest. Similarly, if a caretaker happened to repeatedly visit during rest periods, they may erroneously conclude that the horse has a decreased appetite. For this reason, observing for longer periods of time via video recording, which can be viewed in fast forward, is invaluable in understanding the overall behavior pattern of a horse.
Behaviors that suddenly interrupt ongoing goal-directed behavior, such as foraging or resting, when the horse is otherwise calm, more clearly appear to represent acute discomfort. In our experience, deviation from, or apparent inability to perform, normal sequences of foraging and resting behavior, has been associated with increased levels of discomfort.
Our experience with video observation of such a large number of clinical cases, many with similar sources and types of diagnosed physical discomfort, has allowed us to appreciate that there appear to be individual differences among horses in the expression and the particular combination of various discomfort behaviors for any particular diagnosed condition. Ijichi et al. [46] recently explored the association of pain expression and corresponding severity of musculoskeletal lesion in horses with owner-assessed personality factors. An important finding in that work was that the degree of clinical lameness was not a reliable indicator of the severity of tissue damage as diagnosed on ultrasound or radiographic imaging. The conclusion of their preliminary study was that individual variation in the expression of pain may be associated with certain personality factors.
Many of the discomfort behaviors included in this ethogram can be expressed with slight variation in form. For example, as described in Table 2e, kicking out or back can be performed in several distinct forms, varying in height and degree of extension of the limb or limbs. In this ethogram, our illustration depicts just one common form. We expect that variations will be easily recognizable as that behavior. Experience observing horses over time will likely increase the observer's knowledge of the possible variations of particular behaviors.
We would not consider a single occurrence of any one behavior to be conclusive evidence of discomfort. When viewing video of a horse to assess comfort level, in most instances, we consider the first occurrence of a potential discomfort behavior as an indicator to continue watching for repetitions of that particular behavior. Before making a judgment about the causes of a behavior, or what a specific behavior might indicate regarding discomfort, it is important to be sure that it was not an isolated event with an alternate explanation. Repetitive clusters of behaviors, including one or more non-specific discomfort behaviors (e.g., moving/focusing ears caudally, swishing/flicking tail, rotational shaking head or whole body, restlessness/ill-at-ease), are also helpful in confidently judging that a behavior reflects physical discomfort. For example, if one or more episodes of stomping are observed without any other indicators of discomfort, one might be more suspicious of superficial irritation, such as a biting insect. In our experience, additional specific indicators of limb pain (e.g., non-physiologic locomotion, shifting weight/resting limb, pointing) or multiple nonspecific indicators of discomfort, typically increase confidence that these behaviors reflect pain. Similarly, particular combinations of temporally-associated behaviors often provide insight as to the anatomic source of discomfort. (see supplementary Table S1: Discomfort Behaviors Commonly Associated with Various Anatomic Sources https://doi.org/10.528 1/zenodo.4537909 (accessed on 18 February 2021)).
This ethogram is based on observations of stalled horses at rest. It does not necessarily address discomfort behaviors that may be observed in horses during work. For a recently published pain ethogram for horses while being ridden, see Dyson et al. 2018 [47]. Further, and importantly, our ethogram is not meant to be independently diagnostic, but rather to provide additional detailed information to veterinary professionals. Should routine caretakers or health care professionals observe behavior suggesting discomfort, further veterinary diagnostics are indicated.
Finally, within the context of various horse behavior courses and summer research student training, we have had the opportunity to introduce approximately 50 graduate students, veterinary students, veterinarians, horse owners, and trainers to this method of evaluating discomfort in horses using video recorded behavior of horses in stalls. Our informal measures of training time suggest the rapid acquisition of the skill. Following a review of earlier versions of this ethogram, reliable recognition of these behaviors has been almost immediate. The typical time for a new observer to reach an acceptable inter-observer agreement with highly experienced observers has been less than 3 h of experience with the technique (unpublished observations).

Conclusions
This ethogram, unambiguously describing equine discomfort behaviors, promises to advance welfare of horses by improving recognition of physical discomfort, whether for pain management of hospitalized horses or in everyday husbandry. whole body, restlessness/ill-at-ease), are also helpful in confidently judging that a behavior reflects physical discomfort. For example, if one or more episodes of stomping are observed without any other indicators of discomfort, one might be more suspicious of superficial irritation, such as a biting insect. In our experience, additional specific indicators of limb pain (e.g., non-physiologic locomotion, shifting weight/resting limb, pointing) or multiple non-specific indicators of discomfort, typically increase confidence that these behaviors reflect pain. Similarly, particular combinations of temporally-associated behaviors often provide insight as to the anatomic source of discomfort. (see supplementary Table S1: Discomfort Behaviors Commonly Associated with Various Anatomic Sources https://doi.org/10.5281/zenodo.4537909 (accessed on 18 February 2021)). This ethogram is based on observations of stalled horses at rest. It does not necessarily address discomfort behaviors that may be observed in horses during work. For a recently published pain ethogram for horses while being ridden, see Dyson et al. 2018 [47]. Further, and importantly, our ethogram is not meant to be independently diagnostic, but rather to provide additional detailed information to veterinary professionals. Should routine caretakers or health care professionals observe behavior suggesting discomfort, further veterinary diagnostics are indicated.
Finally, within the context of various horse behavior courses and summer research student training, we have had the opportunity to introduce approximately 50 graduate students, veterinary students, veterinarians, horse owners, and trainers to this method of evaluating discomfort in horses using video recorded behavior of horses in stalls. Our informal measures of training time suggest the rapid acquisition of the skill. Following a review of earlier versions of this ethogram, reliable recognition of these behaviors has been almost immediate. The typical time for a new observer to reach an acceptable interobserver agreement with highly experienced observers has been less than 3 h of experience with the technique (unpublished observations).

Conclusions
This ethogram, unambiguously describing equine discomfort behaviors, promises to advance welfare of horses by improving recognition of physical discomfort, whether for pain management of hospitalized horses or in everyday husbandry. whole body, restlessness/ill-at-ease), are also helpful in confidently judging that a behavior reflects physical discomfort. For example, if one or more episodes of stomping are observed without any other indicators of discomfort, one might be more suspicious of superficial irritation, such as a biting insect. In our experience, additional specific indicators of limb pain (e.g., non-physiologic locomotion, shifting weight/resting limb, pointing) or multiple non-specific indicators of discomfort, typically increase confidence that these behaviors reflect pain. Similarly, particular combinations of temporally-associated behaviors often provide insight as to the anatomic source of discomfort. (see supplementary Table S1: Discomfort Behaviors Commonly Associated with Various Anatomic Sources https://doi.org/10.5281/zenodo.4537909 (accessed on 18 February 2021)). This ethogram is based on observations of stalled horses at rest. It does not necessarily address discomfort behaviors that may be observed in horses during work. For a recently published pain ethogram for horses while being ridden, see Dyson et al. 2018 [47]. Further, and importantly, our ethogram is not meant to be independently diagnostic, but rather to provide additional detailed information to veterinary professionals. Should routine caretakers or health care professionals observe behavior suggesting discomfort, further veterinary diagnostics are indicated.
Finally, within the context of various horse behavior courses and summer research student training, we have had the opportunity to introduce approximately 50 graduate students, veterinary students, veterinarians, horse owners, and trainers to this method of evaluating discomfort in horses using video recorded behavior of horses in stalls. Our informal measures of training time suggest the rapid acquisition of the skill. Following a review of earlier versions of this ethogram, reliable recognition of these behaviors has been almost immediate. The typical time for a new observer to reach an acceptable interobserver agreement with highly experienced observers has been less than 3 h of experience with the technique (unpublished observations).

Conclusions
This ethogram, unambiguously describing equine discomfort behaviors, promises to advance welfare of horses by improving recognition of physical discomfort, whether for pain management of hospitalized horses or in everyday husbandry.

k. Tucked Up Abdomen
Tensing of the abdominal muscles, hollowing out the flank, sometimes with a hunched back. Reduced ingesta and dehydration may contribute to the hollow appearance (also known as "drawn up" or "sucked up"

Stepping in Place
Repeated flexing of a limb, briefly relieving weight-bearing on that limb. The toe may touch or be held slightly above the substrate for up to several seconds, before the hoof is lightly placed down again. Stepping often occurs in a series of several rhythmic steps at about 1 s intervals, with a pause between series.

Kicking Up Toward Abdomen
Flexing a hindlimb, directing the hoof or stifle toward the abdomen, often arcing laterally as the hoof returns toward the substrate, often similar to that seen in response to cutaneous irritation, e.g., insects.

c. Autogrooming
Nibbling, nuzzling, and/or biting at an area of the body, or rubbing one part of the body to another or against an object, similar to that seen in response to cutaneous irritation, e.g., insects. Video S48 https://doi.org/10.5281/zenodo.4541398 (accessed on 18 February 2021) Nibbling, nuzzling, and/or biting at an area of the body, or rubbing one part of the body to another or against an object, similar to that seen in response to cutaneous irritation, e.g., insects. Video S48 https://doi.org/10.5281/zenodo.4541398 (accessed on 18 February 2021)

a. Sighing
Emitting an audible long exhalation following a deep inhalation.

b. Snorting
Emitting an audible sudden forced exhalation through the nares.
c. Whining Emitting a long, high-pitched vocalization.
g. Screaming/Calling Emitting a long loud whinny vocalization, typical of calling to locate distant herd mates.

h. Teeth Grinding
Moving tightly clenched jaws back and forth and grating the upper and lower molars, resulting in a crunching, scraping sound. Institutional Review Board Statement: Procedures for obtaining video recordings of client-owned hospitalized equine patients during the 2017 pain assessment study were approved by the University of Pennsylvania Institutional Animal Care and Use Committee according to all applicable guidelines for humane care and use of animals in research (protocol #806321). Owners of these horses also provided informed consent for this clinical research.