Journal of Veterinary Behavior 17 (2017) 24e31
Contents lists available at ScienceDirect
Journal of Veterinary Behavior
journal homepage: www.journalvetbehavior.com
Canine Research
Assessing stress in dogs during a visit to the veterinary clinic:
Correlations between dog behavior in standardized tests and
assessments by veterinary staff and owners
Ann-Kristina Lind a, Eva Hydbring-Sandberg b, Björn Forkman c, Linda J. Keeling a, *
a
b
c
Department of Animal Environment and Health, Swedish University of Agricultural Sciences, Uppsala, Sweden
Department of Anatomy, Physiology and Biochemistry, Swedish University of Agricultural Sciences, Uppsala, Sweden
Department of Large Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 December 2014
Received in revised form
7 October 2016
Accepted 7 October 2016
Available online 15 October 2016
A visit to a veterinary clinic can be very stressful for the dog, and stress may interact with pain. The aim of
this study was to observe the behavior of dogs in a veterinary clinic and to correlate it with subjective
stress assessments by different persons. Systems have already been developed to assess pain in dogs. We
tested the behavior of 105 dogs, sampled from 233 dog owners who completed our questionnaire and
whose dogs were patients at the clinic. The dog owner, the test leader, the nurse, and the veterinarian
were each asked to assess if the dog was experiencing pain or was stressed and to evaluate, overall, how
the dog experienced the visit. Three behavior tests were also carried out to describe the dog’s reaction in
the veterinary clinic: a “social contact” test, a “play” test, and a “treat” test. The play and treat tests were
carried out both inside and outside the veterinary clinic to see if the dogs reacted differently in the 2
situations. Agreement between observers was good to excellent but generally better when assessing pain
than stress. Dogs rated as more stressed were significantly less likely to engage in social contact with an
unfamiliar person (P < 0.0001). They were significantly more willing to play and eat a treat outside the
veterinary clinic compared to inside the clinic (P < 0.001), implying that the dogs themselves experienced inside the clinic as being more negative. The results also indicated that the type of relationship the
owner has with the dog may influence the dog’s behavior during the clinical examination. The good
agreements between the different measures in this study suggest that there is potential for a system to
be developed to assess the extent to which the dog is stressed in the clinic.
Ó 2016 Elsevier Inc. All rights reserved.
Keywords:
animal hospital
animal welfare
behavior tests
canine
ethology
questionnaire
Introduction
There have been several studies on the assessment of welfare of
dogs in society, particularly in shelters (Hennessy et al., 2002; Sales
et al., 1997; Weiss, 2002) but only a few studies on the assessment
of the welfare of dogs in veterinary clinics (Mariti et al., 2015;
Stanford, 1981; Väisänen et al., 2005). A visit to a clinic may be
stressful for the patient for several reasons. The most obvious is
because the dog is in pain associated with an injury or disease. Pain
may be defined as an unpleasant sensory and emotional experience
* Address for reprint requests and correspondence: Linda J. Keeling, Department
of Animal Environment and Health, Swedish University of Agricultural Sciences, P.O.
Box 7068, 75007 Uppsala, Sweden, Tel: þ46 18 67 1622; Fax: þ46 18 67 3588.
E-mail address: [email protected] (L.J. Keeling).
http://dx.doi.org/10.1016/j.jveb.2016.10.003
1558-7878/Ó 2016 Elsevier Inc. All rights reserved.
associated with actual or potential tissue damage or described in
terms of such damage (Merskey & Bogduk, 1994). In addition, most
animal species both prey and predators commonly try to mask their
pain as a protective mechanism, and it may be difficult to detect
pain at a clinic (Quimby et al., 2011; Scotney, 2010; Sharkey, 2013).
Although there are individual differences, pain and stress usually
generate similar physiological responses (Hellyer et al., 2007) and
are therefore difficult to separate. Visits to a veterinary clinic are
also for routine checks, vaccination, or an operation such as
castration, on an otherwise healthy dog. They can, however, still be
stressful for the dog. Within the veterinary field, studies have
shown that animals can respond to “the white coat effect” (Belew
et al., 1999; Kallet et al., 1997; Quimby et al., 2011), and dogs may
react based on earlier negative experiences in a clinic. Dogs may
even react to other dogs and species in the waiting room (Rodan
et al., 2011; Scotney, 2010; Siracusa et al., 2007) which may be
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
stressful if the dog is scared of, aggressive toward, or curious about
these other animal species. The dogeowner relationship may affect
the dog’s reactions during a visit to the veterinary clinic, especially
if the owner is stressed because of their own previous experiences
of veterinary clinics or anxious about what will happen to their dog
in the current visit (Merola et al., 2012). Finally, isolation and separation from the owner may affect the dog negatively (Siracusa
et al., 2008). Thus, there are many factors, past and present, that
may be contributing to the overall experience of the dog during a
particular visit to the veterinary clinic.
Stress in animals is sometimes defined as when an animal is
suffering a threat to its homeostasis (Moberg, 2000), but with such
a nonspecific response and variety of ways that an individual animal may try to cope (Koolhaas et al., 2011), it is not obvious how
much agreement there should be between people when scoring
stress. The advantage for an owner when assessing the level of
stress in their dog is that they are very familiar with the dog’s
normal behavior and more likely to see deviations. A veterinarian
on the other hand is trained to recognize symptoms of stress but
probably only sees the dog for a limited period, most likely in the
examination room. Thus, the question arises of who would be the
best person to score the dog and whether any of these subjective
scorings would correlate with the behavior of the dogs in standardized tests designed to assess its reactions under pressure.
Rating, as opposed to behavioral coding, is often thought to
reflect more global states and so may be a fruitful way forward to
validate tests in this context (van den Berg et al., 2006; Wilson &
Sinn, 2012; for a longer discussion see Jones & Gosling, 2005;
Meagher, 2009). More specifically, rating could help identify
which of our noninvasive tests should be validated against accepted
behavioral and/or physiological measures of stress in a future study.
Nowadays, most veterinary clinics aim to make the experience
of the visit as positive for the dogs as possible, for example, by offering carefully prepared waiting and examination rooms, not
mixing different species in the waiting rooms, and offering the
patients different types of rewards (Rodan et al., 2011; Siracusa
et al., 2008; Väisänen et al., 2005).
Pain recognition in animals has advanced considerably in the
past decade (Mathews et al., 2014; Viñuela-Fernández et al., 2007;
Viñuela-Fernández et al., 2011). For the assessment of pain, scoring
systems have been developed for use in the veterinary clinic
(Merola and Mills, 2015; Reid et al., 2007; Sharkey, 2013; van Loon
et al., 2010). When measuring stress, it is usually suggested that a
combination of behavioral and physiological measures is the most
reliable approach (e.g. Beerda et al., 1999; Rooney et al., 2007).
However, taking the physiological measures may in turn influence
the dog’s experience of the clinic, and there is a delay before the
results of the analysis are available. There is a clear need for more
research in this area to evaluate the potential of developing a
feasible, reliable scoring system for assessing stress in veterinary
clinics.
There are already several tests available to test the reactions of
dogs in different situations. For example, the “Swedish Working
Dog Association” uses the behavior of dogs in standardized tests as
an indicator of whether they have an appropriate personality to
work in different services, such as rescue, guide dogs, and so forth.
The test is nowadays called the “Dog Mentality Assessment” and
consists of 10 separate subtests, which are performed outdoors in a
specific order (Svartberg and Forkman, 2002). Over the years, the
different subtests have been evaluated and are now considered
sufficiently validated that they are often used by researchers in
other situations (Strandberg et al., 2005; Svartberg, 2002; Svartberg
et al., 2005). Other studies directed specifically at dogs in a clinic
have shown that dogs in pain are less willing to eat (Downing,
2011). One might expect the same also in dogs that are stressed.
25
It is known that the support of the owner can affect the behavior
of the dog, especially in a stressful situation (Rehn et al., 2014). The
relationship can be assessed by Monash DogeOwner Relationship
Scale (MDORS; Dwyer et al., 2006).
This study investigates how dogs experience a visit to the veterinary clinic using a series of standardized tests and assessments
by a veterinarian, a veterinary nurse, a test leader, and the owner.
The study also addresses whether the type of relationship the
owner has with the dog influences how the dog experiences the
visit. The aim is to evaluate the interobserver reliability and see how
well these are correlated with standardized behavior tests. This is
done with the intention of developing a simple stress assessment
protocol in the future, similar to that developed for pain. The future
protocol should be applicable for all different breeds and ages and
as much as possible should work for both healthy and sick dogs. The
selected behavioral tests have to be easy to do and robust enough to
yield valid results under a number of different circumstances. The
benefit of such a stress assessment protocol, once developed and
validated, would be to help veterinary clinics better adapt the way
they receive the dog, thus contributing to dog welfare and handler
safety.
Material and methods
Data collection and study population
The target population was privately owned dog patients at the
Animal Hospital at the Swedish University of Agricultural Sciences.
At the time of the study, the clinic had on average 700-900 dog
patients entering the clinic per month. Questionnaires were placed
at the front desk of the clinic so dog owners could choose themselves whether to participate in the study. Information about the
study was given on posters in the waiting room. The data were
collected in the period from February 2005 to the end of July 2005.
Before filling in the questionnaire or testing the dog, the owner was
asked to read a brief paragraph describing both the background and
the purpose of the study as well as what would happen to the dog
and then to sign an owner approval form.
A total of 233 owners completed the questionnaire. The questionnaire was always available, but the behavior tests were carried
out on certain days, and the 105 dogs that took part in the behavior
tests were randomly selected. In addition to performing the
behaviour tests, the test leader also made pain and stress assessments on these 105 dogs. The veterinarian completed a total of 89
questionnaires, of which 84 were of dogs that had taken part in the
behavior tests, and the veterinary nurse completed a total of 117, of
which 68 dogs had been tested. The Swedish Board of Agriculture
approved the use of nondestination bred dogs, and the Animal
Ethics Committee in Uppsala, Sweden, approved the experimental
procedure.
Questionnaire design
To access information on how the dogs experienced the visit to
the veterinary clinic, 4 different questionnaires were developed,
and for the evaluation questions, a Likert scale from 1 (not at all) to
10 (very much) was used. The first questionnaire was given to the
owner and addressed general information about the dog (such as
sex, age, and breed), other questions about the dog (such as
whether the dog had any behavior problems), and the reason for
visiting the veterinary clinic. The owner was also asked to scale how
the dog was experiencing the visit. The second questionnaire was
filled in by the person receiving the dog, usually a veterinary nurse
(although this could sometimes be a veterinary student). The third
questionnaire was filled in by the veterinarian, who was also asked
26
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
Table 1
Description of the test setup
Order
Data collection
Reporter/assessor
Location
1
2
2
3
4
5
General questionnaire, MDORS pain/stress scoring
Pain/stress scoring, social contact test
Pain/stress scoring
Calming effect of the owner
Play (n ¼ 76) and treat tests (n ¼ 70)
Play (n ¼ 74) and treat tests (n ¼ 68)
Owner (n ¼ 233)
Test leader (n ¼ 105)
Veterinarian (n ¼ 89) and nurse (n ¼ 117)
Veterinarian (n ¼ 87a)
Test leader
Test leader
Waiting room inside the clinic
Waiting room inside the clinic
Examination room
Examination room
Waiting room inside the clinic
Outside the clinic
a
On 2 occasions, the veterinarian commented that the owner was not in the room at the time of the examination.
to rate if the owner had a positive or negative calming effect on the
dog during their examination. For comparison, the owner, veterinary nurse, and veterinarian were all asked to answer the questions
“To what extent do you feel that the dog is stressed?” and “To what
extent do you feel that the dog is in pain?” using the same Likert
scale. The test leader also answered these 2 questions related to the
dog’s level of stress and pain. The exact wording of the questionnaires given to the different categories of people is available on
request. To obtain data on the dogeowner relationship, the MDORS
questionnaire developed by Dwyer et al. (2006) was used.
The questionnaires were completed separately by the owner and
test leader before the veterinary consulting, so that these answers
were not affected by the veterinarian’s examination and the result
of it. The veterinarian and nurse completed their questionnaires
after the clinical examination of the dog, meaning that the nurse did
the preexamination and completed the questionnaire before the
veterinarian (Table 1). The procedure was done to maximize that
each person was blinded to the answers of the other persons.
stranger which is classified according to the standardized scale
from 1 to 5. The test was carried out where the patient was seated in
the waiting room before the veterinary consultation and after the
owner had filled in the questionnaire and owner approval form, to
insure that the dog owners’ answers were not affected by the
behavioral tests and veterinarian consultation. The test leader had
also assessed the dog from a distance before approaching. She
greeted the owner and shook his/her hand, after which she greeted
the dog, by kneeling beside the dog and petting it on the side and
chest for approximately 30 seconds. Afterward, the dog’s willingness to cooperate was assessed by testing the dog’s willingness to
follow a stranger. The test leader took the leash and walked 1520 meters away from the owner with the dog, then paused 5 seconds before returning to the owner. The test leader took the dog in a
direction away from other patients in the waiting room. The walk
lasted for approximately 1-2 minutes. After returning to the owner,
the test leader handled the dog, that is to say physically touched the
body and chest of the dog, before opening its mouth as though
looking at its teeth (Table 2).
Behavior tests
To assess the dog’s experience of the situation, 3 behavior tests
were carried out; a social contact test, a play test, and a treat test
(Table 2). The social contact test was performed inside the waiting
room before the examination by the nurse and the veterinarian. The
dog’s interest in play was tested in a specific test area inside the
waiting room directly after the veterinary examination, followed by
the treat test in the same specific location. The test leader then led
the dog together with the owner to the test area outside the clinic
where the play and treat test were repeated. The social contact and
play tests were the standardized tests used during the Dog Mentality Assessment in Sweden (Svartberg & Forkman 2002). The treat
was a “Frolic” (Mars Nordics) with chicken flavor, a typical dog treat
in Sweden.
Social contact test
The social contact test consists of 3 different subtests, greeting,
cooperation, and handling. It shows the reaction of the dog to a
Play test
In the play test, the interest of the dog to play with a stranger
was tested both inside (in the waiting room) and outside the veterinary clinic. The play test also consists of 3 different subtests:
playfulness, grabbing, and tug-of-war, and these were performed
after the veterinary consultation. The test leader presented 5
different toys for the owner and the dog to choose which toy they
preferred for the test. The choice of toys was either a ball, a rag, a
squeaky toy, a teddy bear, or tug rope. Depending on the toy, the
owner dragged the rag or tug rope on the ground in front of the dog,
squished the squish toy, or played with the ball. If the dog showed
interest in playing, the owner gave the toy to the test leader, who
then tried to start a play with the dog; otherwise, the owner
continued to try to get the dog to play. The dog was free to play if it
wanted to, with grabbing, tug of war, and play behavior. During the
play, the test leader switched between active pulling and passive
holding the toy. After this, the owner, test leader, and dog went
outside, and the same procedure of the play session was repeated.
Table 2
The different behavior tests and the different test variables and their scoring
Test
Subtests
Score
Social contact
Greeting reaction
Score from rejection contact or growl and/or biting attempts (1) to greeting with intense contact to the test leader, can jump and
whine (5)a
Score from refusal to walk with stranger (1) to willing to walk with the stranger, combined with intense greeting toward stranger,
with jumping and whining (5)a
Score from rejects handling, with growl and/or biting attempts (1) to accepts handling with intense social behavior toward the test
leader (5)a
Score from no play at all (1) to plays very active and starts immediately (5)a
Score from no grabbing (1) to immediate and intense grabbing (5)a
Score from no biting in the toy at all (1) to immediate grabbing with twitching and fighting until the test leader lets go of the toy (5)a
Score 1: does not eat the treat. Score 2: takes the treat but spits it out. Score 3: takes the treat and eats it with hesitation. Score 4:
takes the treat, chews, and swallows. Score 5: takes the treat quick and swallows it immediately.
Cooperation
Handling
Play
Treat
a
Interest in play
Grabbing
Tug of war
Eat the treat
Scoring according to the dog mentality assessment (Fält, 1997a, 1997b). For a more detailed description of the scores, see Svartberg and Forkman (2002).
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
27
The reaction of the dog is described on a scale from 1 to 5 according
to behavioral variables from the Dog Mentality Assessment test
(Table 2).
(17%), dermatologic disorders (15%), routine check/treatment (13%),
miscellaneous (12%), gastrointestinal problems (11%), and a few for
behavior, respiratory, and reproductive problems.
Treat test
The treat test was performed directly after each of the play tests
(inside and outside the clinic) in the same specific areas. The test
leader gave the dog a Frolic treat, and the reaction of the dog was
described on a scale from 1 to 5 (Table 2).
Behavior tests
Of the 105 dogs that participated in the behavior tests, 100%
were tested in the greeting and handling subtests of the social
contact test, whereas not all dogs took part on both play tests; 72%
were tested inside and 70% were tested outside. Of the dogs
participating in the play test, 8% did not participate in the treat test
because of dietary restrictions.
Of the dogs who participated in the treat test, 94% ate a treat
and 89% of the dogs ate it at both locations. Of the dogs who ate
the treat, 63% ate it without hesitation (scores 4 or 5) inside the
clinic, but this was 74% when it was presented outside the clinic.
The few dogs with locomotor difficulties did not accompany the
test leader for the walk in the cooperation part of the social contact test or take part on the play test. The main reason dogs did not
participate in the play test was that they had already been sedated
in association with a veterinary procedure. Thus, only nonsedated
dogs took part in the play test, and of these, there was little difference in the percentage of dogs who did not play at all (28%
inside and 29% outside). Nevertheless, more dogs started to play
immediately and to play very actively (score 5) outside the clinic
(15%) compared to inside (7%). Thirty-eight percent played in both
locations.
The average scores for the social contact, play, and treat tests can
be seen in Table 3. In all subtests, scores were given from 1 to 5
except for the subtest “Greeting,” where no dogs scored “1” (score 1
means that the dog rejected contact, growled, or attempted to bite).
Only 1% of the dogs scored “1” in the subtest “Handling.” There was
a positive correlation between the play and treat tests inside and
outside the veterinary clinic. In general, the dogs were more willing
to play and eat a treat outside the clinic (P < 0.001).
Dog’s willingness to play and acceptance of eating treats depend
on a large variety of factors, in addition to breed, and therefore, in
the questionnaire, owners were asked about how often they played
with their dog and gave their dog a treat. Of the 105 behavior-tested
dogs, only 1 owner answered that he/she never played with the dog
and 2 answered that they never gave their dog a treat.
Data analysis
Descriptive statistics were produced for all participating dogs.
The agreement between observers was calculated with prevalenceadjusted bias-adjusted kappa. The interpretation of prevalenceadjusted bias-adjusted kappa is equal to that of the k value which
ranges from þ1.00 to 1.00 (Byrt et al., 1993; Lantz and Nebenzahl,
1996), a k value of >0.80 indicates excellent agreement, 0.61-0.80
good agreement, 0.41-0.60 moderate agreement, 0.21-0.40 fair
agreement, and 0.00-0.20 poor agreement. Values <0 mean that
the agreement is worse than expected by chance (Fleiss, 1981;
Landis and Koch, 1977).
To test the hypothesis of differences in dog behavior inside and
outside the clinic, an analysis of variance was performed (SAS
version 9.2) following confirmation of normality with the histogram and q-q plot options. A 1-way analysis of variance was used
for comparing the results of the dog during the play and treat tests
inside and outside the clinic, with the average ranking by the
people for the dog’s level of pain and stress. Because for pain there
was good to excellent agreement between the owner, veterinarian,
veterinary nurse, and test leader, their average score was used.
Agreement was not as high for assessment of stress, so 2 averages
were used, one being the average of the veterinarian’s and veterinary nurse’s scores, who had good agreement between them, and
the other being the average of the owner’s and test leader’s scores,
who also had good agreement between them.
For each of the 3 subscales from the MDORS, the results were
correlated against the three different behavior tests to investigate if
the dogeowner relationship influenced the outcome of the tests.
Results
Baseline information
Questionnaire study
All breeds, ages of dog, and both sexes (castrated and not castrated) were included in the study; in total, 77 different breeds were
represented among the 233 dogs. The majority of the dog owners
who participated in the questionnaire study were women (79%)
compared to men (17%), and 98% stated that they were primarily
responsible for the care of the dog. A subset of 105 dogs participated
in the behavior tests; of these dog owners, 70% were women and
19% were men, the remaining 11% did not answer the question. The
average age of the owners was 42 years (ranging from 14 to
82 years) for both the owners who participated in the behavior tests
and those who just participated in the questionnaire study.
There were 43 different breeds that were scored in the behavioral
tests. The type of dogs that were represented were gun dogs (44),
working dogs (18), herding dogs (18), terriers (6), lap dogs (2), racing
dogs (2), mongrels (9), and breed missing (4). The average age of
these 105 dogs was 4.25 years (ranging from 3 months to 13 years).
Among these, 42% were females, 41% males, 3% castrated males, 3%
castrated females, and 10% where the sex was not specified.
From the 233 questionnaires, the reasons for the visit to the clinic
were mainly muscle skeletal disorders (19%), eye and ear problems
Correlations between veterinary clinic staff, the owner and the test
leader when assessing pain and stress
When assessing pain, there was excellent agreement between
all observers except between the owner and the veterinarian, but
even here, the agreement was good. When assessing stress, there
was good agreement between the veterinarian and veterinary
Table 3
The average scoring (SD) of the different subtests in the social contact, play, and treat
tests and the correlations and P values between the play and treat tests inside and
outside the clinic
Behavior tests Subtests
Social contact
Play
Treat
N
Average
(SD)
Range Correlation P value
(1-5)
Greeting
105 3.8 (0.93) 2-5
Cooperation
99 3.5 (1.03) 1-5
Handling
105 3.5 (0.96) 1-5
Playfulness In
76 2.4 (1.20) 1-5
Out 74 2.7 (1.42) 1-5
Grabbing
In
76 2.5 (1.29) 1-5
Out 74 2.6 (1.50) 1-5
Tug of war In
76 2.0 (1.32) 1-5
Out 74 2.2 (1.52) 1-5
Treat
In
70 3.5 (0.99) 1-5
Out 68 3.7 (1.03) 1-5
0.762
<0.001
0.692
<0.001
0.716
<0.001
0.881
<0.001
28
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
Figure. The prevalence-adjusted bias-adjusted kappa value for each score of pain and stress obtained by the owner (n ¼ 105), the test leader (n ¼ 105), the veterinary nurse (n ¼
68), and the veterinarian (n ¼ 84). The mean value and the standard deviation (SD) are shown for the individual observers as (mean; SD).
nurse and between the owner and the test leader, but it was only
moderate between these 2 categories of people (Figure). It should
be noted that the veterinarian and the veterinary nurse assessed if
the dog was in pain or was stressed during their examination,
whereas the test leader and the owner assessed if the dog was in
pain or stressed while in the waiting room.
Correlations between dog behavior and estimated levels of pain and
stress
There were negative correlations between the behavior of the dog
in the social contact test and the average estimate of the level of
stress by the owner and the test leader, but no such correlations were
found with the estimates by the veterinarian and the nurse (Table 4).
There was also a negative trend in playfulness inside the clinic and
the level of stress the owner and test leader estimated the dog to be
experiencing. There were no significant correlations between the
social contact, the play, and the treat tests inside and outside the
clinic and the level of pain the dog was estimated to be experiencing.
Correlations between dog behavior and results in the MDORS
questionnaire
Owners who perceived a high emotional closeness with their
dogs had dogs which were more willing to play both inside and
Table 4
Correlations and P values of average pain scores estimated by all observers, average
stress score estimated by the owner and test leader, and the average stress score
estimated by the veterinarian and the veterinary nurse, with the behavioral tests
Subtests
Pain (average all)
Stress (owner/test
leader)
Stress (vet/nurse)
Correlation
P value
Correlation
P value
Correlation
0.639
L0.416
<0.0001
0.019
0.899
0.823
L0.425
<0.0001
0.097
0.527
0.715
L0.413
<0.0001
0.058
0.698
0.900
0.307
0.203
0.056
0.084
0.642
0.104
0.044
0.563
0.813
0.374
0.502
0.103
0.029
0.384
0.807
0.058
0.093
0.749
0.613
0.641
0.966
0.145
0.056
0.246
0.639
0.027
0.011
0.879
0.952
0.911
0.428
0.180
0.180
0.141
0.147
0.230
0.313
0.213
0.092
Greeting
In
0.047
Cooperation
In
0.023
Handling
In
0.036
Playfulness
In
0.015
Out
0.122
Grabbing
In
0.105
Out
0.080
Tug of war
In
0.055
Out
0.005
Treat
In
0.014
Out
0.099
P value
outside the clinic and were more likely to take a treat indoor
(Table 5). There was a positive trend for owners and subscale 2
(perceived emotional closeness) and a negative trend for owners
and subscale 3 (perceived costs of dog ownership; Table 6) to have a
better calming effect on the dog during the examination, according
to the veterinarian.
Correlations between behavior problems and estimated level of pain
and stress
According to the dog owners’ answers in the questionnaire, 40 of
105 behavior-tested dogs (38%) showed behaviors that were
perceived by the owners as being a problem. The most frequent
problems reported by the owners were that the dog chased sticks,
stones, and balls (30%); showed excessive barking and licking (each
14%); chased its tail (10%); jumped up against walls and/or was
biting furniture (6%); and excessive digging (4%). Dogs with a
behavior problem were assessed as being more stressed according
to the average assessment of stress by the owner and the test leader,
compared to dogs that did not have a behavior problem (P ¼ 0.019)
with a mean average score of stress 4.25 (standard deviation ¼ 2.57)
for dogs with a behavior problem and 3.32 (standard deviation ¼
2.39) for dogs with no behavior problems. However, no such correlation was found with the assessment of stress given by the
veterinarian or the nurse (P ¼ 0.115). From the total population of
people who filled in the questionnaire, 52% of owners reported that
their dog showed 1 of the behaviors in the list.
Table 5
The Pearson correlation (Corr) between the 3 subscales in the MDORS questionnaire
with the social contact test and the 2 behavior tests performed inside and outside
the clinic
Tests
Social
Greeting
contact Cooperation
Handling
Play
Playfulness
Grabbing
Tug of war
Treat
Significant results are in bold.
Subtests
Treat
Subscale 1
In
In
In
In
Out
In
Out
In
Out
In
Out
Subscale 2
Corr
P value Corr
0.179
0.117
0.105
0.220
0.432
0.233
0.128
0.187
0.091
0.208
0.146
0.097
0.287
0.334
0.081
0.432
0.064
0.321
0.138
0.480
0.111
0.274
0.033
0.117
0.045
L0.392
L0.372
L0.366
L0.264
L0.311
L0.259
L0.270
0.233
Subscale 3
P value Corr
P value
0.762
0.288
0.677
0.001
0.003
0.003
0.038
0.012
0.042
0.037
0.078
0.666
0.719
0.591
0.423
0.652
0.690
0.664
0.619
0.752
0.154
0.476
0.047
0.040
0.058
0.102
0.058
0.051
0.056
0.063
0.041
0.186
0.095
Significant results are shown in bold.
Subscale 1 ¼ dogehuman interaction (a high score means more interactions), subscale 2 ¼ perceived emotional closeness (a high score means a low emotional
closeness), subscale 3 ¼ perceived costs of dog ownership (a high score means a high
awareness of the cost).
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
Table 6
Correlation between the 3 subscales in the MDORS questionnaire and the owners’
calming effect on the dog compared during the veterinarian examination of the dog,
scored on a scale from 1 (none) to 10 (very much) by the veterinarian
Owners
calming effect
Subscale 1
Subscale 2
Subscale 3
Correlation P value Correlation P value Correlation P value
Scored by the 0.052
veterinarian
0.646
0.210
0.062
0.206
0.066
Subscale 1 ¼ dogehuman interaction (a high score means more interactions), subscale 2 ¼ perceived emotional closeness (a high score means a low emotional
closeness), subscale 3 ¼ perceived costs of dog ownership (a high score means a high
awareness of the cost).
Discussion
The owner and test leader had a similar view of how stressed the
dogs were during the veterinary visit and the veterinarians, and the
veterinary nurses also had a similar view. In general, the veterinarians and veterinary nurses assessed the level of stress to be
higher. This may reflect genuine differences in how different categories of people assess stress, but alternatively, it may also reflect
where the assessment is taking place. The behavior of the dogs in
the social contact test was significantly correlated with the owner
or test leader assessment of its stress. That there was no correlation
between the dogs’ social contact test behavior and the estimates of
stress by the veterinarians and the veterinary nurses is probably
because these were not done at the same location.
Dogs were less likely to play and take a treat inside the clinic
than they were outside, suggesting that the dogs themselves
experienced inside the clinic as being more negative, although the
indoor test was carried out sooner after the examination, whereas
by the time of the outdoor test, the dog had more time to recover
after the examination experience. Together, these findings suggest
that the examination itself may have been experienced more
negatively than the waiting room or it may be an effect of many
negative events following one another.
All 4 categories of people agreed on the assessment of pain;
however, few dogs in this study were perceived as being in severe
pain. It is important to consider that dogs at a clinic may mask their
pain, and earlier studies have suggested that an absence of normal
behavior to be the best pain indicator (Sharkey, 2013). Although for
a few dogs the examination itself may have been painful, in general,
it may be that pain is less likely to vary between the waiting room
and the examination room, whereas stress levels are more likely to
vary over time or between different locations because it is context
specific. This may in part explain the differences in the extent to
which the behavior tests correlated with the assessment of stress by
the 4 categories of people.
Assessment of pain and stress in relation to dog behavior
It is generally assumed that pain affects the behavior of the
animal differently in different individuals, and there is variation
according to species, breed, and age (Sharkey, 2013; Mathews
et al., 2014). There were no significant correlations between any
of the behavior tests and the pain assessments as the dogs were
equally willing to play or eat a treat regardless of the scores given
by any of the assessors. This can be due to the fact that the average
scoring of pain by the 4 categories of observers was between 2.08
and 3.09 (max 10), meaning that most of the dogs were not in pain
according to the observers. Perhaps, this reflects low average pain
scores for all dogs at the clinic, but it may also be that owners who
considered that their dog was in pain were unwilling to participate
in the study. Nevertheless, in severe cases, the behavior of dogs is
29
likely to be biased by illness. This is a potential limitation of the
study. The alternative approach, however, of only using healthy
dogs in the study, would have the limitation of not being applicable to a veterinary clinic where only a proportion of the dogs are
there for routine checks or vaccinations. In our study, there was a
mixture of healthy and sick dogs, and in future studies, any confounding effects of health status should be investigated more
systematically.
The mean values for the average ranking of stress (2.37-4.30) by
the 4 observers were higher than those for pain, and the distribution of dogs being stressed varied more compared to pain. Stress
level, as perceived by the owner, was correlated with reduced social
contact between the dog and the test leader inside the clinic.
However, there was only weak evidence that the dog’s play
behavior was affected by whether the dog was assessed as stressed
as the results in this study only indicated a trend that playfulness
inside was reduced by increased stress scores. Dogs were more
willing to play outside, and no correlations between stress levels
and outdoor play were found. The difference in play activities can
be explained by the 2 different environments (such as possible
distraction from other dogs in the waiting room) but also by the fact
that the dog had already played inside the clinic and may therefore
have wanted to play more and, by the time the outside test was
carried out, the dog was more familiar with the test leader. However, it seems that familiarity with the human play partner is less
important than the type of play with regards to willingness to play
(Tóth et al., 2008), and in the present study, the type of game and
objects were selected by the owner based on the dog’s previous
experience.
The treat test showed that the dogs were more willing to eat
outside compared to inside the clinic, but it is not possible to say if
this depended on the environmental changes or the fact that they
had already received a treat from the test leader inside the clinic. It
was also clear that play behavior was influenced by the type of
relationship that the owner had with the dog.
To conclude, the behavior tests used in the study, in particular
the social contact test, can be used as an indicator for the reactions
of the dogs to the veterinary clinic. The social contact test was the
most stable test as the play behavior can be affected by the breed or
the relationship with the owner, and a few dogs did not participate
in the treat test because of dietary problems, which makes this test
less suitable. Furthermore, the play behavior of the dog was influenced by the dogeowner relationship as the results show that
owners with perceived high emotional closeness with their dogs
(ranked high on subscale 2) had dogs who were more willing to
play both inside and outside the clinic.
It is proposed that simple behavioral tests, especially how the
dog greets an unfamiliar person at the clinic, in combination with
a rating by owner, veterinary nurse, or veterinarian may provide a
practical way to estimate stress levels in dog patients. These types
of tests may provide useful complementary information to the
staff in the same way as does a pain assessment score. Pain may
result in aggressive behavior (Fureix et al., 2010) as probably stress
does also. Such dogs present risks for the examining veterinarian
because they need to handle dogs to examine them properly. Even
if the number of dogs reacting negatively to being greeted or
handled by an unfamiliar person was low in our study, the
awareness of dog’s different reactions is important, and a simple
contact test, as the one presented here, should give both the
veterinarian and the owner a tool to predict the dogs reaction and
thereby give them the possibility to take precautions for their own
safety. Thereby, the overall experience of the clinic can be
improved.
A trend was found suggesting that owners who were less
emotionally close or less concerned about the cost of having a dog
30
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
had a greater calming effect than owners who were emotionally
close or concerned about the costs of pet ownership. To our
knowledge, this is the first time that such results have been
demonstrated in this particular context, but such a finding would
need to be confirmed.
Behavior problems and levels of stress
Over half of the dogs (52%) entering the veterinary clinic had
behavioral problems. We did not specify whether there should be
behavioral problems associated with reduced welfare for the dog
or merely an inconvenience for the owner, so this may account in
part for the high proportion. Dogs that according to the owner
had a behavior problem were also more stressed according to the
average scoring of stress by the owner and the test leader. The
test leader was “blind” to the information in the questionnaire at
the time the stress assessment scoring was made, although
knowledge about the dog’s behavioral problem was of course
known by the owner.
Conclusion
The results of the study indicate that contact test is a better test
for assessing stress compared to play and treat test. Play behavior
can be affected by the relationship with the dog owner and not all
dogs were allowed to eat a treat because of dietary problems.
Subjective stress and pain assessments by different observers were
in good agreement. A follow-up study is required to validate this
against already accepted measures of stress. Nevertheless, this
study implies that there is a potential to develop a reliable,
noninvasive stress scoring system for use in the veterinary clinic,
and we suggest that the social contact test may even be useful as a
preliminary screening for dogs that might be dangerous for the
veterinarian to handle.
Acknowledgment
The authors thank all the dogs and dog owners, veterinarians,
veterinary nurses, and students at the Animal Hospital at the
Swedish University of Agricultural Sciences, Uppsala, Sweden for
their participation. They specially thank Fleur Dwyer for letting
them use the “Monash Dog-Owner Relationship Scale.” The final
stage of this work was carried out within the Centre of Excellence in
Animal Welfare Science, a Swedish collaborative research environment funded by FORMAS.
Author contributions: All authors participated in the study
design. A.-K.L. was responsible for data collection and performed
the statistical analysis. A.-K.L. interpreted the data and drafted the
manuscript. L.J.K. revised the first versions of the manuscript and
corrected the language. All authors revised the manuscript and
have read and approved the final manuscript.
Ethical statement
The Swedish Board of Agriculture approved the use of nondestination bred dogs, and the Animal Ethics Committee in Uppsala,
Sweden, approved the experimental procedure.
Conflict of interest
The authors declare no conflict of interest.
References
Beerda, B., Schilder, M.B., Bernadina, W., van Hooff, J.A., de Vries, H.W., Mol, J.A.,
1999. Chronic stress in dogs subjected to social and spatial restriction. II. Hormonal and immunological responses. Physiol. Behav. 66, 243e254.
Belew, A.M., Barlett, T., Brown, S.A., 1999. Evaluation of the white-coat effect in cats.
J. Vet. Intern. Med. 13, 134e142.
Byrt, T., Bishop, J., Carlin, J.B., 1993. Bias, prevalence and kappa. J. Clin. Epidemiol. 46,
423e429.
Downing, R., 2011. Pain management for veterinary palliative care and hospice
patients. Vet. Clin. North Am. Small Anim. Prac. 41, 531e550.
Dwyer, F., Bennett, P.C., Coleman, G.J., 2006. Development of the Monash Dog
Owner Relationship Scale (MDORS). Anthrozoös 19, 243e256.
Fält, L., 1997a. Anvisningar Mentalbeskrivning. Svenska Brukshundklubben,
Stockholm, Sweden.
Fält, L., 1997b. Kompendium Mentalitet. Svenska Brukshundklubben, Stockholm,
Sweden.
Fleiss, J.L., 1981. Statistical methods for rates and proportions. In: Wiley Series in
Probability and Mathematical Statistics, 2nd Edition. Wiley-Interscience, New
York.
Fureix, C., Menguy, H., Hausberger, M., 2010. Partners with bad temper: reject or
cure? A study of chronic pain and aggression in horses. PLoS One 5, e12434.
Hellyer, P., Rodan, I., Brunt, J., Downing, R., Hagedorn, J.E., Robertson, S.A., 2007.
AAHA/AAFP Pain Management Guidelines for Dogs and Cats. J. Feline Med. Surg.
9, 466e480.
Hennessy, M.B., Voith, V.L., Young, T.L., Hawke, J.L., Centrone, J., McDowell, A.L.,
Linden, F., Davenport, G.M., 2002. Exploring human interaction and diet effects
on the behavior of dogs in a public animal shelter. J. Appl. Anim. Welfare Sci. 5,
253e273.
Jones, A.C., Gosling, S.D., 2005. Temperament and personality in dogs (Canis familiaris): a review and evaluation of past research. Appl. Anim. Behav. Sci. 95, 1e53.
Kallet, A.J., Cowgill, L.D., Kass, P.H., 1997. Comparison of blood pressure measurements obtained in dogs by use of indirect oscillometry in a veterinary clinic
versus at home. J. Am. Vet. Med. Assoc. 210, 651e654.
Koolhaas, J.M., Bartolomucci, A., Buwalda, B., de Boer, S.F., Flügge, G., Korte, S.M.,
Meerlo, P., Olivier, B., Palanza, P., Richter-Levin, G., Sgoifo, A., Steimer, T.,
Stiedl, O., van Dijk, G., Vöhr, M., Fuchs, E., 2011. Stress revisited: a critical
evaluation of the stress concept. Neurosci. Biobehav. Rev. 35, 1291e1301.
Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for categorical data. Biometrics 33, 159e174.
Lantz, C.A., Nebenzahl, E., 1996. Behavior and interpretation of the k statistic: resolution of the two paradoxes. J. Clin. Epidemiol. 49, 431e434.
Mariti, C., Raspanti, E., Zilocchi, M., Carlone, B., Gazzano, A., 2015. The assessment of
dog welfare in the waiting room of a veterinary clinic. Anim. Welfare 24, 299e305.
Mathews, K., Kronen, P.W., Lascelles, D., Nolan, A., Robertson, S., Steagall, P.V.,
Wright, B., Yamashita, K., 2014. Guidelines for recognition, assessment and
treatment of pain. J. Small Anim. Prac. 55, E10eE68.
Meagher, R.K., 2009. Observer ratings: validity and value as a tool for animal welfare
research. Appl. Anim. Sci. 119, 1e14.
Merola, I., Mills, D.S., 2015. Systematic review of the behavioral assessment of pain
in cats. J. Feline Med. Surg. 18, 60e76.
Merola, I., Prato-Previde, E., Marshall-Pescini, S., 2012. Dogs’ social referencing towards owners and strangers. Plos One 7, 1e9.
Merskey, H., Bogduk, N. (Eds.), 1994. Classification of Chronic Pain. IASP Task Force
on Taxonomy. IASP Press, Seattle, pp. 209e214.
Moberg, G.P., 2000. Biological Responses to Stress: Implications for Animal Welfare
in the Biology of Animal Stressdthe Principles and Implications for Animal
Welfare. Edited by Moberg G.P. and Mech J.A.. CABI Publishing, Wallingford, UK,
ISBN 0851993591, pp. 1e21.
Quimby, J.M., Smith, M.L., Lunn, K.F., 2011. Evaluation of the effects of hospital visit
stress on physiologic parameters in the cat. J. Feline Med. Surg. 13, 733e737.
Rehn, T., Lindholm, U., Keeling, L., Forkman, B., 2014. I like my dog, does my dog like
me? Appl. Anim. Behav. Sci. 150, 65e73.
Reid, J., Nolan, A.M., Hughes, J.M.L., Lascelles, D., Pawson, P., Scott, E.M., 2007.
Development of the Short-Form Glasgow Composite Measure Pain Scale (CMPSSF) and Derivation of an Analgesic Intervention Score. Anim. Welfare 16, 97e104.
Rodan, I., Sundahl, E., Carney, H., Gagnon, A., Heath, S., Landsberg, G., Seksel, K.,
Yin, S., 2011. AAFP and ISFM feline-friendly handling guidelines. J. Feline Med.
Surg 13, 364e375.
Rooney, N.J., Gaines, S.A., Bradshaw, J.W.S., 2007. Behavioural and glucocorticoid
responses of dogs (Canis familiaris) to kennelling: investigating mitigation of
stress by prior habituation. Physiol. Behav. 92, 847e854.
Sales, G., Hubrecht, R., Peyvandi, A., Milligan, S., Shield, B., 1997. Noise in dog kennelling: is barking a welfare problem for dogs? Appl. Anim. Behav. Sci. 52, 321e329.
Scotney, R., 2010. Environmental enrichment in veterinary practice. Vet. Nurse 1,
140e149.
Sharkey, M., 2013. The challenges of assessing osteoarthritis and postoperative pain
in dogs. The AAPS J. 15, 598e607.
Siracusa, C., Manteca, X., Cerón, J.J., Cuenca, R., Lavın, S., Garcia, F., Pastor, J.,
2007. Perioperative stress in dogs: neuroendocrine and immune responses
(perioperative stress response in sheltered dogs). J. Vet. Behav.: Clin. Appl.
Res. 2, 94e95.
Siracusa, C., Manteca, X., Cerón, J., Martínez-Subiela, S., Cuenca, R., Lavín, S.,
Garcia, F., Pastor, J., 2008. Perioperative stress response in dogs undergoing
A.-K. Lind et al. / Journal of Veterinary Behavior 17 (2017) 24e31
elective surgery: variations in behavioural, neuroendocrine, immune and acute
phase responses. Anim. Welfare 17, 259e273.
Stanford, T.L., 1981. Behavior of dogs entering a veterinary clinic. Appl. Anim. Ethol.
7, 271e279.
Strandberg, E., Jacobsson, J., Saetre, P., 2005. Direct genetic, maternal and litter effect
on behaviour in German Shepherd dogs in Sweden. Livest. Prod. Sci. 93, 33e42.
Svartberg, K., 2002. Shyness-boldness predicts performance in working dogs. Appl.
Anim. Behav. Sci. 79, 157e174.
Svartberg, K., Forkman, B., 2002. Personality traits in the domestic dog (Canis
familiaris). Appl. Anim. Behav. Sci. 79, 133e155.
Svartberg, K., Tapper, I., Temrin, H., Radesäter, T., Thorman, S., 2005. Consistency of
personality traits in dogs. Anim. Behav. 69, 283e291.
Tóth, L., Gácsi, M., Topál, J., Miklósi, Á., 2008. Playing styles and possible causative
factors in dogs’ behaviour when playing with humans. Appl. Anim. Behav. Sci.
114, 473e484.
Väisänen, M., Valros, A., Hakaoja, E., Raekallio, M., Vainio, O., 2005. Pre-operative
stress in dogsda preliminary investigation of behaviour and heart rate variability in healthy hospitalized dogs. Vet. Anaesth. Analg. 32, 158e167.
31
van den Berg, L., Schilder, M.B.H., de Vries, H., Leegwater, P.A.J., van Oost, B.A., 2006.
Phenotyping of aggressive behavior in golden retriever dogs with a questionnaire. Behav. Genet. 36, 882e902.
van Loon, J.P.A.M., Back, W., Hellebrekers, L.J., van Weeren, P.R., 2010.
Application of a composite pain scale to objectively monitor horses with
somatic and visceral pain under hospital conditions. J. Equine Vet. Sci. 30,
641e649.
Viñuela-Fernández, I., Jones, E., Welsh, E.M., Fleetwood-Walker, S.M., 2007. Pain
mechanisms and their implication for the management of pain in farm and
companion animals. Vet. J. 174, 227e239.
Viñuela-Fernández, I., Weary, D.M., Flecknell, P., 2011. Pain. In: Appleby, M.C.,
Mench, J.A., Plsson, I.A.S., Hughes, B.O. (Eds.), Animal Welfare, 2nd Edition.
Cambridge University Press, Cambridge, UK, pp. 64e77.
Weiss, E., 2002. Selecting shelter dogs for service dog training. J. Appl. Anim.
Welfare Sci. 5, 43e62.
Wilsson, E., Sinn, D.L., 2012. Are there differences between behavioral measurement
methods? A comparison of the predictive validity of two ratings methods in a
working dog program. Appl. Anim. Behav. Sci. 141, 158e172.