Zoo Biology 28 : 545–560 (2009)
Animal-Based Welfare Monitoring: Using Keeper Ratings as an Assessment Tool Jessica C. Whitham and Nadja Wielebnowski Chicago Zoological Society, Brookfield, Illinois Zoological institutions are in urgent need of identifying and implementing welfare assessment tools that allow for ongoing, quantitative monitoring of individual animal well-being. Although the American Zoological Association’s (AZA) Animal Welfare Committee (AWC) promotes the use of such tools in internal review processes, current approaches to institutional welfare assessment are resource-based and outline the resources, environmental parameters and ‘‘best practices’’ recommended for promoting good welfare in a species in general. We highlight the value of incorporating animal-based monitoring tools that capture the individual animal’s perspective and subjective experiences, including positive events and feelings, by validating zookeepers’ qualitative assessments. We present evidence that, across a variety of species, caretakers’ assessments of traits related to the well-being of individual animals can be both reliable and valid. Furthermore, we demonstrate that among researchers investigating the welfare of farm, laboratory, companion and even zoo animals, support already exists for developing and validating instruments that objectively evaluate the qualitative assessments of caretakers. Finally, we outline a process currently being evaluated at Brookfield Zoo for developing, validating and testing a cost-effective, userfriendly monitoring tool that will help to quantify keepers’ qualitative assessments of individual well-being and can be integrated into daily operations. This tool (i.e. species-specific Welfare Score Sheets designed through consultation with animal experts) will result in weekly scores of individual well-being that are expected to provide a first indicator of welfare issues in the collection. Specifically, scores can be reviewed during regular workgroup meetings to identify welfare issues proactively, to assess whether particular conditions, practices or events impact Grant sponsors: The Chicago Board of Trade Endangered Species Fund; The Women’s Board of the Chicago Zoological Society. Correspondence to: Jessica C. Whitham, Chicago Zoological Society, Brookfield Zoo, 3300 Golf Road, Brookfield, IL 60513. E-mail:
[email protected]
Received 29 May 2009; Revised 22 July 2009; Accepted 14 August 2009 DOI 10.1002/zoo.20281 Published online 22 October 2009 in Wiley InterScience (www.interscience.wiley.com).
r 2009 Wiley-Liss, Inc.
546 Whitham and Wielebnowski individual well-being, and finally, to evaluate the effectiveness of efforts to address welfare issues. Upon completion of the tool validation and testing phases, we plan to make the Welfare Score Sheets for our 12 study species available to other institutions, and the methods we applied may serve as a ‘‘blueprint’’ for creating similar tools for additional species and institutions. Zoo Biol 28:545–560, 2009. r 2009 Wiley-Liss, Inc.
Keywords: zoo animal welfare; keeper assessments; animal well-being; animal-based assessments; welfare-monitoring tool
INTRODUCTION In concordance with an increasing public concern about animal welfare and an ethical demand for higher internal welfare standards by zoos and aquariums worldwide [e.g. WAZA, 2005], a growing number of institutions are eager to identify assessment tools to monitor and improve animal welfare. Indeed, since its inception in 2000, the Animal Welfare Committee (AWC) of the Association of Zoos and Aquariums (AZA) has been struggling with the issue of effective welfare monitoring [discussed by Barber, 2009; this issue]. Although animal welfare science adopts a variety of approaches to identify welfare measures, most studies are costly, timeintensive and require extensive multi-institutional and multi-disciplinary data collection. Thus, there is an urgent need for more rapid and regular institutional assessments of animal welfare that may be incorporated into daily management and breeding programs. The development of a ‘‘user-friendly’’ welfare-monitoring tool that can be applied at low cost also may aid zoological institutions with the establishment of their own internal welfare review processes. AZA accreditation now requires that institutions develop an Institutional Animal Welfare Process (IAWP) that addresses possible welfare issues through an internal review committee. In addition, regular welfare monitoring may allow institutions to become more effective at prioritizing the distribution of resources according to welfare needs and quantitatively assessing the impact of management decisions. One common approach to institutional welfare assessments involves quantifying the resources and environmental parameters deemed necessary to provide the potential for good welfare. To this end, AZA has been developing Animal Care Manuals (ACMs) for identifying and satisfying the general biological and physical needs of zoo animals. The manuals include husbandry templates and focus on describing the resources, as well as the environmental and social parameters, assumed to promote good welfare. However, simply focusing on these factors may not be sufficient to ensure good individual animal well-being [Barber, 2009; this issue], especially as husbandry and care guidelines often are developed with a lack of scientific data for a given species. Also, ACMs are written with a broader perspective in mind to allow for the establishment of general species-specific care guidelines. Therefore, although ACMs try to encompass a wide range of needs for a given species, they cannot adequately address all possible cases and circumstances. Recently, the use of animal-based parameters (e.g. physical condition, behavioral and/or physiological indicators), in addition to resource-based and environmental variables, has been recommended for welfare assessment [Hewson, 2003; Whay et al., 2003; FAWC, 2005; Main et al., 2007; Whay, 2007]. In fact, recent Zoo Biology
Monitoring Zoo Animal Well-Being 547
work suggests that animal-based indices that extend beyond the scope of general health and body condition allow us to assess well-being in a manner that more closely reflects an individual animal’s perception of its environment, and considers the needs of each individual [McMillan, 2000, 2003; Wiseman-Orr et al., 2006; Scott et al., 2007; Taylor and Mills, 2007; Timmins et al., 2007]. However, the success of such an enterprise, which involves attempting to capture an individual’s point of view in ‘‘real’’ time (i.e. obtaining rapid welfare assessments that provide timely feedback for daily management decisions), hinges, in part, on validating qualitative assessments of animal welfare using quantitative measures. We use the term qualitative assessment to refer to the holistic approach, adopted by experienced caretakers (e.g. animal keepers), that involves quickly integrating and filtering detailed bits of information to evaluate an animal’s condition (note: this differs from the definition used by some other disciplines, e.g. social sciences). Support for developing approaches that employ qualitative assessments has been blossoming in Europe [e.g. Wemelsfelder and Lawrence, 2001], where public pressure for stricter animal care guidelines has resulted in new and improved welfare regulations being mandated for zoos and other institutions [DEFRA, 2007]. Evidence exists that validating qualitative assessments can provide a cost-effective and reliable approach to monitoring farm, laboratory and companion animal welfare [e.g. McMillan, 2000; Morton, 2000; Wemelsfelder and Lawrence, 2001]. Animal care professionals often inadvertently but regularly perform such informal, qualitative welfare assessments. On a daily basis, they apply accumulated experiences to filter a large amount of information rapidly and arrive at a holistic evaluation of the well-being of individuals [e.g. Wemelsfelder and Lawrence, 2001]. Frequently, experienced animal care professionals detect subtle changes in an individual’s behavior and condition that may indicate significant alterations in wellbeing. There is growing evidence from various fields that qualitatively assessed traits linked to welfare can be measured reliably in a variety of species [reviewed by Gosling, 2001]. For zoos, a key component of developing and utilizing an animalbased monitoring tool is tapping into one of our most valuable resources, keeper expertise and experience. Validating qualitative assessments can be challenging, but may ultimately provide a major step toward the development of a tool for ongoing animal-based welfare monitoring. Below, we present an argument for the regular use of animal-based measures and for the validation of qualitative assessments. We address the general reliability and validity of such assessments and provide an overview of their application in other fields of welfare science. Lastly, we provide a brief outline of a process we currently are evaluating for developing, validating and testing a monitoring tool prototype at Brookfield Zoo. A New Approach to Zoo Animal Welfare Assessment Shifting the focus: assessing individuals using animal-based measures One of the most widely used definitions of animal welfare maintains that the welfare of an individual reflects its ability to cope with challenges in the environment [Broom, 1986]. Unfortunately, two key concepts embedded in this definition traditionally have been overlooked when considering the assessment of animal welfare. Zoo Biology
548 Whitham and Wielebnowski
First, careful consideration is not always given to exactly whose welfare should be assessed. Broom’s [1986] definition suggests that assessments should focus on the individual. However, for farm animals, it is common for assessments to be performed at the group-level [e.g. Whay et al., 2003]. Similarly, AZA’s current approach to institutional welfare assessment is to develop ACMs that outline the resources and environmental parameters recommended for satisfying the biological and physical needs of a particular taxon, and then to attempt to quantify resources and evaluate parameters. Although it is essential to compile and regularly update available information about the ‘‘best practices’’ thought to promote good welfare for a species, it cannot be assumed that providing the recommended resources or following particular husbandry routines will necessarily result in individuals experiencing good well-being. The assumption that providing certain conditions or resources will promote good welfare leads to the second, related component of Broom’s [1986] definition that must be addressed: which measures should be assessed? Although Broom stresses that the focus should be on how animals cope with the environment, animal welfare assessment often removes the animal from the equation and mainly considers its environment. Until recently, certification schemes for farms in the United Kingdom have relied on input-based assessments, similar to AZA’s emphasis on care standards and guidelines. In other words, specific husbandry practices and environmental requirements (e.g. shelter, space, nutrition) are delineated, and then practices are evaluated and provided resources are measured [discussed by Hewson, 2003; Whay et al., 2003; FAWC, 2005; Main et al., 2007; Whay, 2007]. However, evidence exists that animals living on farms that voluntarily participate in programs auditing for strict environmental requirements do not necessarily experience better health or welfare than those on nonparticipating farms [Hewson, 2003; Whay et al., 2003]. This is because welfare assessments that only measure available resources may not be as effective as those that include animal-based measures (e.g. physical state, behavior, temperament), which require direct observations of behavior and/or assessments of physiological and physical condition [Hewson, 2003; Whay et al., 2003; FAWC, 2005; Main et al., 2007; Whay, 2007]. As a result, the inclusion of animal-based measures has been endorsed by the UK’s Farm Animal Welfare Council [2005], as well as companion animal welfare researchers [e.g. Hewson, 2003]. Whay’s statement [2007, p 118], ‘‘animal-based observations give the most direct insight into how animals are coping within their own environment,’’ highlights the need for adopting an approach that focuses on the individual and assesses each animal’s particular situation. Quality of life Recently, in an attempt to shift the focus to an individual animal’s particular situation, animal welfare researchers have proposed embracing a quality of life (QOL) framework, similar to that which has been adopted by those studying well-being in humans [Sandøe, 1996; Mench, 1998; McMillan, 2000, 2003; Wojciechowska and Hewson, 2005; Wiseman-Orr et al., 2006; Morton, 2007; Scott et al., 2007; Taylor and Mills, 2007; Timmins et al., 2007]. Central to this line of research is the concept that QOL is subjective and can only be measured accurately by considering the individual’s perspective [McMillan 2000, 2003; Wiseman-Orr et al., 2006; Scott et al., 2007; Taylor and Mills, 2007; Timmins et al., 2007]. In other words, just as Zoo Biology
Monitoring Zoo Animal Well-Being 549
individual animals receiving adequate resources and proper management do not necessarily experience good well-being, measuring objective life status in humans (i.e. collecting data on income, housing, etc.) will not always yield the same results as measuring subjective life satisfaction [Li et al., 1998]. In fact, Lykken and Tellegen [1996] discovered that predictors such as socioeconomic status, educational attainment and income each explained less than 3% of the variance in self-ratings of happiness. In humans, this discrepancy is addressed by using subjective criteria (e.g. measures of affect or emotion) in conjunction with objective criteria. Many animal welfare researchers have suggested using a similar approach for animals, stressing that welfare is mostly, or even completely, dependent upon an animal’s feelings [Duncan and Dawkins, 1983; Dawkins, 1990; Duncan, 1996, 2006; McMillan, 2000, 2003]. The QOL framework not only supports incorporating criteria that reflect an individual’s subjective feelings but also specifically stresses the importance of considering positive feelings and experiences. When humans perform self-evaluations of well-being, weighing unpleasant emotions and experiences against pleasant emotions and experiences is crucial, with one of the best predictors of life satisfaction being pleasantness (i.e. positive affect or emotions) [Diener and Larsen, 1993]. Similarly, animal welfare researchers have cautioned against simply focusing on suffering, stress and negative emotions, arguing that pleasure and positive experiences are essential components of good welfare [Gonyou, 1993; Fraser, 1995; Sandøe, 1996; Mench, 1998; McMillan, 2000, 2003; Duncan, 2006; Boissy et al., 2007; Morton, 2007; Taylor and Mills, 2007; Yeates and Main, 2008] and that experiencing pleasurable events may offset the effects of negative events at times [Duncan, 2006]. Ultimately, the goal should not be to simply avoid poor QOL, but rather, to attain excellent QOL, and this may be accomplished by considering an individual’s unique perspective, including positive states and experiences [Boissy et al., 2007; Scott et al., 2007; Yeates and Main, 2008]. In sum, adopting a QOL framework promotes the use of a more holistic, animal-centered approach to assessing the well-being of individual animals. As individual animals may have unique perspectives, preferences and needs [McMillan, 2000, 2003; Wojciechowska and Hewson, 2005], it has been even suggested that standards for improving well-being and concepts of ‘‘normality’’ should be established for individuals, not just for groups or species [Clark et al., 1997b; Morton, 2007]. In the very least, welfare assessment tools should include animalcentered measures that attempt to capture the individual’s perspective and subjective feelings. Caretakers as proxies When assessing well-being, the ideal way to obtain information about an individual’s perspective or subjective feelings is to gather input directly from that individual. In humans, when it is not possible for a person to communicate care preferences or participate in clinical decision making, as is the case with infants, severely ill patients or those with cognitive deficiencies, an assessment is made using proxy informants [Sprangers and Aaronson, 1992; Addington-Hall and Kalra, 2001; Eiser and Morse, 2001]. Often, sufficient agreement is found between self-reports of QOL and assessments made by parents, spouses or caregivers [e.g. Addington-Hall and Kalra, 2001]. Although proxies are better at assessing observable, concrete Zoo Biology
550 Whitham and Wielebnowski
dimensions (e.g. physical symptoms, mobility) than those that are less observable (e.g. anxiety, depression) [Sprangers and Aaronson, 1992; Addington-Hall and Kalra, 2001; Eiser and Morse, 2001], evidence exists that caretakers may be better than clinicians when assessing these less observable dimensions [Bryan et al., 2005]. Not surprisingly, then, several animal welfare researchers and veterinarians have suggested that the best way to evaluate well-being may be for the person most familiar with an animal’s temperament, preferences, behavior and routine to be ‘‘the voice’’ for that individual [McMillan, 2000, 2003; Morton, 2000, 2007; Serpell and Hsu, 2001; Hsu and Serpell, 2003; Wojciechowska and Hewson, 2005; Wojciechowska et al., 2005a; Wiseman-Orr et al., 2006; Taylor and Mills, 2007; Timmins et al., 2007; Meagher, 2009]. Recently, animal welfare researchers have highlighted the value of proxy evaluations and the benefits of adopting approaches that require an interpretation of the animal’s subjective experience, encouraging caretakers to embrace rather than shy away from making qualitative assessments [Wemelsfelder, 1997, 2007; McMillan, 2000, 2003; Hewson, 2003; Wemelsfelder et al., 2000, 2001; Wemelsfelder and Lawrence, 2001]. Traditionally, animal caretakers, including those in the zoological community, have been discouraged from engaging in what might be interpreted as anthropomorphic thinking. However, we would like to point out that there has been growing support for applying critical anthropomorphism [coined by Burghardt, 1985], which involves combining various sources of information (e.g. observations, thoughts, feelings) to gain insight into an individual’s state and which allows for generating testable hypotheses [Burghardt, 1991; Clark et al., 1997a; Scott et al., 2007; Taylor and Mills, 2007; Meagher, 2009]. In fact, when hypotheses based on qualitative assessments are evaluated using the scientific method, and such assessments are found to be based on observable and quantifiable criteria, some researchers would argue that the term anthropomorphic is no longer applicable [e.g. Wemelsfelder et al., 2000]. Many zookeepers have decades of experience with a particular species, as well as the opportunity to observe individuals over long periods of time and in a variety of contexts. Therefore, by integrating subtle details such as slight changes in behavior, posture, attitude, expression or movement, experienced caretakers may be able to detect shifts in well-being that might be overlooked when conducting systematic observations that rely on coding discrete behaviors [Block, 1977; Stevenson-Hinde et al., 1980; Feaver et al., 1986; Wemelsfelder, 1997, 2007; Carlstead et al., 1999; Wielebnowski, 1999; Wemelsfelder et al., 2000, 2001; Gosling, 2001; Wemelsfelder and Lawrence, 2001; Wielebnowski et al., 2002; King and Landau, 2003]. Wemelsfelder et al. [2000, 2001; Wemelsfelder and Lawrence, 2001] refer to this whole-animal approach of synthesizing and filtering accumulated knowledge to make rapid, qualitative judgments about an individual’s well-being as the integrated assessment of animal welfare. The value of such assessments also has been recognized in other disciplines, with an entire book dedicated to describing how rapid cognition is employed by firefighters, doctors and business executives [Gladwell, 2007]. Specifically, in the business world, evidence exists that successful executives are highly intuitive, and executive intuition, which is, ‘‘ybased upon a broad constellation of past experiences, knowledge, skills, perceptions, and feelings held tacitly,’’ is viewed as an extremely useful skill as it allows for, ‘‘y an implicit perception of the total problem’’ [Sadler-Smith and Shefy, 2004, p 83]. Similarly, the Zoo Biology
Monitoring Zoo Animal Well-Being 551
value of hunches and intuitions that are based on keeper expertise and allow for whole-animal assessments should not be overlooked. A call for developing qualitative assessment tools The call for the development and validation of instruments that quantify qualitative assessments of individual animal well-being has come from those researching companion, laboratory, farm, and recently, zoo animals. Companion animal welfare researcher McMillan [2000] stressed, ‘‘yit is necessary to develop and validate an instrument for proxy assessment of QOL in animalsy’’ (p 1909). For laboratory animals, Morton [2000] has encouraged animal care staff to use score sheets that include qualitative items (e.g. whether the animal is ‘‘inquisitive’’), to determine whether an individual’s condition is improving or deteriorating [see also Hawkins, 2002]. In the farm animal welfare community, the UK and Scottish agricultural departments financially support research that objectively evaluates qualitative assessments of behavior [e.g. Wemelsfelder et al., 2000]. Finally, King and Landau [2003], who investigated ‘‘happiness’’ or subjective well-being (SWB) in zoohoused chimpanzees stated, ‘‘yrating scales can reach into the higher realms of SWB that are not well assessed by the traditional measures’’ (p 11). Creating and validating tools that quantify caretakers’ qualitative assessments will allow for prioritizing treatment for at-risk individuals [Taylor and Mills, 2007], assessing the effectiveness of treatments and interventions [McMillan, 2000; Wojciechowska and Hewson, 2005; Wojciechowska et al., 2005b; Wiseman-Orr et al., 2006; Christiansen and Forkman, 2007; Morton, 2007; Scott et al., 2007; Taylor and Mills, 2007], making informed decisions regarding euthanasia [Wojciechowska et al., 2005b; Wiseman-Orr et al., 2006; Scott et al., 2007; Morton, 2007] and evaluating the impact of various management systems and environments [McMillan, 2000; Wemelsfelder and Lawrence, 2001; Taylor and Mills, 2007]. Developing and Testing Qualitative Assessment Tools Overview of reliability and validity testing Animal welfare researchers have stressed that tools comprised of nontraditional measures, such as those that rely on qualitative assessments, should be designed using psychometric principles, as is the case for scales created for measuring QOL in humans [Wiseman-Orr et al., 2006; Reid et al., 2007; Scott et al., 2007; Taylor and Mills, 2007; Meagher, 2009]. The psychometric approach requires that careful consideration be given to the content of the tool during development, with items and domains often being selected by consulting experts, and places a strong emphasis on testing the reliability and validity of the tool post hoc [Streiner and Norman, 2008]. Although most evidence for the reliability and validity of qualitative assessments (i.e. trait ratings) primarily has been provided by animal personality and temperament research, it has become increasingly apparent that many of the assessed traits (e.g. calm) are also useful for assessments of individual well-being [e.g. Wiseman-Orr et al., 2006]. Below, we present examples, mostly from research on zoo animals, demonstrating that assessments of traits related to individual well-being are both reliable and valid. However, it should be noted that numerous other studies on companion, laboratory, farm and some zoo animals also have provided ample evidence of the reliability and/or validity of such assessments [e.g. goats: Lyons, Zoo Biology
552 Whitham and Wielebnowski
1989; pigs: Wemelsfelder et al., 2000, 2001; cows: Rousing and Wemelsfelder, 2006; horses: Napolitano et al., 2008; Minero et al., 2009; cats: Feaver et al., 1986; dogs: Serpell and Hsu, 2001; Hsu and Serpell, 2003; Wiseman-Orr et al., 2004, 2006; Wojciechowska et al., 2005a; rhesus macaques: Stevenson-Hinde and Zunz, 1978; Stevenson-Hinde et al., 1980; Capitanio, 1999; Capitanio et al., 2004; chimpanzees: Lilienfeld et al., 1999; gorillas: Gold and Maple, 1994; elephants: Freeman et al., 2004; hyenas: Gosling, 1998]. Reliability Gosling [2001] reviewed studies that employed trait ratings to assess personality and temperament in companion, farm and exotic animals (everything from dogs to sheep to chimpanzees) and found acceptable levels of both test–retest reliability and inter-observer agreement. Test–retest reliability, or measuring the similarity between measures collected at two different time points, has been examined less frequently than inter-observer agreement. However, Gosling [2001] found substantial correlations for four of the five studies examined. To analyze interobserver agreement across diverse studies, weighted grand mean correlations were computed to evaluate both across-subjects agreement (i.e. two observers agree that animal A displays a particular trait more than animal B) using 15 species, as well as within-subject reliability (i.e. two observers agree that a particular animal is better characterized by trait A than trait B) using 9 species. The weighted grand mean correlations were 0.52 and 0.64, respectively, and Gosling [2001] argued that these are comparable to results from studies on human personality. The inter-observer agreement of ratings performed by zookeepers or experienced observers on zoo animals has been tested using a variety of statistical methods, and overall, high levels of agreement have been found. For example, in Wielebnowski’s [1999] study of cheetahs, inter-observer agreement was calculated for 18 behavioral adjectives (e.g. calm, playful), across subjects. The subjects were distributed across four breeding facilities, and for each of the 44 cheetahs, Wielebnowski and two of the cat’s keepers completed questionnaires. Inter-observer agreement for all items was calculated using Kendall’s coefficient of concordance (W) and only three items had to be excluded from further analyses due to low levels of agreement (Wo0.5 for excluded items; range of W 5 0.57–0.98 for the 15 remaining items). Carlstead et al.’s [1999] study of black rhinoceros calculated interobserver agreement within subjects, across items. Keepers and observers distributed across 19 zoos completed questionnaires for rating individual rhinoceros on specific behaviors. Before conducting statistical reliability analyses, behaviors that had been difficult to interpret or clearly had low reliability (as determined by examining absolute differences between ratings for pairs of raters) were removed. Using the remaining behaviors, inter-rater agreement for each of the 21 rhinoceros rated by 3–5 people was estimated by calculating Kendall’s coefficient of concordance (W) (range of W 5 0.31–0.88). Concordance between keepers was significant for 15 of the 21 rhinoceros. Although some studies of personality and temperament may include items that reflect well-being, King and Landau [2003] examined the reliability of ratings that specifically aimed to measure SWB for 128 chimpanzees housed across 13 zoological parks. Raters, who were either zoo employees working directly with the chimpanzees or experienced observers, completed both a personality rating form and an SWB Zoo Biology
Monitoring Zoo Animal Well-Being 553
questionnaire for each chimpanzee. For the SWB questionnaire, the raters were asked to consider four items, all of which were meant to capture a unique aspect of SWB (e.g. estimate the amount of time that the chimpanzee is happy, contented, enjoying itself or otherwise in a positive mood). Inter-rater reliabilities were estimated using intra-class correlations [see also King and Figueredo, 1997], and overall, were found to be high for the SWB items, as well as for the personality factors that resulted. In fact, King and Landau [2003] argued that the ratings of chimpanzee SWB may be more reliable than ratings of human SWB performed by proxy informants. Recently, Weiss et al.’s [2006] study of zoo-housed orangutans also used intra-class correlations to estimate inter-rater reliabilities of ratings made by experienced staff, and it was found that SWB and personality measures could be rated reliably. Interestingly, Wemelsfelder and Lawrence [2001] discovered that even when people assumed to have differing personal attitudes toward animals were asked to assess the behavioral expressions of individual pigs, without using a list of pre-fixed terms, they agreed in their general qualitative assessments. Three groups of observers, farmers/stockpeople, veterinarians and animal protectionists were asked to perform free choice profiling, which requires observers to generate their own descriptive terminologies for scoring pig behavioral expression and then score individual pigs using these terms [see Wemelsfelder et al., 2001, for a detailed description of all methodology]. Next, a multivariate statistical technique (Generalized Procrustes Analysis) was used as a pattern detection mechanism to calculate inter-observer reliability. Not only was agreement found in the way observers used terms (i.e. their scoring patterns), but there was evidence for strong semantic convergence, with observers often choosing similar terms to describe pig behavioral expression. Validity After the reliability of qualitative assessments has been demonstrated, their validity also must be tested. Generally, three types of validity are recognized: content validity, criterion validity and construct validity [Streiner and Norman, 2008; see Meagher, 2009, for a thorough discussion of the validity of observer ratings of animals]. Some degree of content validity, which considers the extent to which the items chosen for assessment describe and represent the construct of interest, can be established by consulting experts to generate and select agreed upon items for assessment [Streiner and Norman, 2008]. Yet, traditionally, many researchers have selected items based on their own knowledge of the species, by reviewing the behavioral repertoire of the species and then choosing related traits, and/or by sampling items from a related species [e.g. Feaver et al., 1986; Gold and Maple, 1994]. However, in farm and companion animal welfare research, relying on expert opinion and consensus for the development of welfare assessment tools has become increasingly common [e.g. farm: Scott et al., 2003; Whay et al., 2003; Rousing et al., 2007; companion: Holton et al., 2001]. For instance, Whay et al. [2003] employed the Delphi technique, a technique frequently applied in human health-related research, that involved consulting a panel of experts (each panel representing a particular farm animal species) to help identify the most appropriate measures for the assessment of well-being [see also Anonymous, 2001]. First, the selected experts completed a Zoo Biology
554 Whitham and Wielebnowski
questionnaire that instructed them to identify at least five issues they believed to influence welfare status in the species being represented. Next, they listed three animal-based measures for each issue and then rated ‘‘how good’’ and ‘‘how practical’’ each measure would be for on-farm assessment. All proposed information was compiled into a second questionnaire and re-distributed to the panel, who then rated the relative importance of each major welfare measure category. The panel was also given the opportunity to propose additional categories and rank the five measures they believed to be most important. Based on this two-step questionnaire process, Whay et al. [2003] were able to create appropriate, animal-based protocols for on-farm welfare assessments for dairy cattle, pigs and hens. Establishing the criterion validity of trait ratings currently is not feasible for studies of animal well-being, because this type of validation requires a ‘‘gold standard’’ of measurement against which a tool’s effectiveness at measuring a construct can be compared. Although some established disciplines have been able to identify useful gold standards for validating certain tests or tools (e.g. career/aptitude tests in the field of human psychology), no such standard measure or well-established tool currently exists for rating traits related to animal well-being. Therefore, it is more common and feasible to demonstrate construct validity, which usually involves comparing the tool that is being developed to existing tests that are known to measure similar constructs, even though these tests have not been identified as the best possible measure or gold standard for the construct of interest. Gosling [2001] provided an extensive review of studies that have tested for construct validity by correlating trait ratings with behavioral measures (observations or behavioral tests) and/or physiological measures. Most commonly, trait ratings are validated using data from behavioral observations. King and Landau [2003] demonstrated the construct validity of their SWB questionnaire by observing 49 chimpanzees using instantaneous sampling during behavioral observations. The authors found that the four items they used to assess SWB formed a single ‘‘SWB’’ factor and that SWB scores significantly and negatively correlated with standardized scores for behaviors that occurred in a submissive context. However, the behavioral observations and ratings were performed by the same people, which generally should be avoided in order to ensure the independence of measures [Gosling, 2001]. Ratings of personality and temperament also correlate with animals’ responses to behavioral tests. Wielebnowski [1999] exposed cheetahs to a mirror [the mirrorimage stimulation test, Gallup, 1968] and found relationships between behavioral responses and ratings of behavioral adjectives. For example, individuals who were slow to initially approach the mirror received high ratings on items such as fearful of conspecifics and tense, whereas individuals who approached more rapidly received higher ratings on items such as calm, curious and self-assured. Similarly, Carlstead et al. [1999] performed standardized behavioral tests on black rhinoceros and found that keepers’ scores of all behavior traits (behaviors from the initial questionnaire were grouped into six behavior traits based on intercorrelations) were associated with behavioral reactivity to novel stimuli. For instance, individuals with high fear scores took more time to approach a novel scent and were less likely to interact with a novel object (traffic cone) than rhinos rated as less fearful. It is important to note that in this study, ratings were completed by keepers, curators and observers, whereas the behavior tests were performed by outside researchers, ensuring the independence of measures [Carlstead et al., 1999]. Zoo Biology
Monitoring Zoo Animal Well-Being 555
There is also evidence that ratings correlate with physiological measures. In Wielebnowski et al.’s [2002] study of clouded leopards, keepers rated individuals on a series of behavioral questions (e.g. how fearful does this individual react toward keepers?), and ratings were found to be associated with fecal corticoid metabolite concentrations. For example, ratings of the behavior tense were positively correlated with mean overall, baseline and peak fecal corticoid metabolite concentrations. Similarly, in Carlstead and Brown’s [2005] study of black and white rhinoceros, keepers’ scores of behavior traits correlated with mean fecal corticoid metabolite concentrations and corticoid variability. For example, for white rhinos, the behavior trait friendly to keeper was negatively associated with mean corticoid metabolite concentrations. Finally, Capitanio et al.’s [1999] study of laboratory-housed rhesus macaques is noteworthy as it demonstrates that trait ratings may be predictive of physiological measures that reflect an individual’s health status. After being inoculated with the simian immunodeficiency virus, individuals rated higher on sociability showed a faster decline in plasma cortisol concentrations, a stronger antibody response and lower levels of viral load than those rated lower on sociability. In summary, there is substantial evidence suggesting that animal caretakers’ assessments of traits that reflect individual well-being are both reliable and valid. Integrating such assessments into regular monitoring of individual well-being, as promoted by laboratory, companion and farm animal welfare researchers [e.g. Morton, 2000; Wiseman-Orr et al., 2006; Scott et al., 2007], would provide a major step toward more comprehensive institutional welfare assessments. Therefore, we recommend that zoological institutions systematically apply the expertise of animal care staff to develop appropriate, species-specific, animal-based tools that allow quantification of qualitative assessments for ongoing monitoring of individual wellbeing. Below, we outline a process by which this can be accomplished. A Process for Development, Validation and Testing of a Monitoring Tool Prototype At the Chicago Zoological Society (CZS) we recently embarked on the process of developing, validating and testing a low cost, animal-based monitoring tool that allows us to quantify keepers’ qualitative assessments of individual well-being. With the expertise of animal care staff, we developed 12 species-specific Welfare Score Sheets for ongoing monitoring. We will briefly describe our three-step process that is currently being tested and evaluated. Step 1: Development of species-specific Welfare Score Sheets So far, we have developed Welfare Score Sheets for the following species: aardvark, African elephant, black rhino, clouded leopard, fennec fox, Goeldi’s monkey, green-winged macaw, leopard gecko, okapi, polar bear, red-tailed hawk and western lowland gorilla. Species were chosen based on their availability for regular data collection, the welfare concerns of managers and an attempt to represent various taxa. To design appropriate score sheets for monitoring individual well-being in each study species, we employed the Delphi technique, a method that has been used successfully by farm animal welfare researchers, to identify the most useful measures for welfare assessment [Anonymous, 2001; Whay et al., 2003] (see Validity). First, for each species we chose at least three animal care experts (full-time Zoo Biology
556 Whitham and Wielebnowski
keepers/managers, some from other institutions, with at least 3 years of experience providing daily care to the species of interest) to complete a detailed, three-part Welfare Assessment Questionnaire. Experts were asked: (1) several open-ended questions regarding welfare (e.g. how do you determine whether an individual is comfortable?), (2) to list and define the most useful indicators of poor well-being, good/great well-being and change in welfare status, and (3) to rate qualitative traits from previous studies of animal personality, temperament and well-being on their usefulness for monitoring changes in individual well-being (e.g. stressed, playful). We compiled all responses, and then to ensure agreement, experts reviewed and edited all measures and definitions proposed by those representing the same species. Finally, by combining agreed upon measures and definitions with highly rated qualitative traits, we designed species-specific Welfare Score Sheets consisting of approximately 10 items presented on a 5-point Likert scale. Step 2: Cross-validating specific items from the Welfare Score Sheets For a 6-month period, currently underway, study animals are being rated on a weekly basis by two to four full-time, experienced keepers (i.e. keepers who work with the individuals/species on a regular basis and have done so for several years) using the Welfare Score Sheets. We also are conducting concurrent quantitative behavioral and fecal glucocorticoid monitoring, as well as compiling regularly collected health/ medical/nutrition data (e.g. weight, consumption charts). Study animals are focally observed three times per week in 10-minute sessions using: (1) instantaneous sampling of states (e.g. pacing) at 1-minute intervals, and (2) all-occurrence sampling of events (e.g. vocalizations) [Martin and Bateson, 1993]. Fecal samples are collected from each animal three times per week for subsequent glucocorticoid metabolite analyses. Once data collection has been completed we will use concordance analyses to test for keeper agreement [Sokal and Rohlf, 1995] in order to determine which score sheet items to include in further statistical analyses. To validate reliable items, we will test for correlations between average weekly score sheet ratings, average weekly fecal glucocorticoid metabolite concentrations, average weekly proportion of time spent in relevant behavioral states, average frequency of relevant behavioral events and available health/medical/nutrition data. If significant correlations are found for particular items (thus helping to establish construct validity), this will provide evidence that keepers’ ratings of well-being measures can serve as a first indicator of underlying physical, physiological and/or behavioral problems reflecting changes in welfare status. Also, score sheets will be refined by eliminating items with low keeper agreement and/or no cross-validation. Step 3: Testing score sheet management application During this phase, we will determine how score sheets can be effectively integrated into daily management practices to proactively identify and address welfare issues. Specifically, we will help managers and keepers use score sheets during regular workgroup meetings to flag cases of apparent changes in well-being, identify management issues that may be reducing welfare and evaluate whether interventions to resolve these issues are successful. Individuals will be flagged if: (1) scores change frequently, dramatically or cumulatively, (2) scores consistently decline/increase (more than three weeks in a row), or (3) sudden keeper disagreement occurs on items previously showing high agreement. We will bring these cases to the attention of the Zoo Biology
Monitoring Zoo Animal Well-Being 557
managers, who will meet with keepers to discuss specific management issues that may be impacting well-being and to propose feasible changes in the environment or husbandry routine. We will continue monitoring as changes are applied, reporting our findings to the workgroups, to assess whether scores improve and are correlated to measurable behavioral, physiological and/or physical changes. If both validation and testing of the Welfare Score Sheets are successful, we expect the sheets to be applicable at most other zoological institutions. We hope that our methods of tool development and validation may serve as a ‘‘blueprint’’ for other species and institutions. CONCLUSIONS 1. Identifying effective and user-friendly welfare assessment tools is becoming a priority for many zoos and aquariums worldwide. 2. Institutional welfare assessments and internal welfare-monitoring processes already are being conducted in some farm and lab animal industries, and are increasingly implemented at zoos and aquariums in the UK. 3. To provide appropriate assessments of individual well-being at any given institution, animal-based welfare assessments should be used in combination with any existing resource-based assessments. 4. Animal-based measures can be integrated systematically into ongoing institutional welfare-monitoring processes by validating assessments based on keeper expertise and experience against quantiative measures. 5. Simple scoring tools, once carefully developed and validated for a given species, may be employed at many institutions for long-term monitoring of individual animal well-being and integrated into routine management practices. ACKNOWLEDGMENTS We are extremely grateful to all managers and keepers who have contributed to the development process of the welfare-monitoring tool prototype. Specifically, we thank our expert consultants for completing the Welfare Assessment Questionnaire, which allowed us to design appropriate, species-specific Welfare Score Sheets. Finally, we thank the Chicago Board of Trade Endangered Species Fund and the Women’s Board of the Chicago Zoological Society for providing funding for various aspects of the tool development and validation steps. REFERENCES Addington-Hall J, Kalra L. 2001. Measuring quality of life: who should measure quality of life? Br Med J 322:1417–1420. Anonymous. 2001. Scientists’ assessment of the impact of housing and management on animal welfare. J Appl Anim Welfare Sci 4:3–52. Barber J. 2009. Programmatic approaches to assessing and improving animal welfare in zoos and aquariums. Zoo Biol, this issue. Block JH. 1977. Advancing the psychology of personality: paradigmatic shifts or improving the quality of research? In: Magnusson D,
Endler NS, editors. Personality at the crossroads: current issues in interactional psychology. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. p 37–63. Boissy A, Manteuffel G, Jensen MB, Moe RO, Spruijt B, Keeling LJ, Winckler C, Forkman B, Dimitrov I, Langbein J, Bakken M, Veissier I, Aubert A. 2007. Assessment of positive emotions in animals to improve their welfare. Physiol Behav 92:375–397. Broom DM. 1986. Indicators of poor welfare. Br Vet J 142:524–526.
Zoo Biology
558 Whitham and Wielebnowski Bryan S, Hardyman W, Bentham P, Buckley A, Laight A. 2005. Proxy completion of EQ-5D in patients with dementia. Qual Life Res 14:107–118. Burghardt GM. 1985. Animal awareness: current perceptions and historical perspective. Am Psychol 40:905–919. Burghardt GM. 1991. Cognitive ethology and critical anthropomorphism: a snake with two heads and hognose snakes that play dead. In: Ristau CA, editor. Cognitive ethology: the minds of other animals. San Francisco: Lawrence Erlbaum Associates, Inc. p 53–90. Capitanio JP. 1999. Personality dimensions in adult male rhesus macaques: prediction of behaviors across time and situation. Am J Primatol 47:299–320. Capitanio JP, Mendoza SP, Baroncelli S. 1999. The relationship of personality dimensions in adult male rhesus macaques to progression of Simian Immunodeficiency Virus disease. Brain Behav Immun 13:138–154. Capitanio JP, Mendoza SP, Bentson KL. 2004. Personality characteristics and basal cortisol concentrations in adult male rhesus macaques (Macaca mulatta). Psychoneuroendocrinology 29:1300–1308. Carlstead K, Brown JL. 2005. Relationships between patterns of fecal corticoid excretion and behavior, reproduction, and environmental factors in captive black (Diceros bicornis) and white (Ceratotherium simum) rhinoceros. Zoo Biol 24:215–232. Carlstead K, Mellen J, Kleiman DG. 1999. Black rhinoceros (Diceros bicornis) in U.S. zoos: I. Individual behavior profiles and their relationship to breeding success. Zoo Biol 18:17–34. Christiansen SB, Forkman B. 2007. Assessment of animal welfare in a veterinary context—a call for ethologists. Appl Anim Behav Sci 106:203–220. Clark JD, Rager DR, Calpin JP. 1997a. Animal well-being I. General considerations. Lab Anim Sci 47:564–570. Clark JD, Rager DR, Calpin JP. 1997b. Animal well-being II. Stress and distress. Lab Anim Sci 47:571–579. Dawkins MS. 1990. From an animal’s point of view: motivation, fitness, and animal welfare. Behav Brain Sci 13:1–61. Department for Environment Food and Rural Affairs [DEFRA]. 2007. Zoos Forum Handbook. Chapter 4: Animal welfare and its assessment in zoos. hhttp://www.defra.gov.uk/ wildlife-countryside /pdf/protection/zoo-handbook/ 4.pdfi. Accessed 2009. Diener E, Larsen RJ. 1993. The experience of emotional well-being. In: Lewis M, Haviland JM, editors. Handbook of emotions. New York: The Guilford Press. p 405–415. Duncan IJH. 1996. Animal welfare defined in terms of feelings. Acta Agric Scand Sect A, Animal Sci Suppl 27:29–35. Duncan IJH. 2006. The changing concept of animal sentience. Appl Anim Behav Sci 100:11–19.
Zoo Biology
Duncan IJH, Dawkins MS. 1983. The problem of assessing ‘‘well-being’’ and ‘‘suffering’’ in farm animals. In: Smidt D, editor. Indicators relevant to farm animal welfare. The Hague: Martinus Nijhoff Publishers. p 13–24. Eiser C, Morse R. 2001. Quality of life measures in chronic diseases of childhood. Health Tech Assess 5:1–156. Farm Animal Welfare Council [FAWC]. 2005. Report on the welfare implications of Farm Assurance Schemes. London: FAWC. Feaver J, Mendl M, Bateson P. 1986. A method for rating the individual distinctiveness of domestic cats. Anim Behav 34:1016–1025. Fraser D. 1995. Science, values and animal welfare: exploring the ‘‘inextricable connection’’. Anim Welfare 4:103–117. Freeman EW, Weiss E, Brown JL. 2004. Examination of the interrelationships of behavior, dominance status, and ovarian activity in captive Asian and African elephants. Zoo Biol 23:431–448. Gallup Jr GG. 1968. Mirror-image stimulation. Psychol Bull 70:782–793. Gladwell M. 2007. Blink: the power of thinking without thinking. Boston: Back Bay Books. 320p. Gold KC, Maple TL. 1994. Personality assessment in the gorilla and its utility as a management tool. Zoo Biol 13:509–522. Gonyou HW. 1993. Animal welfare: definitions and assessment. J Agr Environ Ethics 6:37–43. Gosling SD. 1998. Personality dimensions in spotted hyenas (Crocuta crocuta). J Comp Psychol 112:107–118. Gosling SD. 2001. From mice to men: what can we learn about personality from animal research? Psychol Bull 127:45–86. Hawkins P. 2002. Recognizing and assessing pain, suffering and distress in laboratory animals. Lab Anim 36:378–395. Hewson CJ. 2003. Can we assess welfare? Can Vet J 44:749–753. Holton L, Pawson P, Nolan A, Reid J, Scott EM. 2001. Development of a behaviour-based scale to measure acute pain in dogs. Vet Rec 148:525–531. Hsu Y, Serpell JA. 2003. Development and validation of a questionnaire for measuring behavior and temperament traits in pet dogs. J Am Vet Med Assoc 223:1293–1300. King JE, Figueredo AJ. 1997. The five-factor model plus dominance in chimpanzee personality. J Res Pers 31:257–271. King JE, Landau VI. 2003. Can chimpanzee (Pan troglodytes) happiness be estimated by human raters? J Res Pers 37:1–15. Li L, Young D, Wei H, Zhang Y, Zheng Y, Xiao S, Wang X, Chen X. 1998. The relationship between objective life status and subjective life satisfaction with quality of life. Behav Med 23:149–159. Lilienfeld SO, Gershon J, Duke M, Marino L, de Waal FBM. 1999. A preliminary investigation of
Monitoring Zoo Animal Well-Being 559 the construct of psychopathic personality (psychopathy) in chimpanzees (Pan troglodytes). J Comp Psychol 113:365–375. Lykken D, Tellegen A. 1996. Happiness is a stochastic phenomenon. Psychol Sci 7:186–189. Lyons DM. 1989. Individual differences in temperament of dairy goats and the inhibition of milk ejection. Appl Anim Behav Sci 22:269–282. Main DCJ, Whay HR, Leeb C, Webster AJF. 2007. Formal animal-based welfare assessment in UK certification schemes. Anim Welfare 16:233–236. Martin P, Bateson P. 1993. Measuring behaviour: an introductory guide, 2nd ed. Cambridge, UK: Cambridge University Press. 222 p. McMillan FD. 2000. Quality of life in animals. J Am Vet Med Assoc 216:1904–1910. McMillan FD. 2003. Maximizing quality of life in ill animals. J Am Anim Hosp Assoc 39:227–235. Meagher RK. 2009. Observer ratings: validity and value as a tool for animal welfare research. Appl Anim Behav Sci 119:1–14. Mench JA. 1998. Thirty years after Brambell: whither animal welfare science? J Appl Anim Welfare Sci 1:91–102. Minero M, Tosi MV, Canali E, Wemelsfelder F. 2009. Quantitative and qualitative assessment of the response of foals to the presence of an unfamiliar human. Appl Anim Behav Sci 116:74–81. Morton DB. 2000. A systematic approach for establishing humane endpoints. ILAR J 41:80–86. Morton DB. 2007. A hypothetical strategy for the objective evaluation of animal well being and quality of life using a dog model. Anim Welfare 16:75–81. Napolitano F, De Rosa G, Braghieri A, Grasso F, Bordi A, Wemelsfelder F. 2008. The qualitative assessment of responsiveness to environmental challenge in horses and ponies. Appl Anim Behav Sci 109:342–354. Reid J, Nolan AM, Hughes JML, Lascelles D, Pawson P, Scott EM. 2007. Development of the short-form Glasgow Composite Measure Pain Scale (CMPS-SF) and derivation of an analgesic intervention score. Anim Welfare 16:97–104. Rousing T, Wemelsfelder F. 2006. Qualitative assessment of social behaviour of dairy cows housed in loose housing systems. Appl Anim Behav Sci 101:40–53. Rousing T, Jakobsen IA, Hindhede J, Klaas IC, Bonde M, Sørensen JT. 2007. Evaluation of a welfare indicator protocol for assessing animal welfare in AMS herds: researcher, production advisor and veterinary practitioner opinion. Anim Welfare 16:213–216. Sadler-Smith E, Shefy E. 2004. The intuitive executive: understanding and applying ‘‘gut feel’’ in decision-making. Acad Manag Exec 18:76–91. Sandøe P. 1996. Animal and human welfare–are they the same kind of thing? Acta Agric Scand Sect A, Animal Sci Suppl 27:11–15. Scott EM, Fitzpatrick JL, Nolan AM, Reid J, Wiseman ML. 2003. Evaluation of welfare state
based on interpretation of multiple indices. Anim Welfare 12:457–468. Scott EM, Nolan AM, Reid J, Wiseman-Orr ML. 2007. Can we really measure animal quality of life? Methodologies for measuring quality of life in people and other animals. Anim Welfare 16:17–24. Serpell JA, Hsu Y. 2001. Development and validation of a novel method for evaluating behavior and temperament in guide dogs. Appl Anim Behav Sci 72:347–364. Sokal RR, Rohlf FJ. 1995. Biometry, 3rd ed. New York: WH Freeman Company. 88p. Sprangers MAG, Aaronson NK. 1992. The role of health care providers and significant others in evaluating the quality of life of patients with chronic disease: a review. J Clin Epidemiol 45:743–760. Stevenson-Hinde J, Zunz M. 1978. Subjective assessment of individual rhesus monkeys. Primates 19:473–482. Stevenson-Hinde J, Stillwell-Barnes R, Zunz M. 1980. Subjective assessment of rhesus monkeys over four successive years. Primates 21:66–82. Streiner DL, Norman GR. 2008. Health measurement scales: a practical guide to their development and use, 4th ed. New York: Oxford University Press. 428p. Taylor KD, Mills DS. 2007. Is quality of life a useful concept for companion animals? Anim Welfare 16:55–65. Timmins RP, Cliff KD, Day CT, Hart BL, Hart LA, Hubrecht RC, Hurley KF, Phillips CJC, Rand JS, Rochlitz I, Serpell JA, Zawistowski SL. 2007. Enhancing quality of life for dogs and cats in confined situations. Anim Welfare 16:83–87. Weiss A, King JE, Perkins L. 2006. Personality and subjective well-being in orangutans (Pongo pygmaeus and Pongo abelii). J Pers Soc Psychol 90:501–511. Wemelsfelder F. 1997. The scientific validity of subjective concepts in models of animal welfare. Appl Anim Behav Sci 53:75–88. Wemelsfelder F. 2007. How animals communicate quality of life: the qualitative assessment of behaviour. Anim Welfare 16:25–31. Wemelsfelder F, Lawrence AB. 2001. Qualitative assessment of animal behaviour as an on-farm welfare-monitoring tool. Acta Agric Scand Sect A, Animal Sci Suppl 30:21–25. Wemelsfelder F, Hunter EA, Mendl MT, Lawrence AB. 2000. The spontaneous qualitative assessment of behavioural expressions in pigs: first explorations of a novel methodology for integrative animal welfare measurement. J Appl Anim Welfare Sci 67:193–215. Wemelsfelder F, Hunter TEA, Mendl MT, Lawrence AB. 2001. Assessing the ‘‘whole animal’’: a free choice profiling approach. Anim Behav 62:209–220. Whay HR. 2007. The journey to animal welfare improvement. Anim Welfare 16:117–122. Whay HR, Main DCJ, Green LE, Webster AJF. 2003. Animal-based measures for the assessment of welfare state of dairy cattle, pigs and laying
Zoo Biology
560 Whitham and Wielebnowski hens: consensus of expert opinion. Anim Welfare 12:205–217. Wielebnowski NC. 1999. Behavioral differences as predictors of breeding status in captive cheetahs. Zoo Biol 18:335–349. Wielebnowski NC, Fletchall N, Carlstead K, Busso JM, Brown JL. 2002. Noninvasive assessment of adrenal activity associated with husbandry and behavioral factors in the North American clouded leopard population. Zoo Biol 21:77–98. Wiseman-Orr ML, Nolan AM, Reid J, Scott EM. 2004. Development of a questionnaire to measure the effects of chronic pain on health-related quality of life in dogs. Am J Vet Res 65:1077–1084. Wiseman-Orr ML, Scott EM, Reid J, Nolan AM. 2006. Validation of a structured questionnaire as an instrument to measure chronic pain in dogs on the basis of effects on health-related quality of life. Am J Vet Res 67:1826–1836.
Zoo Biology
Wojciechowska JI, Hewson CJ. 2005. Quality-oflife assessment in pet dogs. J Am Vet Med Assoc 226:722–728. Wojciechowska JI, Hewson CJ, Stryhn H, Guy NC, Patronek GJ, Timmons V. 2005a. Development of a discriminative questionnaire to assess nonphysical aspects of quality of life of dogs. Am J Vet Res 66:1453–1460. Wojciechowska JI, Hewson CJ, Stryhn H, Guy NC, Patronek GJ, Timmons V. 2005b. Evaluation of a questionnaire regarding nonphysical aspects of quality of life in sick and healthy dogs. Am J Vet Res 66:1461–1467. World Association of Zoos and Aquariums [WAZA]. 2005. Building a future for wildlife— the world zoo and aquarium conservation strategy. Bern: WAZA Executive Office. Yeates JW, Main DCJ. 2008. Assessment of positive welfare: a review. Vet J 175:293–300.