Introduction
Across contract research organisation (CRO) facilities it can be difficult to have uniform harmonisation of animal facility management and associated Standard Operating Procedures (SOPs) due to differences in client needs and study types, regional or national oversight body requirements, availability of resources for facility design and equipment, and differences in cultural practices across sites (Underwood Reference Underwood2007). Some harmonisation of animal care practices can be accomplished through local oversight bodies, such as animal ethics committees or through international third party animal welfare assessment organisations, such as AAALAC, International (Kendall et al. Reference Kendall, Petervary, Bergdall, Page and Baneux2018; Bayne & Turner Reference Bayne and Turner2019; Mohan & Huneke Reference Mohan and Huneke2019). However, these committees and organisations largely provide guidance and do not go into detail concerning welfare assessment or behavioural management programmes (Turner & Bayne Reference Turner and Bayne2023). This can result in a wide range of behavioural management practices and programmes and asynchronous improvements, which may contribute to variability in the scientific data outcomes as well as animal welfare (Everitt & Berridge Reference Everitt and Berridge2017).
Approximately 100,000 primates are used in biomedical research around the world every year not including primates kept for breeding purposes (Lankau et al. Reference Lankau, Turner, Mullan and Galland2014). The most common research primate species are cynomolgus macaques (Macaca fascicularis), rhesus macaques (Macaca mulatta), and common marmosets (Callithrix jacchus) (Lankau et al. Reference Lankau, Turner, Mullan and Galland2014; Paterson & Turner Reference Paterson and Turner2022). Research primates across facilities may live in different housing conditions and undergo various study- and veterinary-related procedures (Carlsson et al. Reference Carlsson, Schapiro, Farah and Hau2004; Johnsen et al. Reference Johnsen, Johnson, Whitney, Abee, Mansfield, Tardif and Morris2012; Wolf & White Reference Wolf, White, Abee, Mansfield, Tardif and Morris2012). It is important to have accurate and ongoing assessments of the individual and group animal welfare states in these different circumstances to fulfil moral, legal, and scientific obligations (Turner Reference Turner2020; Canadian Council on Animal Care [CCAC] 2021).
When assessing animal welfare, measures should be inspired by the framework of “a life worth living”, which incorporates positive welfare indicators (Mellor Reference Mellor2016; Nunamaker et al. Reference Nunamaker, Davis, O’Malley and Turner2021). Primates have complex requirements to meet their behavioural management needs, such as appropriate biological health and functioning, positive affective states, and the ability to display natural behaviours (Fraser et al. Reference Fraser, Weary, Pajor and Milligan1997; Howell & Cheyne Reference Howell and Cheyne2019; Testard et al. Reference Testard, Tremblay and Platt2021). Under natural conditions, many primate species, such as macaques and marmosets, live in small- to medium-sized social groups in which there is an established hierarchy and behavioural synchronisation (Lehmann et al. Reference Lehmann, Korstjens and Dunbar2007). These interactions can impact access to resources, such as food (Hambali et al. Reference Hambali, Ismail and Munir Md-Zain2012), and complex social behaviours, such as huddling during rest that can affect overall energy expenditure for thermoregulation (Campbell et al. Reference Campbell, Tkaczynski, Lehmann, Mouna and Majolo2018). Human-primate interactions can also impact welfare. Implementing behaviour-focused practices, such as positive reinforcement training (PRT) contributes to improved affective state (Prescott & Buchanan-Smith Reference Prescott and Buchanan-Smith2003; Perlman et al. Reference Perlman, Bloomsmith, Whittaker, McMillan, Minier and McCowan2012; Turner & Bayne Reference Turner and Bayne2023). PRT is a form of positive human-animal interaction that allows animals some choice and control, and its implementation can reduce overall daily stress related to husbandry and study procedures (Schapiro et al. Reference Schapiro, Perlman, Thiele and Lambeth2005). In a study looking at PRT in chimpanzees (Pan troglodytes), animals who voluntarily presented for an anaesthetic injection had significantly lower white blood cell counts, absolute segmented neutrophil counts, and glucose levels indicative of less stress compared to the chimpanzees anaesthetised using traditional methods (Lambeth et al. Reference Lambeth, Hau, Perlman, Martino and Schapiro2006).
Welfare assessments are intended as a holistic evaluation of the impact of animal behaviour management and care programmes and underpin a continuous quality improvement model (i.e. plan-do-check-act). The assessments are different from daily observations that may be conducted as part of animal husbandry and humane intervention point assessments (CCAC 2021; Nunamaker et al. Reference Nunamaker, Davis, O’Malley and Turner2021) and ideally are formalised and documented, with definition of action plans. Animal welfare assessment should consider input- and outcome-based measures (Barnett & Hemsworth Reference Barnett and Hemsworth2009; CCAC 2019, 2021). Input-based measures focus on the resources that animals are provided (Vasseur et al. Reference Vasseur, Pellerin, de Passillé, Winckler, Lensink, Knierim and Rushen2012). These measures are reliable and have good inter- and intra-rater reliability; however, they can lack validity. Input-based measures are an indirect reflection of animal welfare as individual animals will perceive the environment differently, leading to a spectrum of welfare states (Vasseur et al. Reference Vasseur, Pellerin, de Passillé, Winckler, Lensink, Knierim and Rushen2012; Prescott et al. Reference Prescott, Leach and Truelove2022). For example, providing the adequate resources in the environment to encourage species-typical behaviour. Outcome-based measures are a direct reflection of one’s state (Vasseur et al. Reference Vasseur, Pellerin, de Passillé, Winckler, Lensink, Knierim and Rushen2012; Prescott et al. Reference Prescott, Leach and Truelove2022). For example, quantifying the use of the provided resources through behavioural observations. Another important outcome-based measure at the human level in a laboratory setting is the culture of care (Klein & Bayne Reference Klein and Bayne2007; Robinson et al. Reference Robinson, Sparrow, Williams, Decelle, Bertelsen, Reid and Chlebus2019; Bayne & Turner Reference Bayne and Turner2019). Having a good culture of care means that the institutional team (e.g. caregivers and technical personnel, researchers, animal ethics committee members, veterinarians, managers, etc) understand the importance of their work, and they are working together towards a common goal of promoting high-quality science and proactive animal care based on science-based performance standards that exceed basic regulatory requirements, and that the employees working directly with the animals are satisfied in their job in items such as comprehensive training to be able to care for animals well, work-life balance to avoid stress and rushing, resources related to compassion fatigue and resiliency building to promote good mental health for employees, and that employees feel valued and heard in their role of promoting good animal welfare (Bayne & Turner Reference Bayne and Turner2019; Robinson et al. Reference Robinson, Sparrow, Williams, Decelle, Bertelsen, Reid and Chlebus2019).
Welfare assessment tools have been developed for specific species such as elephants (Yon et al. Reference Yon, Williams, Harvey and Asher2019), horses (Long et al. Reference Long, Dürnberger, Jenner, Kelemen, Auer and Grimm2022), reptiles (Benn et al. Reference Benn, McLelland and Whittaker2019), for animals housed in zoos (Sherwen et al. Reference Sherwen, Hemsworth, Beausolei, Embury and Mellor2018), for livestock species in different stages of production (see, for example, Kirchner et al. Reference Kirchner, Schulze Westerath, Knierim, Tessitore, Cozzi, Pfeiffer and Winckler2014; Buijs et al. Reference Buijs, Ampe and Tuyttens2017; Kang et al. Reference Kang, Bae and Lee2022), but only recently has attention turned towards primates in biomedical research settings (Truelove et al. Reference Truelove, Martin, Langford and Leach2020; Prescott et al. Reference Prescott, Leach and Truelove2022). The main purpose of an animal welfare assessment tool is to have an objective and quantitative measure of animal welfare that permits regular assessment with the goal of taking action to improve animal welfare (Honess & Wolfensohn Reference Honess and Wolfensohn2010; CCAC 2021).
The goal of this project was to develop a novel primate welfare assessment tool that could be applied across sites and countries to assess research primate (predominantly macaque) welfare at the facility level, and globally, refine primate welfare through knowledge of ‘current state’ and creating a culture of continuous improvement. The aim of implementing the tool is to provide a means for formalising primate welfare assessments and to harmonise considerations and approaches to primate care and welfare across sites and countries, regardless of intended research primate use.
Materials and methods
Focus group development
In May 2020, members of Charles River’s animal welfare oversight group, Global Animal Welfare and Training (GAW&T), hosted a two-day global internal and virtual 3Rs primate workshop. A significant deliverable of this workshop was to develop a Primate Welfare Assessment Tool (PWAT). Internal primate experts (i.e. primate behaviourists, primate behaviour champions, and veterinarians) were invited to participate on the project. Ten individuals were recruited to form a focus group from across sites in Canada (n = 4), Europe (n = 2), and the United States (n = 4). Two of the participants were the focus group co-leaders (EAP, PVT) with > 6 and > 28 years experience, respectively, of working with macaques in various research settings. After agreeing upon a charter, the focus group met virtually every three weeks for approximately eight months.
Preliminary tool development
The preliminary tool was based on an extensive list of animal welfare descriptors to be used in a zoo setting as well as other settings, such as pigs on-farm (Association of Zoos and Aquaria [https://www.aza.org/accred-materials]; Kagan et al. Reference Kagan, Carter and Allard2015; Courboulay et al. Reference Courboulay, Meunier-Salun, Stankowiak and Pol2020). The two main researchers (EAP, PVT) refined the list of descriptors based on the current literature and experience with macaques in a range of research facilities, including primate quarantine and procurement, discovery or exploratory settings, and safety assessment. The refined list of descriptors were categorised into physical, behavioural, environmental, training, procedural, and culture of care, which were further divided into subcategories (categories, category aims, subcategories, and maximum score that could be achieved per subcategory are provided in Table 1. For the full list of descriptors by category and subcategory, see Table 2. Every descriptor was attributed a weighted score from 1 to 5 based on the degree of impact on primate welfare based on the literature, with 1 being a very low impact on welfare and 5 very high (welfare weights per descriptor are provided in Table 2). A numerical scale was created for scoring, which included scores of 0, 1, and 2. A score of 0 was representative of something or an activity that was rarely present (< 25%), 1 was indicative of something moderately present (25–75%) and a score of 2 was representative of an item or activity that was obviously present (> 75%). A nominal scale was created for descriptors for which ‘yes’ (2) or ‘no’ (0) responses were applicable. A ‘non-applicable’ option was also available for certain descriptors and, if selected, the points from that descriptor were not added to the final score. Non-applicable was only used if the descriptor was not physically possible (e.g. maternal-offspring rearing opportunities at a facility in which no animal breeding occurred) or could not be evaluated during the assessment (e.g. assessment of animal behaviour during procedure when no procedures were occurring). Some descriptors were provided as bonus points if they were harder to achieve but were important for animal welfare (e.g. access to outdoor spaces, use of remote animal monitoring systems). One bonus descriptor subtracted points from the final score based on the number of singly housed animals within a room (maximum score reduction was set at 20 points). The tool was first created in Microsoft Excel® (Microsoft® Corporation, Redmond, Washington, USA, version 2019). The preliminary tool was presented to the focus group and consensus was achieved on the descriptors, how descriptors were best categorised, the welfare weight of each descriptor, and the numerical/nominal scale for each descriptor. The group suggested applying the tool every six months – a time-frame that would allow sites to select and make progress on several findings while not so long that the tool and its purpose could be forgotten.
Aim and description of each category and subcategory included in the Primate Welfare Assessment Tool, as well as the maximum score, calculated by multiplying the weight of the descriptor with the highest score possible for that indicator
Full list of Primate Welfare Assessment Tool descriptors, including welfare weight (1–5, with 1 being low impact on welfare and 5 being high impact on welfare) and assessment form (room, site, or culture of care personnel interviews). The table is divided by category: 2(a) Physical table, 2(b) Behaviour, 2(c) Environmental, 2(d) Training, 2(e) Procedural and 2(f) Culture of care
Beta-testing
The preliminary tool was beta-tested at six different facilities (two sites in Canada, one in UK, one in France, and one in the US) in December 2020 by six different participants. All facilities met or exceeded country-specific animal regulations and legislation and all were accredited by AAALAC International. Participants at small facilities reported that the assessment took approximately 2 h to complete while participants at larger facilities reported that it took 5–8 h. The most important and universal feedback from participants during beta-testing was that input- and outcome-based measures were hard to obtain when combined in the categories and that users wanted an easier platform for the assessment that minimised lateral scrolling. To address this challenge, the tool was divided into three assessment forms including room-level, in which animal outcome-based measures are evaluated, site-level, in which records reviews and input-based measured are evaluated, and a culture of care assessment, in which facility personnel feedback was sought and evaluated via an anonymous survey. Materials were translated into French to allow for multi-language participation.
Pilot launch of finalised tool and implementation
A finalised tool was created in Smartsheet® (Smartsheet Inc, Bellevue, WA, USA, version 2020), a practical platform allowing for easy data collection and automated tabulation and analysis of results. The answers from the three forms were collected into data-sheets with equations formatted to multiply answers by the designated welfare weight. The raw data were then referenced to metric sheets per category where the data were summed to give a total score for each category and subcategory. The summations were then referenced to dashboards where data were visualised as graphs for each facility and presented to senior global management for approval. The finalised tool in Smartsheet® was piloted at one facility in each of Canada and France. After focus group discussion and pilot-testing, several descriptors were eliminated due to redundancy or reworded for clarity. The final tool had 133 descriptors (Table 2). In March (Q1) and September (Q3) 2022, the PWAT was launched to 13 sites globally and each site was given 30 days to complete all forms. Sites were responsible for selecting whom would complete the assessment at their facility. Sites were asked to assess approximately10% of their occupied primate rooms (minimum of three rooms, maximum of ten depending on total number of occupied primate rooms) including a variety of housing types, animal purposes, study types, and study lengths, with one form submitted per room. Sites were also asked to submit one form for the site-level assessment, and to conduct 3–6 culture of care interviews depending on number of personnel working with primates (one form submitted per interview). The initial launch in 2022 served as benchmarking of current programmes to allow for monitoring over time for future assessments. This was done by averaging data across the two assessments (where two assessments were conducted in 2022). If a site only completed one of the two assessments, that one assessment was used as the overall 2022 benchmarking score for a year-end total result and report. Facilities were also asked to review the results with relevant stakeholders at their facility and identify up to three goals for enhancing their primate behavioural management programme in 2023 based on their specific results from the benchmarking assessments. The welfare assessment process is outlined in Figure 1.
Primate welfare assessment process to be completed semi-annually to track welfare progress over time for facilities working with primates.
Tool-use training
English and French version user training guides were created that included a knowledge check (i.e. quiz with five questions) upon completion. The training was presented as an e-learning module and took approximately 10–15 min to complete. Each site designated 1–3 primary behaviour contacts for completing the tool and retained the same primary contacts in Q1 and Q3; however, some sites with three assessors added or changed their third assessor in Q3. The training guide included the learning outcomes, an explanation of the purpose of the PWAT, description of what was included in the tool, expectations and instructions for using the PWAT including the number of rooms, different types of studies, and employees to interview based on the facility size and function. Also included was a demonstration of how to access the results, and information on what to do with the results once received, such as communicating results to relevant stakeholders and developing discussions on next steps and top priorities to address, and finally a five question knowledge check. Participants needed to achieve 100% on this quiz to gain access to the tool.
Results
Ten sites fully completed both assessments in 2022, although all 13 sites fully completed at least one of the two assessments. Misunderstandings regarding when completion was due or whom was completing the assessment contributed to the missing assessments for three facilities. For each assessment period, across all 13 facilities, an averaged total of 2,615 primates (all cynomolgus macaques) were assessed from within 62 rooms, and an averaged total of 66 employees completed the culture of care surveys.
The overall scores for Q1 and Q3 2022 are provided in Figure 2(a) and for each category in Figure 2(b). The breakdown of scores by category and subcategory for each facility for the combined 2022 benchmarking score are provided in Table 3. For Q1 2022, the scores ranged from 63–95%. For Q3 2022, the scores ranged from 53–90%. When looking at the categories assessed, sites performed best in the physical category (88.5% averaged across sites) and had the most room for improvement in the environmental category (65.2% averaged across sites). Two sites were outliers overestimating their welfare score (89 and 95%) due to misunderstandings in how to use the non-applicable score options. A follow-up meeting with PWAT site representatives was conducted to provide additional training. A summary report for each facility was provided for each assessment including the averaged 2022 results for overall PWAT score, welfare categories, and welfare subcategories. General findings are discussed below per category.
Total PWAT scores by site for the 2022 benchmarking for March (Q1 2022) and September (Q3 2022) (n = 13) including (a) the overall scores and (b) the relative scores by category (total possible score is 600).
Results of the 2022 welfare assessment, presented as the averaged percentage scores from the Q1 2022 and Q3 2022 assessments for each category, subcategory, and total score, presented for each facility (A–M) and averaged across facilities
Physical
The physical category aimed to evaluate physical and clinical health of primates as well as factors that may influence health. Overall sites scored the highest in the physical category compared to other welfare categories. Scores ranged from 72.5 to 98% (Table 3). The subcategory with the highest average score was general body condition (90.2%). The general condition descriptors where sites lost the most points were coat quality, animals reacting normally to external stimuli and personnel, animals having appropriate muscular development and fat deposits for sex and age, proper hydration, and signs of health conditions. The subcategory generally achieving the lowest score was records (81.7%), with lower scores seen for descriptors relating to the appropriateness of the bodyweight history of the animals based on their sex and age, and whether animal procedure history is readily available in the animal rooms. Feedback from personnel included that electronic records are generally available, thus records are not kept in the rooms but are readily accessible to employees. This suggested that the tool needed rewording to capture this possibility. For nutrition, lack of ad libitum feeding of base diet was the descriptor that most influenced scores for this category. Many of the sites record food intake as part of their study activities and thus provide a specific amount of diet for each animal, and all sites indicated that there are procedures in place to manage and observe animals with reduced appetite. For pain assessment and mitigation, the descriptor that was not met at all sites was “the facility has a specific policy or Standard Operating Procedure (SOP) concerning primate pain management practices based on current veterinary practices.” Feedback received on this descriptor was that some sites may have several policies or procedures for pain management and thus were not sure how to respond, or that the pain management policy was not specific to primates, but was generally applicable across species.
Behavioural
The behavioural category aims to evaluate normal and abnormal primate behaviours and behaviour management inputs and sites scored from 60 to 97.5% in this category (Table 3). The subcategory with the highest score was social behaviour (85.2%), but within that subcategory, there were indicators that social group management can be challenging for sites, with some sites reporting evidence of fighting and bullying in social groups, and only six of 13 sites regularly assessed social groups and had established procedures to address incompatibility. The subcategory with the lowest average score across sites was ability to cope (74.9%), which has one descriptor that assessed “when separated from social partners for a procedure, primates do not perform abnormal behaviours.” For behavioural assessment, only seven of 13 sites indicated that personnel are specifically trained to identify normal and abnormal primate behaviours and that regular behavioural assessments are conducted. However, eleven of 13 sites had a team or individual specialised in primate behaviour. Sites reported the occurrence of abnormal behaviours in the animal behaviour category in 31.7% of the assessed rooms in Q1 2022 and 49.2% of the assessed rooms in Q3 2022. For feeding behaviour, there was only one descriptor: “primates are expressing natural feeding behaviours,” and only seven of 13 sites scored above 90% for this category, indicating some room for improvement in providing foraging opportunities. Only one site conducted breeding and had the ability to offer parental behaviour.
Environmental
The environmental category aims to evaluate the quality of the space provided as well as resources and furnishings that permit primates to express species-typical behaviours and postures. Environmental was the lowest scoring category overall. One site scored 104% due to bonus points, while the rest ranged in score from 37 to 84.5% (Table 3). The three highest scoring sites were based in the UK and EU, and employ EU pen-style housing, rather than cage-style housing, which is more common in the US. The scores for the environmental category also revealed gaps in resources being provided to primates to provide comfort (e.g. thermoneutral or elevated resting surfaces) and encourage species-typical behaviours (e.g. foraging, social housing) and only a few sites provided regular out-of-cage exercise opportunities for animals.
Training
The training category aims to assess procedures in place to prepare animals for study. The overall PWAT score for the training category ranged from 49 to 97% (Table 3). In general, acclimation periods are well established throughout facilities, with two sites indicating no procedure for environment acclimation upon arrival (because they are quarantine facilities) and four sites indicating that acclimation periods shorter than 14 days were permissible following animal arrival and depending on the source of the animals (e.g. local quarantine). There were a number of areas for improvement identified across sites in implementing more comprehensive behavioural management programmes. There were a number of descriptors within the subcategory habitution, desensitisation, and counter-conditioning for which sites scored below 2, such as habituation being conducted in a quiet environment and not paired with study activities, maintaining habituation over time, and documenting animal progress. In particular, the subcategory positive reinforcement training had an average score of 35.4%, indicating a need for more formalised training programmes to be implemented with primates to aid in co-operation and human-animal interactions. Additionally, in discussion with personnel, there were assumptions that positive reinforcement training referred specifically to clicker training. Clicker training is a method used in positive reinforcement training in which a device is used to mark the behaviour of interest with a click sound prior to providing the reinforcement. More clarification is needed on the different forms positive reinforcement can take in working with primates, such as simply providing a food reward for calm behaviour or for approaching personnel without the need for additional tools. For human interactions, a number of sites indicated that primates showed fear or stress towards personnel when staff are near the enclosure, emphasising a need to work on positive human-animal interactions.
Procedural
The procedural category assesses refinements to research protocols. The overall PWAT score for procedure ranged between 37 to 90% (Table 3). The highest scoring subcategory was recovery (88.5%), which has one descriptor ‘following procedures, animals are monitored for pain indicators and to ensure animals return to normal state. Monitoring is performed by an individual familiar with the animals.’ The subcategories with the lowest score were capture (60.1%; ‘animals are comfortable and compliant with removal from home enclosure’) and ambience, which refers to noise levels and procedural space. In other categories there were a few descriptors for which sites consistently scored lower than 2, such as offering manipulanda during moderate restraint, providing rewards during or after procedures, training animals to take test articles voluntarily, minimising animal disturbance and personnel rushing by using strategic scheduling, and having sufficient space to allow animals and people to move without risk of injury, indicating issues that could be addressed across multiple sites.
Culture of care
The culture of care category aims to assess employee satisfaction and training to prepare them for their responsibilities working with primates. The overall PWAT score for culture of care ranged from 67.5 to 95.5% (Table 3). The subcategory with the highest score was initial training (86.6%), which assessed the efficacy of the introductory training materials and learning environment. The subcategory with the lowest score was compassion fatigue and resiliency building programmes and activities (58.9%), highlighting an important gap for sites to enhance their compassion science programmes. Descriptors for which sites consistently scored below 2 were related to work-like balance, identifying programme leaders, and implementing internal programmes to honour the research animals.
Based on the results, a number of global trends were observed and general recommendations were created for how to address those gaps. These recommendations are provided in Table 4.
General recommendations for improvement in each category based on the results of the 2022 Primate Welfare Assessment Tool benchmarking exercise
Discussion
The aim of this project was to develop a primate welfare assessment tool (PWAT) as a cross-facility collaborative effort to benchmark current primate management programmes and monitor forward animal welfare progress. Using a global focus group allowed the tool to be applied across facilities and helped build consensus and buy-in that aided in successful tool implementation. The tool was ultimately developed in a Smartsheet® platform, enabling semi-automated data management and visualisation. The PWAT incorporates input- and outcome-based welfare indicators as well as aspects of employee satisfaction, providing a holistic representation of primate management. The tool was used to benchmark primate programmes in 2022, and results were discussed and used by site-level stakeholders to develop goals to work towards and to ultimately create positive change for primate welfare.
Focus groups, as used in the development of the PWAT, have been used in a variety of fields such as education (Stathopoulou et al. Reference Stathopoulou, N-T and Christodoulides2019), health (Brouwers et al. Reference Brouwers, Joosen, van Zelst and Van Weeghel2019), and animal science (Ritter et al. Reference Ritter, Russell, Weary and von Keyserlingk2021). Focus groups generally include participants recruited based on experience and expertise in a subject matter (Tausch & Menold Reference Tausch and Menold2016). Stakeholder focus groups are beneficial in driving animal welfare changes due to the complexity of animal welfare issues and the need to achieve buy-in from different levels of stakeholders to create lasting changes (Fernandes et al. Reference Fernandes, Blache, Maloney, Martin, Venus, Walker, Head and Tilbrook2019). To address complex welfare issues and create lasting change, there are five key areas to consider, including reflexivity of considering multiple perspectives, responsiveness of being able to adjust to changing expectations, revitalisation of reducing conflict by redirecting stakeholders to a common goal, resilience by maintaining flexibility, and relational capital by maintaining collaborations between stakeholders (Termeer et al. Reference Termeer, Dewulf, Breeman and Stiller2015; Fernandes et al. Reference Fernandes, Blache, Maloney, Martin, Venus, Walker, Head and Tilbrook2019). Utilising a stakeholder focus group in developing the PWAT addressed these key areas. The tool was developed to be flexible across sites, regions, and business purposes, and designed to require minimal training by making descriptors simple, self-explanatory, and well-detailed (CCAC 2019). Ongong discussions and collaborations occur through internal primate behaviour group listserves and quarterly meetings. The PWAT was created as a collaborative effort to ensure relevance and applicability in the research environment by individuals from various global sites, job titles, and experiences and built to be flexible and long-lasting to meet changing expectations.
The principle behind the PWAT is similar to the Extended Welfare Assessment Grid (EWAG) which uses a matrix to assesses animal welfare and cumulative suffering in research animals (Honess & Wolfensohn Reference Honess and Wolfensohn2010). The EWAG uses similar welfare categories including clinical condition, behavioural deviations, environment, and experimental/clinical events. The main difference is that the EWAG focuses on assessing welfare at the individual level and monitoring cumulative suffering over time based on research use, whereas the PWAT is focused on assessing the overall primate behavioural management programme. Primates in research are used for more long-term studies compared to other research species, further emphasising the importance of a multifaceted approach to assessing welfare over time and considering cumulative suffering (Honess & Wolfensohn Reference Honess and Wolfensohn2010; Paterson et al. Reference Paterson, O’Malley, Moody, Vogel, Authier and Turner2023). Using a broader tool such as the PWAT could help identify gaps in the primate management programme that could result in animal-based indicators of poor welfare or cumulative suffering, while implementing a purely animal-based measure of welfare, such as the EWAG, could be used to monitor specific at-risk animals based on procedural severity and cumulative use, which could then be used to make decisions on humane endpoints (Honess & Wolfensohn Reference Honess and Wolfensohn2010; Nunamaker et al. Reference Nunamaker, Davis, O’Malley and Turner2021).
The PWAT was developed through focus group discussions but alternative approaches exist. Recently, Truelove et al. (Reference Truelove, Martin, Langford and Leach2020) identified 115 research macaque welfare indicators using a Delphi consultation process in which anonymous expert participants complete surveys over multiple rounds to achieve consensus on a topic. There are certain similarities between the Truelove and colleagues (Reference Truelove, Martin, Langford and Leach2020) results and the PWAT. In Truelove et al. (Reference Truelove, Martin, Langford and Leach2020), the welfare indices were split into six categories including enrichment, environment, health and management practices, appearance and health measures, behaviour, and physiology and genetics. In comparison to the PWAT, there are a similar number of indicators (PWAT = 133) and categories are comparable except that there was no integration of culture of care in the Delphi approach. For the Delphi approach, subject matter experts agreed that social enrichment and self-injurious behaviour were the most important indices of welfare (Truelove et al. Reference Truelove, Martin, Langford and Leach2020). Within the PWAT, there is a significant emphasis on social housing for which sites lose points for singly housing primates and indicators related to social housing are heavily weighted. In the Delphi method study, welfare indices were rated based on validity, reliability, and feasibility (Truelove et al. Reference Truelove, Martin, Langford and Leach2020). In the PWAT, the two main researchers (EAP, PVT) created the welfare descriptors and weighted them based upon welfare impact, then the focus group reached consensus by adding and removing some descriptors and adjusting the scoring weights. In the Delphi method study, it is suggested that assessing environmental-based measures (i.e. room temperature) is more feasible than animal-based measures (i.e. animal behaviour or health indicators) (Truelove et al. Reference Truelove, Martin, Langford and Leach2020). This is also true for the PWAT given that animal-based measures are direct measures of animal state and can be difficult to obtain, invasive, lengthy, and somewhat subjective (CCAC 2019). For example, quantifying behavioural abnormalities can be time-consuming and may differ between observers (Jirkof et al. Reference Jirkof, Abdelrahman, Bleich, Durst, Keubler, Potschka, Struve, Talbot, Vollmar, Zechner and Häger2020). However, in recent discussions on animal welfare assessment, there has also been greater focus on including individual animal assessments rather than just assessing at the group level to have holistic representation of overall animal welfare (Spangenberg & Keeling Reference Spangenberg and Keeling2016; CCAC 2019; Winkler Reference Winckler2019), which was why it was deemed important to maintain animal-based measures at the room-level in the PWAT.
More recently, Prescott et al. (Reference Prescott, Leach and Truelove2022) took the 115 welfare indices identified by Truelove et al. (Reference Truelove, Martin, Langford and Leach2020) and used a modified Delphi method to narrow the indices down to 56 and create a usable tool (GEN-MAC). The tool was used with a hypothetical scenario involving 500 primates but has not been tested in an animal facility (Prescott et al. Reference Prescott, Leach and Truelove2022). In comparison to the PWAT, the GEN-MAC incorporates fewer and more simplified descriptors, which may make the tool faster to use. The scoring systems are similar, using scores of 0, 1, and 2. The GEN-MAC tool is currently available in Microsoft Excel®. For the PWAT, Smartsheet® was used to allow for the tool to be semi-automated such that personnel have immediate access to their raw data inputs, descriptive statistics, data visualisation, and historic data trends. These data are provided per category and subcategory, and at the site and company level allowing for a holistic and detailed view of the primate programmes in real time and historically.
The PWAT was specific enough to distinguish between facilities at a rather detailed level. For example, when comparing facility G and facility I in Table 2 (sites that have the same final score; 78.5%), one can see differences between facilities at the subcategory level in descriptors involving records, feeding behaviour, resources/enrichment provided, and restraint. This specificity allows for tailored recommendations to be made for each facility. Additionally, the PWAT was also able to identify global trends for more challenging aspects of research primate management as noted in Table 3, which will allow for development of global training resources to support site improvements. For example, for the ‘physical’ welfare category, to improve pain assessment and management, using tools such as the cynomolgus macaque grimace scale and specific behavioural indicators could be an important consideration for improving detection (Paterson & Turner Reference Paterson and Turner2022).
A unique component of the PWAT not found in most other welfare assessment tools is inclusion of a culture of care component. Creating a work environment in which employees feel valued and are satisfied with their jobs leads to better animal care and quality of science (Klein & Bayne Reference Klein and Bayne2007). In healthcare, and more recently, in veterinary medicine, culture of care in the work environment as well as compassion satisfaction and resilience, work-related satisfaction, and feeling valued have been recognised for their importance (Newsome et al. Reference Newsome, Clemmons, Fitzhugh, Gluckman, Creamer-Hente, Tambrallo and Wilder-Kofie2019; LaFollette et al. Reference LaFollette, Riley, Cloutier, Brady, O’Haire and Gaskill2020; Randall et al. Reference Randall, Moody and Turner2021). Areas that have been identified as beneficial to compassion satisfaction are work-life balance, positive interactions with research animals, involvement in animal-related refinements, continuing education opportunities provided during working hours, and feeling recognised and valued for their work (Randall et al. Reference Randall, Moody and Turner2021; O’Malley et al. Reference O’Malley, Moody, Foster and Turner2022). Using a formalised tool to evaluate and monitor the work environment for employees working with animals will benefit both animals and employees. In the PWAT, areas such as appropriate species-specific training and continuing education, access to compassion fatigue and resiliency building resources, the ability to be involved in animal welfare initiatives and voice concerns about animal welfare, work-life balance, and being recognised and feeling valued for their role in caring for animals and ensuring good animal welfare were all included to get a well-rounded assessment of the well-being of employees working with primates.
A limitation of the tool is that it was designed and tested in a contract research organisation environment in which cynomolgus macaques are the primary species worked with. It has not been tested for use with other primate species or in other environments with primates, such as in a zoo setting. Another limitation of the tool as it was designed and implemented is that while its use is required for sites working with primates, how the tool outcomes are used to change primate management programmes is not specified. Due to significant differences in facility function and operations (e.g. quarantine facility vs safety assessment facility) it is recognised that each facility will have unique strengths and areas for improvement, and that there will be asynchronous progress. The goal of the tool was not to diagnose issues and mandate changes, but to empower sites to evaluate their programmes, self-identify areas of improvement, and to prioritise and implement changes based on resource availability. As the sites assessed are all AAALAC accredited, welfare standards are already high. Therefore, the PWAT is meant to be an additional tool to encourage targeted refinement in primate management through discussions between various stakeholders and collaboration across facilities to share knowledge and experiences.
Research facilities are encouraged to incorporate welfare assessment tools and increasingly oversight bodies are requiring these as one means of ensuring continuous improvement in animal care and behavioural management programmes (CCAC 2021; Turner & Bayne Reference Turner and Bayne2023). As in the PWAT, assessments should include input- and outcome-based measures to ensure a holistic approach. Tools should be tested prior to implementation to ensure validity and feasibility and once the welfare assessment tool is implemented, the results should be shared with various stakeholders including animal care, behaviour, veterinary, and management personnel. Communication, discussion, and follow-up of the results among key stakeholders is important to promote ongoing refinements. The tool should be evaluated periodically with subject matter experts to ensure that the tool remains relevant and practical in assessing research primate programmes.
Animal welfare implications
Welfare assessment is important for monitoring and improving animal care and use programmes within research environments. The tool includes input- and outcome-based measures for assessing research primate welfare, as well as measures of employee satisfaction and culture of care. The aim of the tool is to provide a means for holistic assessment of primate management programmes, which will allow facilities to identify gaps in their programmes to define and prioritise needed refinements. The tool was developed and launched in a global environment, and therefore has the potential to improve research primate welfare globally.
Conclusion
Through an internal collaborative effort of primate experts, a primate welfare assessment tool (PWAT) was created and used at facilities that housed primates across a global business. The primate welfare assessment tool has six welfare categories that include physical, behavioural, environmental, procedural, training, and culture of care evaluations with a total of 133 welfare descriptors. The tool is composed of three forms: one that evaluates the behaviour management programme at the site level, one that evaluates individual animal states at the room level, and one that assesses the culture of care within a facility based on employee feedback. The PWAT successfully differentiated between programmes at different sites and identified areas for improvement at the facility and corporate level. Future directions for the PWAT will be to measure programmes over time, identify needed refinements, and overall improve research primate welfare.
Acknowledgements
The authors would like to acknowledge site personnel who participated in the development, testing, and finalisation of the Primate Welfare Assessment Tool, as well as management and operations personnel for supporting the development and launch of the tool. EAP was supported by scholarships from NSERC, the Ontario Graduate Scholarship program, and the Ontario Veterinary College.
Competing interest
The authors work for or have worked for Charles River. Charles River was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it to Animal Welfare for publication.
Open access
