Animal welfare is a multifaceted concept that increasingly focuses on the cumulative physical, psychological, and behavioral states of individual animals (1). It encompasses scientific, ethical, economic, cultural, and religious dimensions with varying perspectives among scholars (2, 3). Initially, animal welfare science focused on enhancing the welfare of production and laboratory animals (4, 5). The Farm Animal Welfare Advisory Committee (FAWAC) took a significant step in 1965, developing the Five Freedoms model to address farm animal welfare concerns (6). It stated that animals should be free from hunger, discomfort, pain, and fear, and able to express natural behaviors. The model dominated discussions on animal welfare in Europe for decades (7), serving as a comprehensive framework while acknowledging the operational constraints of the livestock industry (8). However, criticisms surfaced, questioning its practicality and minimal emphasis on positive welfare experiences (9), prompting the development of alternative frameworks. The Five Domains Model offers a holistic approach focusing on affective terms and recognizing the subjectivity in measuring mental experiences (10). This emphasis on mental well-being aligns with broader ethical, policy, and legal considerations in contemporary animal welfare science. The model integrates the concept of agency within Domain 4 (Behavioral Interactions), enabling the evaluation of animal engagement in voluntary, self-generated, and goal-directed behavior (11) and human-animal interactions (10). Widely accepted in farm and zoo communities, the Five Domains Model has been adopted by organizations like WAZA (12) and the Zoo and Aquarium Association, Australasia (13) to uphold high welfare standards. However, to effectively utilize the model as a welfare assessment tool, attention should be given to using well-validated measures, ensuring transparency in expert panel selection, and implementing a clear welfare grading system (14).

Numerous welfare assessment frameworks have emerged by incorporating the Five Freedoms and Five Domains models. For example, Welfare Quality builds upon the Five Freedoms Model and integrates scientific expertise and ethical considerations from various stakeholders, including the general public, industry, and political bodies, to evaluate welfare (4). It prioritizes animal-based measures and follows a bottom-up approach, assigning scores based on four crucial principles: proper nourishment, suitable housing, good health, and appropriate behavior. These principles serve as the basis for evaluating overall welfare, and their scores are combined to determine the final assessment. The Opportunity to Thrive Program flips the concept of the Five Freedoms to focus on achieving a positive welfare state, with a particular emphasis on reintegrating animals back into their natural habitats (15). The framework offers a comprehensive method for managing animals, incorporating formulated diets, environmental design, healthcare, enrichments, choice and control, and access to species-typical behavior. These inputs ultimately aim to achieve desired outputs, resulting in an overall animal welfare assessment. A 24/7 approach was proposed to evaluate zoo animal welfare, utilizing the 12 welfare assessment criteria from the Welfare Quality framework (16). This approach considers the natural history, biology, ecology, diet, habitat, social structure, and activity patterns of animals throughout both day and night, providing a thorough understanding of their welfare. Finally, the Universal Animal Welfare Framework is an institutional-level welfare assessment framework based on the Five Domains Model (17). Developed by the Detroit Zoological Society in 2015, it examines zoo practices, policies, resources, and measures related to housing, routine, and behavior.

Both species-specific and species-general welfare assessment tools have utilized these welfare models and frameworks. Generalized tools work under the assumption that animals have the same basic needs, so management should be based on natural history. However, these tools face challenges in addressing species-specific nuances. A few have been developed for specific species and include the giant pacific octopus (Enteroctopus dofleini) (18), pygmy blue-tongued skink (Tiliqua adelaidensis) (19), bottle-nosed dolphin (Tursiops truncates) (20), waterfowl (Anseriformes) (21), dorcas gazelle (Gazella dorcas) (22), and elephant (Elephas maximus) (23).

These offer a refined and more precise evaluation of animal welfare by tailoring assessments to the unique needs, behaviors, and physiological aspects of a particular species (24). For some species with special spatial, environmental, social, or cognitive needs, a “one-size-fits-all” strategy to assess welfare may not be appropriate; rather, species-specific (25, 26) and if possible, context-specific assessment tools are needed.

Elephants, characterized by their large body size, complex social lives, varied food requirements, and extensive wild home ranges (27–29), pose challenges to meeting physical, psychological, and physiological needs in human-created environments (30). Ensuring good welfare for these animals involves allowing some degree of choice and control. Additionally, wild elephants spend about 80% of the time foraging and are highly social (31). Denying these freedoms can result in maladaptation, chronic stress, poor welfare (32, 33), and abnormal stereotypic behaviors (34, 35). Good zoos provide health care, safety from predation, and food security (30) and aim to meet exercise (36), foraging, and social complexity (37) needs. However, high mortality, low birth rates, limited reproduction, and health problems continue to hamper zoo elephant population sustainability (38–42), igniting worldwide concerns over animal welfare (43, 44). In 2016, a series of epidemiological studies of elephants in North American zoos revealed problems associated with ovarian acyclicity (45), health and musculoskeletal function (46), stereotypic behaviors (35), and high body condition scores (47). A similar set of studies on tourist elephants in Thailand found similar problems associated with elevated stress hormones (48), excessive body condition and metabolic derangements (49), and stereotypies (50). Finally, surveys of thousands of elephants in hundreds of tourist venues across Thailand, India, Nepal, Sri Lanka, Laos, Cambodia, and Malaysia suggest that more than half (63%) are kept in inadequate conditions (51–53). There is little doubt that comparable situations exist for logging (54), temple (55, 56), and circus (57) elephants as well. Thus, while the problematic state of captive elephant welfare across diverse conditions is now well-known, the solutions have proven far more elusive.

Considerable attention has been directed toward improving zoo elephant welfare, while the unique conditions and challenges faced by captive elephants in non-zoo settings are often overlooked (58). There are over 14,000 captive Asian elephants outside traditional zoo environments across 13 range countries, primarily in tourist or logging camps and temples (58). There are notable differences in the management of zoo and camp elephants. Zoo elephants are typically managed in protected contact systems, minimizing direct interaction with humans and other practices that adhere to standardized regulations (59). Staff are responsible for feeding, bathing, training, and veterinary care; however, because of limited space, socialization and exploration can be limited. In contrast, camp elephants are often managed in free contact where elephants and people share the same space, including with tourists (60). Daily routines involve tourist-related tasks such as shows, riding, walking, bathing, feeding, or observation (60). Some elephants participate in cultural activities like religious rituals and festivals (61). Welfare can be better in camps situated in natural environments, with forests and rivers providing more natural foraging and exercise opportunities than zoos (62). However, restraint methods like chaining and using an ankus (also known as a bullhook or guide) to control elephants are significant concerns (63). In addition, the lack of enforceable standards results in varied management practices across and even within camps (64), which ultimately poses challenges in addressing the welfare needs of camp elephants. Animal activists continually voice concerns regarding the welfare and management of tourist camp elephants (65, 66). Thus, there is a need for a holistic, evidence-based welfare assessment approach to identify potential welfare risks, inform management decisions, and record welfare changes over time (67, 68). It also can contribute to elephant welfare standardization and policy-making processes crucial for properly managing elephants in range countries.

This review examines available generic and elephant-specific welfare assessment tools and methodologies and discusses applicability to tourist camp elephants (Tables 1, 2). While there have been several reviews of animal welfare frameworks (24–26), to our knowledge, this is the first overview of welfare assessment tools specific to elephants. Predefined criteria guided the selection of welfare assessment tools for this review to ensure a representative and comprehensive overview. Those included relevance to captive elephant welfare, recognition and adoption in the scientific community, and diversity of approaches. The featured tools were carefully chosen to provide readers with meaningful insights into the diversity and applicability of current welfare assessment practices for captive elephants, acknowledging that the selection may not encompass every existing tool. Our ultimate goal is to synthesize a new welfare assessment tool specific to elephants used in tourism, considering the strengths and limitations of existing tools and challenges faced by tourist camps.

Within the two main welfare models used today, Five Freedoms and Five Domains, a range of methodologies exist for comprehensive welfare evaluations. To satisfy accreditation criteria, zoos and aquariums regularly evaluate the welfare of animals under their care (89), often using tools designed for multiple species. Each tool offers distinctive features that often serve different functions, such as complete behavioral and space utilization monitoring of ZooMonitor, global data sharing features of ZIMS, numerical and visual data representation of AWAG, or reliable and valid captive elephant measures presented by EBWAT. Tools range from digital formats to more traditional pen and paper for data recording and monitoring. However, overall, the trend is for institutions to use digital tools and advanced technologies to improve welfare standards (100). The tools differ in assessment levels, from assessments to understand individual variation to institutional level assessments that can inform on prioritization of resources and broadly benchmark progress in advancing welfare standards. These tools mostly rely on keeper ratings as a proxy for quantitative behavior assessments because keepers spend more time with the animals and can detect subtle changes that might be overlooked by others less familiar (81). Most also use a relative grading system; for example, AWRAP implements a 0–2 scale, Welfare Discussion Tool a 4-point scale, Ackonc-AWA a 3-point scale, and AWAG has a 10-point scale. Likert scales are also commonly utilized when evaluating behavioral indicators (25). ZIMS is flexible to allow users to select the grading in binary, numeric, and percentile values. Objective welfare scores allow the recording of welfare changes over time (25) and assist accreditation schemes in determining if an organization meets welfare requirements (67).

Across the tools, inputs range from observing behaviors to measures of health and stress indicators to provide comprehensive assessments across different welfare domains. Observational behavioral assessments emerged as a standard in all of the existing tools. Some tools use CCTV or cameras, while others rely on direct observations by keepers or other experts. In one study, ZooMonitor was used along with 18 closed–circuit cameras and five camera traps to record behavior states, habitat use, and social interactions of seven zoo Asian elephants (79). That study highlighted the benefits of combining ZooMonitor with other assessment methodologies for comprehensive welfare interpretations. Tools are increasingly using behavioral indicators associated with comfort, play, affiliation, foraging, and sociality to evaluate mental and overall welfare states, in addition to commonly used and validated negative welfare indicators like stereotypies, poor health and reproduction, and high mortality and morbidity (23, 44, 101, 102). To that end, the score sheet of WelfareTrak consists of positive (e.g., calm-relaxed) and negative (e.g., self-mutilating) behaviors. EBWAT includes stereotypies, social interactions, feeding, comfort, social behaviors, interactions with the environment, vocalizations, and others to measure mental health. AWAG also evaluates stereotypies, social affiliations, enrichment utilization, and responses to training as measures of psychological welfare.

Many tools also incorporate health evaluations as animal-based measures of physical condition. Stool and urine appearance, body coat condition, wounds, skin lesions, locomotion, micturition behaviors, general illness, teeth condition, and coat condition are all included in the health domain of Ackonc-AWA. Physical assessments in the AWAG include factors such as body condition scores, appetite, drinking and feeding behaviors, and activity levels, while the AWRAP tool includes body condition and an overall general health score, and the WDT overlays behavioral data with cortisol (feces, urine, etc.) analyses. In the case of elephants, cortisol or its metabolites can be measured in blood, saliva, urine, feces, and hair (103). Indeed, a study in India found zoos and sanctuaries with low welfare scores tended to have elephants with higher urinary cortisol and stereotypy rates (71). Immunoglobulin A (IgA) is among the novel biomarkers used as a positive welfare indicator and also in assessments of immune function (104). Like cortisol, IgA fluctuations can indicate positive and negative welfare states (105, 106) and be measured non-invasively. Combining analyses of glucocorticoids and IgA with behavioral indicators like stereotypies can further validate assessment findings (107, 108). Methods like allostatic load indexes are gaining attention because of their ability to capture cumulative stress (109), and so could potentially be used to predict mortality and morbidity risks. Other indices to consider could include evaluations of preference/avoidance, displacement, vocalization, startle/vigilance behaviors, salivary or urinary epinephrine, heart rate variability, and cardiovascular function.

Digital tools play a significant role in zoological institutions, enhancing efficacy, data visualization, and multi-institutional collaborations (81, 100). However, implementing these tools institutionally in non-zoological settings will be challenging. Elephant mahouts may have limited knowledge of technological devices to use ZooMonitor or WelfareTrak, and most camps do not have research staff or volunteers to input data. Thus, paper-based assessment methodologies might be more appropriate. It also can be challenging for low-budget venues in range countries to afford CCTV cameras and access to software to analyze data. In tourist camps, where elephants are engaged in activities like bathing, riding, and walking in natural forests (60), CCTV monitoring is impractical and could raise privacy concerns. Constantly updating information in digital tools like ZIMS could also be a logistical challenge for camp staff. Finally, most of these tools are only available in English, making them less useful for range countries.

The current reliance on keeper ratings or expert opinions in welfare assessment tools for captive elephants in range countries also has limitations. Although intimately familiar with their elephants, mahouts (i.e., elephant keepers) might not consistently identify stereotypic behaviors or have a comprehensive understanding of the full spectrum of elephant behaviors (50). Studies have highlighted instances where mahouts, despite their proximity to the animals, could not identify certain behaviors accurately, leading to discrepancies between direct observations and keeper assessments (92). Moreover, mahouts often face time constraints in non-zoological settings due to engaging in tourist interactions, impeding their ability to monitor behaviors continuously. The potential for positive bias in mahout ratings, influenced by personal attitudes and care for specific animals, also raises concerns about the objectivity of assessments (67). A more effective approach might involve a collaborative model that combines the expertise of mahouts and trained observers. This hybrid approach utilizes both perspectives synergistically, with mahouts offering unique insights into individual elephant social interactions and preferences. At the same time, trained observers conduct focused, objective behavioral assessments, especially when evaluating stereotypies.

Moving forward, there is a need to develop a new welfare assessment tool specific to elephants used in tourism. Tools should go beyond mere adaptability from zoo-centric models to incorporate components that address the specific dynamics, challenges, stressors, and ethical considerations found in tourist camps. The tool should integrate a balance of animal and resource-based measures and avoid the narrow focus on single behavior or health indicators (23, 55) to provide a comprehensive welfare risk assessment (67). With an increasing focus on using welfare assessment frameworks for developing assessment tools, the Five Domain Model can be adapted to develop the welfare assessment tool. Despite criticisms against the Five Domain Model (14), it is the most widely used model in animal welfare science and is important because of its focus on mental states. If limitations such as reliable and valid measures focusing on the overall mental and welfare state of captive elephants, and a structured welfare grading system are considered (14), the Five Domain Model can be adapted to develop a new welfare assessment tool. Previously established behavioral measures for captive elephants (23, 101, 102) and welfare factors associated with tourist camps can be integrated and adapted for further testing. The tool must be rapid, adaptable, undemanding in resources, non-invasive, and easy to complete, considering financial limitations, feasibility, and ethical concerns associated with invasive techniques (23, 25). Impractical measures like cognitive bias that require experimental setups (26) and measures such as safety from predators can be omitted, acknowledging their minimal impact on captive elephants in tourist camps. Despite recent efforts to enhance efficiency through technology (100), the practical constraints of tourist camps necessitate a focus on direct observation and questionnaires with mahouts. In the case of developing countries, a lack of understanding and awareness of animal welfare among mahouts makes it more challenging (94). To address this, a tool should integrate the perspectives of both mahout and experts, ensuring a more comprehensive and objective evaluation of elephant welfare. The tool should be designed to be executed by a trained individual familiar with the methodology, metrics, and relevant evaluation tools, intending to expand training to allow stakeholders and medical staff for in-house evaluation and assessments. Ensuring the tool’s validity, reliability, and practicality is paramount (23). Achieving validity involves integrating existing literature, expert consultancy, and adapting established and validated assessment measures (25). Reliability can be tested through inter-rater, reliability, repeatability, and internal consistency assessments. The tool should be able to track welfare changes over time, integrating objective and quantitative welfare scores. This integration facilitates the comparison of welfare levels for future evaluations, enabling institutions to meet accreditation. It provides a quantifiable means to interpret individual welfare states, reduce inter-observer variability, and the potential for intra- and inter-group comparisons to establish best practices in elephant welfare across diverse tourist camps. A range of factors, such as age, health status, reproductive status, and life history, need to be accounted for in welfare assessments of captive elephants. Animals of different ages may react differently to the same scenario or resource allocation (110). Having baseline data for specific age groups for later comparison will contribute to developing a credible tool (25). For example, in the U.S. most captive elephants have experienced at least one inter-zoo transfer (111), which is associated with stereotypic behaviors (35). Similarly, seasonality in cortisol or its metabolites is evident in African (112) and Asian (113) elephants and so must be considered when evaluating the physiological significance of fluctuations as stress indicators. For example, in Thailand tourist camp elephants, higher fecal glucocorticoid concentrations were observed during winter (November–February), presumably due to colder temperatures (49), but during an international travel ban in Thailand during the Covid-19 pandemic, the highest concentrations were in the rainy season, suggesting it is tourist activities that are the most likely cause of increased glucocorticoid excretion during the winter, high tourist season months (114).

Implementing tools developed by ZIMS (Care and Welfare module) and Wild Welfare might reveal issues that require costly improvements, conflicting with the profit-oriented approach of elephant tourism, making camps hesitant to adopt changes. Welfare concerns surrounding captive elephants in Asia encompass various activities such as the use of ankus, chaining, riding, performing in shows, logging work, training methods, weaning, participation in religious rituals or festivals, and even involvement in polo tournaments. These activities provoke international concern, but the upright dismissal of such practices could lead to tension between local communities and outside experts. Thus, establishing collaborations among all stakeholders is vital for informed management adaptations.

Addressing welfare challenges and implementing assessment methodologies also demands clear objectives, heightened awareness, robust legal frameworks, and collaborative endeavors involving governmental bodies (115). Organizations like the Asian Elephant Specialist Group (AsESG) (116), WAP, ABTA, and MAZPA are developing conservation action plans, guidelines, and manuals for elephants managed in range countries. However, governmental concerns are often overlooked. Thailand, for example, initiated efforts to improve elephant welfare in 2002 with welfare standards for elephant camps, later supplemented by additional standards in 2009 (64). However, compliance was low due to non-enforceability and limited incentives. Thailand passed the Cruelty Prevention and Welfare of Animals Act in 2014 to prevent cruelty and improve animal welfare, but it has yet to be implemented. The Asian Captive Elephant Standards (ACES) were created to promote the well-being of elephants in Southeast Asia but require sincere participation from elephant camps and strict welfare monitoring by governmental bodies (64). Hopefully, the elephant camp standards launched by the Thailand National Bureau of Agricultural Commodity and Food Standards implemented in August 2024 will bring positive changes regarding the welfare of elephants in tourist camps. Similar issues are evident in other regions, like India (56, 117, 118), where many captive elephants are kept in temples under dismal conditions. Unlike conventional zoo or sanctuary environments, these settings operate under distinct governance structures that are often less restrictive and more culturally influenced. Therefore, a tailored welfare assessment tool must navigate the delicate balance between traditional and modern welfare standards, recognizing the diversity of beliefs and practices surrounding captive elephant management.

Addressing the welfare concerns of elephants in non-zoological settings, particularly tourist camps, presents a pressing challenge. Existing animal welfare assessment tools, although flexible, often lack essential components for effectively monitoring and enhancing elephant welfare in these contexts. Many tools were initially designed for zoological settings, rendering them less practical for non-zoological environments. Digital tools and methodologies such as keeper ratings encounter difficulties when applied to tourist venues because mahout knowledge of elephant biology and behavior is more limited. Tools should consider critical factors like reliability, validity, practicality, and recent advances in animal welfare science for comprehensive assessments. By doing so, we can better identify welfare risks, inform management decisions, track welfare changes over time, and contribute to standardizing elephant welfare practices and policy-making processes in non-zoological settings. This review proposes that there is a need to develop holistic, context-specific, evidence-based, and practical assessment tools tailored to the unique needs of tourist camp elephants across Asia. Recognizing the limitations of current approaches, we are actively engaged in developing a novel assessment tool specifically designed for assessing the welfare of elephants in tourist camps. This initiative aims to fill the gaps identified in existing methodologies and promote higher welfare standards for elephants across Asian tourist venues. By employing a comprehensive and tailored approach, we aspire to foster positive welfare outcomes for elephants and contribute to the broader efforts aimed at enhancing animal welfare across diverse captive settings in Asia.

RG: Conceptualization, Writing – original draft, Writing – review & editing. JB: Conceptualization, Writing – original draft, Writing – review & editing. CT: Conceptualization, Writing – original draft, Writing – review & editing. PB: Conceptualization, Writing – original draft, Writing – review & editing.