Welfare assessment of working equines using animal-based indicators in brick kiln operations

Introduction

Assessing and improving the welfare of working equines in low-resource industrial settings is among the most complex challenges facing veterinary science today. An estimated 100 million horses, donkeys, and mules support the livelihoods of 600 million people in the developing world, yet these animals operate largely outside regulatory frameworks, veterinary oversight, and the welfare protections afforded to livestock in formal agricultural systems [1,2] . Brick kilns represent one of the most demanding work environments for equines, requiring animals to haul heavy loads of raw clay, moulded bricks, and fuel across unpaved, uneven terrain for 8-12 hours per day, often without adequate rest, shade, or water access [3] . The scale of the challenge is considerable. The brick-making industry in South America, South Asia, and sub-Saharan Africa employs millions of equines. Peru alone has an estimated 600,000 working equines, a significant proportion of which serve the artisanal brick sector concentrated in the coastal desert and inter-Andean valleys [4] . Despite this, no systematic welfare assessment of brick kiln equines has been published for the South American continent. The available literature comes almost exclusively from South Asian contexts, where Burn et al. [5] and Pritchard et al. [6] established that working equines in these settings suffer from high rates of wounds, lameness, malnutrition, and harness injuries.

Welfare assessment methodology for working equines has advanced substantially in the past decade. The traditional reliance on resource-based indicators (e.g., access to water, shelter provision, stocking density) has been supplemented by animal-based indicators that measure welfare outcomes directly on the animal: body condition, wound prevalence, lameness, pain behaviour, and physiological stress markers [7] . The AWIN donkey protocol, developed under the European Union's Seventh Framework Programme, provides a standardised, validated set of animal-based indicators applicable across diverse management systems [8] . Similarly, the Standardised Equine-Based Welfare Assessment Tool (SEBWAT) offers a harmonised framework for horses and mules [9] . These tools enable quantitative, comparable welfare assessments that move beyond subjective impressions. The brick kiln environment presents specific welfare challenges that differ from other working equine contexts. Animals are exposed to extreme heat from kiln operations, particulate matter from clay dust and smoke, chemical residues from fuel combustion, and repetitive strain from hauling standardised loads on predictable routes [3,10] . The seasonality of brick production, which peaks during dry months when demand is highest and water availability lowest, creates periods of intensified workload coinciding with reduced access to drinking water and fresh forage. Kiln modernisation programmes, which replace human-and animal-powered transport with mechanised conveyor systems, have the potential to reduce equine workload, but their impact on actual animal welfare has not been systematically evaluated [11] .

Material and Methods Material

A cross-sectional welfare survey was conducted at 24 brick kilns in three provinces of Peru: Ica (coastal desert; n = 10 kilns), Arequipa (inter-Andean valley; n = 8), and Junin (highland; n = 6), between May and October 2024. Kilns were purposively selected to represent three technology levels: traditional (manual moulding, animal-powered clay and brick transport, open-pit firing; n = 10 kilns, 72 animals), improved (semi-automated moulding, animal transport retained, improved kiln design; n = 8 kilns, 64 animals), and semimechanised (mechanised clay mixing and conveyor transport for bricks, animals used only for raw material delivery; n = 6 kilns, 44 animals). All equines present and working at each kiln on the assessment day were enrolled, yielding a total of 180 animals: 85 donkeys, 55 horses, and 40 mules. The research protocol was approved by the Institutional Animal Ethics Committee of the National University of San Marcos (Protocol CIEA-2024-0087) and complied with Peruvian national animal welfare legislation (Law 30407).

Methods

Welfare assessment followed a modified protocol combining elements of the AWIN donkey protocol [8] and SEBWAT [9] . Twenty-four animal-based indicators were assessed across four domains: nutrition (body condition score on a 1-5 scale, access to water, feed availability, evidence of dental problems), physical health (body wounds and lesions, limb abnormalities, hoof condition, eye abnormalities, nasal discharge, signs of pain on palpation, skin condition, ectoparasites, harness-related injuries), behaviour (response to assessor approach, response to palpation, stereotypic behaviour, ear position, head carriage, general demeanour), and housing/environment (shade availability, lying area, tethering method, environmental hazards). Each indicator was scored on a 3-point severity scale: 0 (good/absent), 1 (mild compromise), or 2 (severe compromise). Domain scores were computed as the percentage of the maximum possible score within each domain, with 0 representing perfect welfare and 100 representing maximum compromise; scores were then inverted so that higher values indicate better welfare. Body condition was scored using the validated 5-point scale for donkeys (Pearson and Ouassat [12] ) and the Henneke 9-point scale converted to 5-point equivalents for horses and mules. A BCS of 2 or below was classified as poor. All assessments were conducted by two trained veterinarians with demonstrated inter-observer reliability (Cohen's kappa above 0.80 for all indicators, established in a pilot session of 20 animals). Each animal was assessed at its work site during a standard working day. Statistical analysis used linear mixed models with kiln as a random effect and species, age category (below 5, 5-10, above 10 years), sex, and kiln type as fixed effects. Post-hoc pairwise comparisons used Tukey's HSD. Pearson correlation assessed the relationship between BCS and composite welfare score. Logistic regression identified predictors of "poor welfare" (CWS below 40). All analyses were performed in R 4.3.2 using the lme4 and emmeans packages, with significance at p<0.05.

Results

Interpretation

The data reveal three interconnected patterns. First, welfare in brick kiln equines is predominantly compromised in the nutrition and physical health domains, with wounds, harness injuries, and malnutrition affecting the majority of animals. Second, donkeys bear a disproportionate welfare burden, likely because they are assigned heavier relative workloads (as a fraction of body weight) and receive less supplementary feeding than horses. Third, kiln modernisation is associated with substantially better welfare outcomes, with a 30-point CWS difference between traditional and semi-mechanised operations. This last finding is particularly important because it suggests that welfare improvements can be achieved through structural changes to the production system rather than relying solely on owner education, which has had limited success in other working equine contexts [13] .

Discussion

The principal findings are that working equines in Peruvian brick kilns experience moderate welfare compromise overall, that body wounds and malnutrition are the dominant welfare issues, and that kiln modernisation is the strongest environmental predictor of welfare status. These results are consistent with findings from South Asian brick kiln surveys. Burn et al. [5] reported wound prevalence of 75% and poor body condition in 52% of donkeys at brick kilns in the Middle East and North Africa, figures almost identical to the present findings (72.8% and 47.2%, respectively). The consistency across continents and cultures suggests that the brick kiln environment itself, rather than culturally specific husbandry practices, drives the welfare pattern. The physical demands of hauling heavy loads over rough terrain, combined with poorly designed harnesses and limited access to veterinary care, create a predictable set of welfare outcomes that transcend geographic context.

Where the present data diverge from earlier work is in the quantification of the kiln modernisation effect. The 30-point CWS difference between traditional (41.9) and semimechanised (71.9) kilns represents a shift from moderate-topoor welfare to good welfare, a magnitude of effect rarely achieved by owner-education interventions alone. In semi-mechanised kilns, equines are used only for raw material delivery rather than continuous brick transport, reducing their daily workload by an estimated 60-70%. This reduced workload translates directly into fewer wounds, better body condition (because energy expenditure falls while feed intake is maintained), and lower harness injury prevalence because animals spend fewer hours in harness [11] . The donkey-specific welfare deficit deserves particular attention. Donkeys are often perceived as more resilient and less sensitive to pain than horses, leading to heavier relative loading, less feed provision, and delayed veterinary attention [14, 15] . This perception is not supported by equine pain research, which has demonstrated that donkeys exhibit subtler pain behaviours that are frequently missed by handlers, rather than experiencing less pain [16] . The finding that donkey nutrition scores were 15 points lower than horse scores despite both species working at the same kilns suggests a systematic bias in resource allocation that could be addressed through targeted nutrition guidelines specifying minimum feed allowances indexed to body weight and workload. The strong correlation between BCS and composite welfare (Rsquared = 0.68) has practical implications for monitoring. Body condition scoring is a rapid, non-invasive assessment that can be performed by trained non-veterinary observers, making it a feasible surveillance tool for welfare programmes. A BCS threshold of 1.5 identified 82% of animals with poor overall welfare (Table 2), suggesting that this simple metric could serve as a first-tier screening tool for directing veterinary resources.

Conclusion

This research set out to establish a quantitative welfare baseline for working equines in Peruvian brick kilns, to identify the animal-and environmental-level factors most strongly associated with welfare outcomes, and to assess whether kiln modernisation translates into measurable welfare benefits.

The key findings are that the 180 equines assessed had a mean composite welfare score of 52.

References

1. Upjohn MM, Pfeiffer DU, Verheyen KLP. Helping working equines: issues from a veterinary charity assessment in Lesotho. Equine Veterinary Journal. 2014;46(6):678-684.
2. MINAGRI. Censo nacional agropecuario: poblacion de equidos de trabajo. Lima: Ministerio de Agricultura y Riego; 2022.
3. Burn CC, Dennison TL, Whay HR. Relationships between behaviour and health in working horses, donkeys, and mules in developing countries. Applied Animal Behaviour Science. 2010;126(3-4):109-118.
4. Pritchard JC, Lindberg AC, Main DCJ, Whay HR. Assessment of the welfare of working horses, mules and donkeys, using health and behaviour parameters. Preventive Veterinary Medicine. 2005;69(3-4):265-283.
5. Dalla Costa E, Murray L, Dai F, Canali E, Minero M. Equine on-farm welfare assessment: a systematic map of the literature. Animals. 2014;4(3):455-483.
6. Dalla Costa E, Minero M, Lebelt D, Stucke D, Canali E, Leach MC. Development of the Horse Grimace Scale as a pain assessment tool in horses. PLoS ONE. 2014;9(3):e92281.
7. Sommerville R, Brown AF, Upjohn M. A standardised equine-based welfare assessment tool used for six working equid populations. PLoS ONE. 2018;13(1):e0190653.
8. Regan FH, Hockenhull J, Pritchard JC, Waterman-Pearson AE, Whay HR. Clinical abnormalities in working donkeys and their associations with behaviour. Veterinary Record. 2016;179(6):137.
9. Brooke Hospital for Animals. State of the World's Donkeys. London: Brooke Hospital; 2019.
10. Pearson RA, Ouassat M. A Guide to Live Weight Estimation and Body Condition Scoring of Donkeys. Edinburgh: CTVM; 2000.
11. Upjohn MM, Attfield K, Pfeiffer DU, Verheyen KLP. Establishing a baseline: welfare assessment of working equids in Costa Rica. Proc 7 th Int Colloq Working Equids. 2014:156-164.
12. Sells PD, Pinchbeck G, Mezzane H, Ibourki J, Crane M. Pack type and working conditions influence welfare of working donkeys in Moroccan brick kilns. Equine Veterinary Journal. 2010;42(Suppl 38):644-649.
13. Regan FH, Hockenhull J, Pritchard JC, Waterman-Pearson AE, Whay HR. Identifying behavioural differences in working donkeys in response to stimuli. Equine Veterinary Journal. 2014;46(2):163-170.
14. Burden FA, Thiemann AK. Donkeys are different. Journal of Equine Veterinary Science. 2015;35(5):376-382.
15. McLean AK, Navas Gonzalez FJ. Can scientists influence donkey welfare? Historical perspective and future outlook. Journal of Equine Veterinary Science. 2018;65:25-32.
16. Popescu S, Diugan EA. The relationship between behavioural and other welfare indicators in working horses. Journal of Equine Veterinary Science. 2013;33(1):1-12.
17. Tadich T, Escobar A, Pearson RA. Husbandry and welfare aspects of urban draught horses in the south of Chile. Archivos de Medicina Veterinaria. 2008;40(3):267-273.
18. de Aluja AS, Tapia Perez G, Lopez Perez F, Rubio Lozano MS. Assessment of the welfare of working donkeys in Mexico City. Equine Veterinary Journal. 2005;37(2):7882.
19. Minero M, Dalla Costa E, Dai F, Murray LAM, Canali E, Wemelsfelder F. Use of Qualitative Behaviour Assessment as an indicator of welfare in donkeys. Applied Animal Behaviour Science. 2016;174:147-153.