Assessment of the welfare of working horses, mules and donkeys, using health and behaviour parameters

1 Preventive Veterinary Medicine xxx (2005) xxx xxx www.elsevier.com/locate/prevetmed 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Assessment of the welfare of working horses, mules and donkeys, using health and behaviour parameters J.C. Pritchard a,b, *, A.C. Lindberg a, D.C.J. Main a, H.R. Whay a a Department of Clinical Veterinary Science, University of Bristol, Langford, Bristol BS40 5DU, UK b Brooke Hospital for Animals, Broadmead House, 21 Panton Street, London SW1Y 4DR, UK Abstract Received 11 March 2004; received in revised form 9 February 2005; accepted 9 February 2005 Working animals provide an essential transport resource in developing countries worldwide. Many of these animals are owned by poor people and work in harsh environments, so their welfare is a cause for concern. A protocol was developed to assess the welfare of working horses, mules and donkeys in urban and peri-urban areas, using direct observation of health and behaviour parameters. In this study, 4903 animals used for draught, pack and ridden work in Afghanistan, Egypt, India, Jordan and Pakistan were assessed between December 2002 and April 2003. The data showed that donkeys were more likely than mules or horses to demonstrate avoidance or aggressive behaviour towards an observer, while horses were most likely to make a friendly approach. Fewer than 8% of working equines had abnormal mucous membranes, ectoparasites or poor coat condition. Body lesions occurred predominantly in the areas of the breast/shoulder, withers and girth in all three species, with mules having the highest prevalence of lesions in these areas (22.5, 21.3 and 28.4%, respectively). Among horses and donkeys, the prevalence of these lesions was influenced by the type of work carried out. Lesions on the head, neck, ribs, flank and tail base were seen in less than 10% of animals. Across all three species approximately 70% of animals were thin, having a body condition score (BCS) of 2 or less on a scale of 1 5 (1, very thin; 5, very fat) and more horses were in very thin condition (BCS 1) than mules or donkeys. Over 75% of animals demonstrated limb deformities and abnormalities of gait. The results of this study are being used as the initial stage of a long-term strategy to inform priorities for welfare interventions in working equines and to establish a welfare benchmark. Subsequent stages will rank the welfare concerns identified, assess the contributing risk factors and implement specific interventions to address these risks. Following intervention, success in * Corresponding author. Tel.: +44 117 928 9340; fax: +44 117 928 9582. E-mail address: joy.pritchard@bristol.ac.uk (J.C. Pritchard). 0167-5877/$ see front matter # 2005 Published by Elsevier B.V. doi:10.1016/j.prevetmed.2005.02.002

2 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 33 34 35 improving welfare will be measured by repetition of this protocol and comparison with the benchmark. # 2005 Published by Elsevier B.V. 36 38 37 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 Keywords: Working equine; Horse; Mule; Donkey; Welfare assessment; Developing countries 1. Introduction Brooke Hospital for Animals ( the Brooke ) is a UK charity founded in 1934 to provide veterinary treatment for equines working in developing countries and to advise their owners on good management practices. The Brooke has a network of static and mobile veterinary clinics based in major cities in Afghanistan, Egypt, India, Jordan and Pakistan. This paper describes collaboration between the Brooke and the University of Bristol to develop and implement a working equine welfare assessment system. Welfare assessment systems can be broadly categorised into animal-based or resource-based measures and different applications tend to draw from one or both of these types of measure (Main et al., 2003). The long-term aim of this collaboration is to build a strategic working methodology for the charity, starting with assessment of working equine welfare using animal-based measures (i.e. direct observations of health and behaviour parameters) as an alternative to recording resource inputs. At the next stage, the welfare issues identified will be ranked using expert consultation, followed by a comprehensive assessment of risks contributing to the most important issues identified. This information will be used as the basis for interventions to improve welfare. Knowledge of differences in the prevalence of each health and behaviour parameter between species and across work types will allow interventions to be targeted at specific problems in specific groups of animals. The success, or otherwise, of interventions will be evaluated using the same direct observations to identify the outcomes most relevant to the animals themselves, such as fewer or smaller body lesions, more positive behavioural interactions with humans or a lower prevalence of lameness within a population. There are an estimated 90 million equines in the developing world, with the highest population concentrations in central Asia and North and East Africa (FAO statistical database, 2003). Over 95% of all donkeys and mules and 60% of all horses are found in developing countries (Fielding, 1991) and the majority of these will be used for work. Recent information regarding the contribution of draught animal power to the economies of developing countries is scarce, although in 1988 it was estimated that working animals, including equines, produced 75% of traction energy in the developing world (US. Congress, Office of Technology Assessment, 1988) and it has been suggested that more than half of the world s population depends on animal power as its main energy source (Wilson, 2003). Today, draught animals and humans provide an estimated 80% of the power input on farms in developing countries (Pearson, 2005), but traction animals are often neglected in the allocation of resources such as food, shelter and appropriate equipment, because they belong to members of the poorest sections of society, who cannot afford motorised transport. Working equines are used for carting goods and people, carrying packs, bricks and other construction materials, riding, tillage, weeding, water-

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 3 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 carrying, rubbish collection, tourism and for ceremonial purposes such as weddings and festivals. A wide spectrum of welfare issues is encountered, including limb disorders, skin lesions and malnutrition. However, the prevalence of these problems within large populations and their relationship to work type are not known. Indirect methods of evaluating the welfare of animals are based on measuring the adequacy of inputs, such as resources and management provision (Wood et al., 1998; Bartussek, 1999). These methods indicate a risk of welfare problems rather than an actual measure of welfare state (Rousing et al., 2001). The advantage of such input-based assessment methods is that they are usually objective and repeatable; however, a positive score does not guarantee good welfare (Winckler and Willen, 2001; Whay et al., 2003). The use of direct, animal-based measurements to assess the welfare of farm and laboratory animals has increased in recent years. Scoring systems have been developed to assess lameness in dairy cattle (Whay, 2002), skin lesions in pigs (Leeb et al., 2001) and lameness in broiler chickens (Kestin et al., 1992). Direct observations provide the measure of welfare status that is most relevant to the animal itself. Although animal-based observations are often assessed subjectively, they should provide a more direct, and therefore more valid, assessment of welfare than resource measurements. Repeatability (precision) of the observations is an important consideration and subjective health and behaviour assessments can be very repeatable. Main et al. (2000) demonstrated that locomotion scoring of pigs could achieve a high level of repeatability between observers, while Hansen and Møller (2001) demonstrated that, after 30 min of training, six farmers were able to perform a temperament test on farmed mink and achieve 74 100% agreement in results with an experienced tester. Previous studies of equine welfare have used a combination of direct and indirect indicators. Most include body condition, sometimes with other animalbased measures, plus indirect measures in the form of resource examination and/or an owner questionnaire (Christie et al., 2003; Zanella et al., 2003). Animal-based measurements are particularly appropriate to situations where resource examination is not practical, as in the case of working equines. Housing, feed provision and other inputs cannot be measured during the working day; this would require a home visit for each animal. Owners do not keep animal health records for inspection and questionnaires carry a risk of social or cultural bias. Consideration of both health and behaviour is important when assessing welfare. Behaviour is the expression of an animal s perception of, and interaction with, its environment. In horses, direct observations of behaviour have been used to assess recovery from intestinal surgery (Durham et al., 2003) and arthroscopic surgery (Price et al., 2003). Behavioural observations used in farm animals include social behaviour, comfort behaviour, such as rising behaviour in dairy cattle, and standardised fear tests to measure the human-animal relationship (Sørensen et al., 2001). Rushen et al. (1999) concluded that rough or aversive handling of farm animals could substantially reduce their productivity and welfare. Health issues affecting an animal s welfare include acute diseases and disorders causing immediate suffering and long term, progressive conditions causing chronic pain (Rousing et al., 2001). Physical observations of particular relevance to equines include body condition score (Henneke et al., 1983; Carol and Huntingdon, 1988), hoof horn quality (Zenker et al., 1995) and skin turgor as an indicator of hydration status (Freeman et al., 1999).

4 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 The objective of this work was to develop a protocol for assessment of the welfare of working equines, using direct observations of their health and behaviour. A major requirement was that the protocol should be suitable for practical use under field conditions, so it was designed to take less than 10 min per animal, be applicable during the animals normal working day and most measures did not require the observer to touch the animal. 2. Methods 2.1. Welfare assessment protocol Using information obtained from published literature, field experience, a preliminary consultation with 15 experts in the field of working equine welfare (Lindberg et al., 2003) and an international meeting of Brooke Hospital field staff, a list of health and behaviour parameters was devised. Eight descriptors were used, including the date, observer, geographical region, animal and work type, followed by 41 observations of health and behaviour and a space for additional observations. Observations were recorded either as present/absent or as scores of severity. 2.2. Selection of animal-based welfare indicators A range of indicators was selected to be representative of both behaviour and health status. Measures of behaviour in relation to human approach and handling were considered to be important to the future planning of welfare intervention strategies. In this study, novel tests were used to assess the reaction of horses, mules and donkeys to human approach, proximity and touch. Clamping down the tail or tucking in the hindquarters ( tail tuck )in response to an observer walking down their side was recorded only for donkeys, because this response was not seen in mules or horses during previous field-observations. Indicators of health were selected to gain information about the three major issues identified by the experts: wounds, lameness and poor body condition. In addition, indicators of heat stress were proposed by the Brooke s veterinary staff. Some lesions, such as those at the commissures of the lips, on the cranial aspect of the carpus and at the point of the hock, were considered to be markers for particular risk factors; these were assessed independently of other lesions on the corresponding body areas. Pathognomonic lesions relating to two risk factors of concern, firing and tethering/hobbling, were also assessed. Gait abnormality was assessed over twelve paces at walk. The sole of the foot was examined on the right fore only. Where the sole surface of the hoof could not be examined due to a closed shoe (sole and frog covered with leather, rubber or a metal plate), this was recorded separately. 2.3. Field testing the checklist The draft welfare assessment checklist was categorised into observations of behaviour, general health parameters, body condition, limb disorders and lesions of skin and/or deeper

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 5 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 tissues. It was field-tested in Cairo, Egypt in July 2002. The practicality of assessing each indicator was tested under field conditions, using a wide range of situations where working equines are encountered. Over 7 days the checklist was used by two pairs of observers to assess 211 animals, including donkeys transporting goods to and from markets, riding donkeys, tourist horses, cart horses transporting goods or people and working equines being sold at markets. The two pairs of observers worked independently from each other on the same animals, and made continuous modifications to the checklist, in order to improve the ease and accuracy of observations. The optimum balance of speed and accuracy (compared with JCP as a gold standard ) occurred when observers worked in pairs, with one person making observations and the other prompting observations from the checklist and recording the results. Using this method, scoring a single animal took 5 7 min. 2.4. Baseline welfare assessment in Jordan, Afghanistan, India, Pakistan and Egypt As there is no accurate information available about the population of animals in the countries and regions visited, local knowledge was used to make a best estimate of the working equine population in the area accessible to each of the Brooke s clinics and outreach teams. This population of approximately 88,600 animals was sub-divided by species and type of work done. The sampling frame aimed to represent the proportions of species and work types within each area. Where large populations of a species and work type were present, the sampling frame aimed to assess at least 5% of the population. Where small sub-groups of species or work type were present, such as with ceremonial horses, the entire sub-group was assessed. Sampling was carried out in streets, markets, brick kilns (factories) and at tonga (carriage/cart) stands between December 2002 and April 2003. The work areas of different animal groups did not overlap and sampling was carried out in a different work area each day. A handbook of photographs and descriptions for each indicator, produced by ACL and JCP, was provided to observers. After initial training by JCP, HRW or both, the data were collected by eight observers working in pairs. 2.5. Data handling In each country, data were recorded by hand and entered into a dedicated web-based database (Neil Ambrose, Smart Tuna Ltd.). In the UK, the database was transferred to SPSS (SPSS Inc.) and analysed for prevalence of each health and behaviour parameter. Two animals were eliminated from the analysis due to missing data relating to work type and two groups of animals (pack mules, n = 9, and other donkeys, n = 3) were eliminated due to small sample size. Data were collated according to species and work type. The twelve work types identified in the sample were aggregated into four categories: draught, pack, ridden and other, which were mainly ceremonial horses used for weddings (see Table 1). Table 2 gives a description of the sex, age group and work type of the animals observed. Severity thresholds were considered as presence/absence of a health or behaviour parameter for analysis purposes. The observations of individual animals were compiled and expressed as the proportion of animals demonstrating the observed behaviour or health parameter within each species sampled.

6 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 199 200 201 202 203 204 205 206 207 208 209 Table 1 Types of work carried out by 4903 equine animals assessed in Afghanistan, Egypt, India, Jordan and Pakistan Aggregated category Components Draught Agriculture (e.g. ploughing, weeding) Brick kiln bricks transported by cart Tourism (carriage) Transport of goods by cart (including agricultural goods being taken to market) Transport of people (non-tourists, carriage) Pack Brick kiln bricks transported by pack Transport of goods by pack (including agricultural goods being taken to market) Ridden Riding (non-tourists) Tourism (riding) Other Ceremonial (e.g. wedding) Foal of working mother Other work type not covered elsewhere Groups of observations belonging to similar categories were also aggregated for further analysis. Lack of responsiveness to environment/handling was examined by aggregating scores for apathy/severe depression, lack of response to observer approach and to the observer walking down the animal s side. A measure of limb problems was derived from aggregated scores for firing lesions, tether/hobble lesions, swelling of tendons/joints, limb deformities, hoof walls too long or short, abnormal sole surface and abnormal hoof horn quality. The interactions between different measurements were examined using the Chi-squared test with significance measured using either the Monte-Carlo estimation or the exact test when computational limits allowed. Spearman rank correlation was then used to relate the measurements. The overall level of significance was set to P < 0.05. Table 2 Description of work type, sex and age group of 4889 equine animals assessed in Afghanistan, Egypt, India, Jordan and Pakistan, total number of animals = 4889 Species Donkey Mule Horse Work type (n) Work type (n) Work type (n) Total (2596) Draught (1773) Pack (272) Ridden (551) Draught (222) Total (2071) Draught (1659) Pack (50) Ridden (177) Other (185) Sex a Stallion 1972 1487 175 310 115 1071 938 31 50 52 Gelding 14 6 1 7 32 27 13 0 12 2 Mare 610 280 96 234 73 973 708 19 115 131 Age group (years) b Less than 5 552 329 73 150 30 260 165 20 40 35 5 15 1687 1217 167 303 146 1390 1113 27 122 128 Over 15 357 227 32 98 44 418 379 3 14 22 a Missing sex data for two mules. b Missing age group data for two mules and three horses.

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 7 210 211 3. Results 3.1. Animals sampled 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 Of the 4903 animals sampled, 4889 were included in all analyses; this comprised 2596 donkeys, 222 mules and 2071 horses. Across all species, the study sampled five to 18% of animals from the target populations. The largest number of animals (3654) were involved in draught work and mules in this sample were not used for any other type of work. Within the sample there were more entire male donkeys and mules than mares or geldings, across all work types. Among the horses, stallions were used more for draught and pack work and mares were used more for ridden and ceremonial work. Fewer than 3% of male animals were gelded; the number of geldings was higher among mules than horses or donkeys. Across all species the majority of animals observed were in the 5 15 year age group. 3.2. Assessment of welfare The observations of individual animals shown in Table 3 are expressed as the proportion of animals demonstrating the observed behaviour or health parameter within each species. The significance of differences in the prevalence of each observation between species is also shown. In response to an observer approaching the animal s head, 12.1% of donkeys showed a friendly approach (turning head towards the observer), 44.3% responded with avoidance (turning head away or moving away from the observer) or aggressive behaviour and 43.6% showed no response at all. Horses demonstrated significantly higher proportions (P < 0.001) of friendly approach (19.1%) and no response to the observer (54.7%) and a lower proportion of avoidance or aggression (26.0%) than donkeys. Twenty-eight percent of donkeys responded to the observer walking down their side and back again by clamping down the tail and in some cases tucking in the hindquarters as well. The proportion of horses avoiding chin contact by the observer s cupped hand was significantly higher (P < 0.001) than that of donkeys or mules. The observations of body condition showed that across all species approximately 70% of animals were thin, having a body condition score (BCS) of 2 or less on a scales of 1 5 (1, very thin; 5, very fat). More horses were in very thin condition (BCS 1) than mules or donkeys. The observations of general health showed that fewer than 8% of working equines had abnormal mucous membranes, missing teeth, ectoparasites or poor coat condition. A high proportion of animals (37.1% of donkeys, 45.9% of mules and 50.2% of horses) showed an increased skin tent duration, despite fewer than 4% of all species showing overt signs of heat stress. Limb-associated abnormalities were highly prevalent across all species, with 94.7% of working donkeys, 90.9% of working mules and 89.6% of working horses showing some degree of gait abnormality, ranging from mildly abnormal to severely lame. Cow hocked conformation, congenital and acquired angular limb deformities, swelling of tendons and joints and abnormal hoof horn quality were present in over 75% of all animals observed. The sole surface of the right fore foot was abnormal in over 80% of animals observed and in a further 4% it could not be assessed due to presence of a closed shoe.

8 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx Table 3 Behaviour and health parameters of 4889 working horses, mules and donkeys, expressed as a proportion within each species Species (n) % Within each species Significance of difference in Donkey (2596) Mule (222) Horse (2071) proportion between species (P) Observations of behaviour (% of animals) General attitude Alert 88.5 90.5 91.9 <0.01 Apathetic/severely depressed 11.5 9.5 8.1 Response to observer approach a No response 43.6 55.4 54.7 <0.001 Friendly approach 12.1 14.9 19.3 Avoidance/aggression 44.3 29.7 26.0 Walk down side b No response 9.4 9.5 10.6 <0.05 Responds 90.6 90.5 89.4 Tail tuck (donkeys only) c 28.2 n/a n/a n/a Avoids chin contact d 18.3 18.9 24.4 <0.001 General observations (% of animals) Body condition score (scales 1 5) 1 21.5 27.5 31.4 <0.001 2 48.8 49.1 38.4 3 5 29.7 23.4 30.2 Observations of health e Mucous membranes abnormal 2.9 1.8 4.7 <0.01 Lesions at commissures of lips f 30.1 54.8 31.9 <0.001 Teeth missing 3.0 1.9 1.8 <0.001 Molar hooks or sharp edges 76.5 85.1 80.3 <0.001 Eye(s) abnormal g 86.4 64.4 66.4 <0.001 Coat staring/matted/dry/uneven 2.5 1.8 2.3 Not significant Ectoparasites 7.6 1.4 5.1 <0.001 Diarrhoea under tail 8.7 19.4 18.2 <0.001 Skin tent (loss of elasticity) 37.1 45.9 50.2 <0.001 Heat stress h 1.0 1.8 3.9 <0.001 Firing lesions or scars f 21.7 8.1 19.2 <0.001 Tether/hobble lesions or scars f 88.8 79.3 62.1 <0.001 Carpal lesions or scars f 63.9 59.5 53.4 <0.001 Hock lesions or scars f 49.8 27.5 34.5 <0.001 Swelling of tendons/joints 80.0 79.7 89.1 <0.001 Limb deformity i 92.4 76.6 85.4 <0.001 Cow hocked conformation 87.0 84.2 77.3 <0.001 Hoof wall(s) too long 52.0 46.8 55.5 <0.05 Hoof wall(s) too short 25.8 26.1 30.9 <0.01 Hoof horn quality abnormal 88.6 76.6 81.7 <0.001 Sole surface abnormal (RF) 92.6 87.7 80.3 <0.001 Gait abnormal j 94.7 90.9 89.6 <0.001 Lesions of skin and/or deeper tissues k Head 5.5 1.8 4.7 <0.05

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 9 251 252 253 254 255 256 257 258 259 260 261 262 Table 3 (Continued ) Species (n) % Within each species Significance of difference in Donkey (2596) Mule (222) Horse (2071) proportion between species (P) Ears 0.7 0 0.3 Not significant Neck 1.5 1.8 1.5 Not significant Breast/shoulder 11.5 22.5 7.7 <0.001 Withers 10.2 21.3 13.2 <0.001 Spine 7.5 1.4 7.0 <0.05 Girth 18.3 28.4 19.2 <0.05 Belly 1.1 3.2 2.5 <0.01 Ribs/flank 3.2 2.7 2.9 Not significant Hindquarters 12.1 5.9 5.3 <0.001 Tail/tail base 5.5 9.1 5.3 <0.05 Forelegs (except carpus) 4.1 4.6 8.1 <0.001 Hindlegs (except hock) 5.3 6.4 6.5 Not significant n/a: not applicable. a Response to observer approaching the animal s head from 3 to 5 m away, at angle of approximately 458 (more acute if animal is wearing blinkers). Friendly approach: animal turns head towards observer. Avoidance/ aggression: animal does one or more of following: turns head away, moves away, flattens ears, attempts to bite or kick. b Response to observer walking down side of animal s body at distance of 30 cm from its side, turning at tail and walking back to head. Responds: any acknowledgment of observer s presence, e.g. ear turn, head turn, move away, kick. c Proportion of donkeys clamping down tail and/or tucking in hindquarters when observer was level with hindquarters during walk down side. d Proportion of animals avoiding contact or withdrawing head when hand was placed lightly under the chin. e Proportion of animals with signs of each condition. f Proportion of animals with lesion of any kind including hair loss, healed lesion, scar. g Proportion of animals with any abnormality of eye including ocular discharge. h Proportion of animals showing most or all of the following: flared nostrils, increased respiratory rate, increased respiratory depth with head movement, apathy. i Proportion of animals showing lateral or flexural abnormalities of the limbs, excuding cow hocked conformation. j Proportion of animals showing abnormalities of gait or overt lameness. k Proportion of animals with full thickness skin or deeper lesions measuring at least 2 cm 2cm or 1cm 4 cm. Firing, tether, carpus, hock and lip lesions scored previously were not included. Lesions of skin and deeper tissues were most prevalent on the breast/shoulder, withers and girth regions of all three species, with mules having a significantly higher frequency than others. Donkeys had a relatively high prevalence of lesions on the hindquarters (12.1%) compared with mules and horses. Tables 4 and 5 illustrate significant associations between type of work and measures of welfare in the donkeys and horses, respectively. In this sample, mules only carried out draught work so no work type comparisons could be made. Thirty-five of the 41 welfare observations made in working donkeys showed a significant difference (P < 0.01) in prevalence between work types (Table 4). Pack donkeys were more likely to be alert, friendly and responsive and less likely to tuck in their tails in response to an observer walking down their side than those doing draught or ridden work. They were also most likely to be thin, with 82% of pack donkeys in body condition 2 or 1, compared with 75.6%

10 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx Table 4 Significant associations between work type and behaviour and health parameters in 2596 working donkeys, expressed as a proportion within each work type Work type (n) % Within each work type Significance of difference in Draught (1773) Pack (272) Ridden (551) Observations of behaviour (% of animals) General attitude Alert 87.6 95.6 87.7 <0.01 Apathetic/severely depressed 12.4 4.4 12.3 Response to observer approach a No response 45.1 61.0 30.3 <0.001 Friendly approach 11.3 15.8 12.9 Avoidance/aggression 43.7 23.2 56.8 Walk down side b No response 8.8 15.1 8.5 <0.01 Responds 91.2 84.9 91.5 Tail tuck (donkeys only) c 29.6 9.2 33.8 <0.001 Avoids chin contact d 17.9 17.0 20.5 <0.05 General observations (% of animals) Body condition score (scales 1 5) 1 25.5 20.6 9.3 2 49.1 61.4 41.6 3 5 25.4 18.0 49.2 <0.001 Observations of health e Mucous membranes abnormal 2.9 1.1 3.4 Not significant Lesions at commissures of lips f 40.8 3.3 8.9 <0.001 Teeth missing 3.1 1.8 3.5 <0.05 Molar hooks or sharp edges 77.5 84.2 69.3 <0.001 Eye(s) abnormal g 88.5 53.3 95.6 <0.001 Coat staring/matted/dry/uneven 1.3 9.6 2.7 <0.001 Ectoparasites 7.5 3.3 9.8 <0.05 Diarrhoea under tail 11.8 4.0 1.1 <0.001 Skin tent (loss of elasticity) 42.0 11.8 33.8 <0.001 Heat stress h 1.0 1.9 0.5 <0.01 Firing lesions or scars f 24.0 5.1 22.2 <0.001 Tether/hobble lesions or scars f 86.4 85.7 98.0 <0.001 Carpal lesions or scars f 67.3 49.3 59.9 <0.001 Hock lesions or scars f 50.3 37.1 54.5 <0.01 Swelling of tendons/joints 80.9 78.3 77.7 <0.01 Limb deformity i 93.9 88.6 89.5 <0.001 Cow hocked conformation 88.7 77.2 86.4 <0.001 Hoof wall(s) too long 51.8 62.1 47.5 <0.001 Hoof wall(s) too short 26.6 40.1 16.0 <0.001 Hoof horn quality abnormal 65.0 5.2 18.3 <0.001 Sole surface abnormal (RF) 95.4 94.9 89.8 <0.001 Gait abnormal j 95.5 98.2 90.4 <0.001 Lesions of skin and/or deeper tissues k Head 6.9 0.4 3.6 <0.001 proportion between work types (P)

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 11 263 264 265 266 267 268 269 270 271 272 273 274 Table 4 (Continued ) Work type (n) % Within each work type Significance of difference in Draught (1773) Pack (272) Ridden (551) proportion between work types (P) Ears 0.8 0.7 0.2 Not significant Neck 2.1 0 0.4 <0.001 Breast/shoulder 16.2 1.8 0.2 <0.001 Withers 8.2 1.8 0.2 <0.001 Spine 5.1 1.8 0.6 <0.001 Girth 16.9 1.1 1.5 <0.001 Belly 1.4 1.1 0.2 Not significant Ribs/flank 3.6 6.3 0.5 <0.001 Hindquarters 13.6 27.3 0 <0.001 Tail/tail base 6.4 7.8 1.6 <0.001 Forelegs (except carpus) 5.3 0.7 2.0 <0.001 Hindlegs (except hock) 5.4 14.0 0.7 <0.001 n/a: not applicable. a Response to observer approaching the animal s head from 3 to 5 m away, at angle of approximately 458 (more acute if animal is wearing blinkers). Friendly approach: animal turns head towards observer. Avoidance/ aggression: animal does one or more of following: turns head away, moves away, flattens ears, attempts to bite or kick. b Response to observer walking down side of animal s body at distance of 30 cm from its side, turning at tail and walking back to head. Responds: any acknowledgment of observer s presence, e.g. ear turn, head turn, move away, kick. c Proportion of donkeys clamping down tail and/or tucking in hindquarters when observer was level with hindquarters during walk down side. d Proportion of animals avoiding contact or withdrawing head when hand was placed lightly under the chin. e Proportion of animals with signs of each condition. f Proportion of animals with lesion of any kind including hair loss, healed lesion, scar. g Proportion of animals with any abnormality of eye including ocular discharge. h Proportion of animals showing most or all of the following: flared nostrils, increased respiratory rate, increased respiratory depth with head movement, apathy. i Proportion of animals showing lateral or flexural abnormalities of the limbs, excuding cow hocked conformation. j Proportion of animals showing abnormalities of gait or overt lameness. k Proportion of animals with full thickness skin or deeper lesions measuring at least 2 cm 2cm or 1cm 4 cm. Firing, tether, carpus, hock and lip lesions scored previously were not included. of draught and 50.9% of ridden donkeys. Draught donkeys were most likely to have lesions at the commissures of the lips, on the cranial aspect of the carpus, on the head, breast/ shoulder, withers, spine (all significant at P < 0.001) and girth (P < 0.01). Those carrying packs were most likely to have lesions on the hindquarters, tail/tail base and hindlegs (P < 0.001). Draught donkeys also had the highest prevalence of swollen tendons and joints, deformed limbs, cow hocked conformation, abnormal hoof horn quality and abnormal sole surface. However, pack work was more associated with hoof walls being too long or too short and abnormal gait. Table 5 summarises the horses observed. Those which were ridden or doing other work types (mainly ceremonial) were most likely to be alert, to avoid chin contact by an observer and to be in good to fat body condition (BCS 3 5). Like donkeys, draught horses had a relatively high prevalence of lesions at the commissures of the lips, on the cranial

12 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 275 276 277 278 279 280 281 282 283 284 Table 5 Significant associations between work type and behaviour and health parameters in 2071 working horses, expressed as a proportion within each work type Work type (n) % Within each work type Significance of difference Draught (1659) Pack (50) Ridden (177) Other (185) Observations of behaviour (% of animals) General attitude Alert 90.5 92.0 97.7 99.5 <0.001 Apathetic/severely depressed 9.5 8.0 2.3 0.5 Response to observer approach a No response 58.3 48.0 22.0 55.1 <0.001 Friendly approach 17.4 30.0 32.2 21.1 Avoidance/aggression 24.3 22.0 45.8 23.8 Walk down side b No response 8.9 12.2 14.4 22.2 <0.001 Responds 91.1 87.8 85.6 77.8 Avoids chin contact c 23.9 12.0 24.1 32.4 <0.05 General observations (% of animals) Body condition score (scales 1 5) 1 38.2 6.0 2.3 4.9 <0.001 2 37.3 62.0 36.4 43.8 3 5 24.6 32.0 61.9 50.8 Observations of health d Mucous membranes abnormal 5.4 2.0 3.4 0.5 <0.01 Lesions at commissures of lips e 35.9 20.0 14.7 16.3 <0.001 Teeth missing 2.0 2.1 0.6 0.6 <0.001 Molar hooks or sharp edges 82.5 72.9 64.2 76.5 <0.001 Eye(s) abnormal f 71.5 20.0 71.8 27.7 <0.001 Coat staring/matted/dry/uneven 2.1 18.0 0 1.6 <0.001 Ectoparasites 4.8 14.0 10.8 0 <0.01 Diarrhoea under tail 21.8 4.0 1.7 6.0 <0.001 Skin tent (loss of elasticity) 54.3 12.0 35.4 37.5 <0.001 Heat stress g 4.6 0 1.1 0.5 <0.05 Firing lesions or scars e 22.8 2.0 6.8 3.3 <0.001 Tether/hobble lesions or scars e 64.0 50.0 63.6 47.0 <0.001 Carpal lesions or scars e 57.1 28.0 38.9 39.9 <0.001 Hock lesions or scars e 34.0 20.0 48.6 29.5 <0.01 Swelling of tendons/joints 91.3 84.0 64.3 91.3 <0.01 Limb deformity h 89.6 86.0 55.7 73.2 <0.001 Cow hocked conformation 81.2 95.9 40.9 72.0 <0.001 Hoof wall(s) too long 56.1 56.0 54.8 50.5 <0.01 Hoof wall(s) too short 30.7 46.0 11.4 47.8 <0.001 Hoof horn quality abnormal 87.0 26.0 55.1 72.1 <0.001 Sole surface abnormal (RF) 88.3 81.6 80.3 0.7 <0.001 Gait abnormal i 93.7 93.9 61.0 78.1 <0.001 in proportion between work types (P) Lesions of skin and/or deeper tissues j Head 5.5 0 0 3.3 <0.01 Ears 0.4 0 0 0 Not significant

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 13 275 276 277 278 279 280 281 282 283 284 285 286 287 Table 5 (Continued ) Work type (n) % Within each work type Significance of difference Draught Pack Ridden Other in proportion between (1659) (50) (177) (185) work types (P) Neck 1.8 0 0 0.5 Not significant Breast/shoulder 9.0 0 0 5.4 <0.001 Withers 14.8 20.0 4.0 5.4 <0.001 Spine 7.7 10.0 5.1 1.1 Not significant Girth 22.8 2.0 3.2 6.5 <0.001 Belly 3.0 0 0 0.5 Not significant Ribs/flank 3.2 0 3.4 0.5 Not significant Hindquarters 6.2 6.0 1.1 1.6 <0.01 Tail/tail base 6.4 0 0 1.6 <0.01 Forelegs (except carpus) 9.4 0 5.1 1.7 <0.01 Hindlegs (except hock) 7.6 0 2.8 2.2 <0.01 a Response to observer approaching the head from 3 to 5 m away at angle of approximately 458 (more acute if animal wearing blinkers). Friendly approach: animal turns head towards observer. Avoidance/aggression: animal does one or more of following: turns head away, moves away, flattens ears, attempts to bite or kick. b Response to observer walking down side of animal s body at distance of 30 cm from its side, turning at tail and walking back along side to head. Responds: any acknowledgment of observer s presence, e.g. ear turn, head turn, move away, kick. c Proportion of animals avoiding contact or withdrawing head when hand was placed lightly under the chin. d Proportion of animals with signs of each condition. e Proportion of animals with lesion of any kind including hair loss, healed lesion, scar. f Proportion of animals with any abnormality of eye including ocular discharge. g Proportion of animals showing most or all of the following: flared nostrils, increased respiratory rate, increased respiratory depth with head movement, apathy. h Proportion of animals showing lateral or flexural abnormalities of the limbs, excuding cow hocked conformation. i Proportion of animals showing abnormalities of gait or overt lameness. j Proportion of animals with full thickness skin or deeper lesions measuring at least 2 cm 2cm or 1cm 4 cm. Firing, tether, carpus, hock and lip lesions scored previously were not included. aspect of the carpus and on the head, breast/shoulder and girth, but the prevalence of lesions on the withers was higher in pack horses than draught horses and those doing ridden or other work. Abnormal gait and sole surface, cow hocked conformation and limb deformities were more likely to be seen in pack or draught horses than ridden or other work types. Table 6 shows correlations between aggregated behaviour and health parameters in the working equines observed. Correlation coefficients exceeding r = 0.3 included low body condition score with systemic health abnormalities (r = 0.316), low body condition score with lesions of skin and deeper tissues (r = 0.371) and limb problems with abnormal gait (r = 0.337). 4. Discussion The 4903 horses, mules and donkeys assessed do not necessarily represent the welfare status of all working equines in developing countries worldwide because conditions vary

14 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 Table 6 Correlations between aggregated behaviour and health parameters of 4889 working equine animals assessed in Afghanistan, Egypt, India, Jordan and Pakistan Behaviour and health parameters Correlation coefficient P Lack of responsiveness to environment/handling a Low body condition score 0.150 <0.001 Lesions of skin and deeper tissues 0.102 <0.001 Abnormal gait 0.058 <0.001 Limb problems 0.019 Not significant Systemic health abnormalities 0.002 Not significant Low body condition score Lesions of skin and deeper tissues 0.371 <0.001 Systemic health abnormalities b 0.316 <0.001 Abnormal gait 0.132 <0.001 Limb problems c 0.047 0.001 Lesions of skin and deeper tissues Systemic health abnormalities b 0.203 <0.001 Abnormal gait 0.096 <0.001 Limb problems 0.019 Not significant Limb problems c Abnormal gait 0.337 <0.001 Systemic health abnormalities b 0.046 0.001 Significance at P < 0.01 level is reported. a Aggregated score: general attitude + responsiveness to observer approach + responsiveness to observer walking down side. b Aggregated score: mucous membranes + coat condition + diarrhoea + skin tent + heat stress. c Aggregated score: firing lesions + tether/hobble lesions + swelling of tendons/joints + deformed limbs + hoof too long + hoof too short + sole surface abnormal (RF) + hoof horn quality. within and between countries. Accurate population demographics are not known so fully representative sampling is difficult. Brooke clinics are based in and around large towns and cities where work pressures on equines are thought to be greatest, therefore agricultural draught animals are under-represented in the sample. The four work type categories used in the analysis represent all major uses of working equines in urban and peri-urban areas and this study is probably the largest assessment of the welfare of working equines carried out to date. In order to maintain a standardised assessment, the observations were made by eight assessors who were trained and supplied with detailed guidance notes and photographs illustrating the scores. The assessors also received training in the field in their own countries. Two inter- and intra-observer repeatability tests carried out on this group of observers showed good agreement and will be reported elsewhere, as will an analysis of observations according to country and region. The welfare assessment protocol was based on direct observations of animals and designed for use in any situation where working equines can be observed, regardless of their breed or work type. The requirement for disturbance to the animals work was kept to a minimum, although it was necessary to look under harness and remove pack saddles. The major health and behaviour parameters to be assessed were identified by consultation with

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 15 306 307 308 309 310 311 311 312 312 313 314 313 315 314 316 317 315 318 316 319 320 317 321 318 322 319 323 320 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 15 experts in the field and by Brooke s international field staff, to ensure a broad range of opinion on its content and validity. The protocol meets the requirements for satisfactory welfare indicators listed by Sørensen et al. (2001), who stated they will: 1. Describe relevant and significant aspects of what matters from the point of view of the animals. 2. Express changes over time (i.e. at subsequent assessments, changes in welfare could be demonstrated using the indicators selected). 3. Be capable of being influenced by decisions taken by the individual farmer (in this case, the equine owner or user). 4. Be measurable in a relatively cheap and easy manner. The novel behaviour observations used in this protocol were developed to give indications of the responsiveness of an animal to the environment and to attempt to identify fearfulness. Fear is a negative motivational affective state and in a strong or prolonged form constitutes suffering (Fraser and Duncan, 1998). Working equines may be unresponsive due to disease, exhaustion, over-stimulation by a crowded and noisy city environment or to avoid soliciting harsh handling. Animals displaying fear behaviour are often exposed to adverse handling procedures because they react inappropriately to handling (Rousing et al., 2001). In equine species, fear behaviour presents a serious risk of injury to handlers, resulting in a cycle of increasing severity of restraint and increasing fear. Where the person who works the animal is not the owner, as with brick kiln donkeys in Cairo or hired riding mules in north India, the bond of interdependence is broken, so the risk of adverse handling is thought to increase. A limitation of these behaviour observations is that the response to an unfamiliar person may differ from the response to the regular handler, as observed in comparable tests on dairy calves (de Passillé et al., 1996). An assessment of the quality of the human-animal bond is relevant to the development of welfare interventions for working equines because, without a degree of social bonding between the owner or user and his animal and in the absence of enforced legal protection for the animal s welfare, there is little motivation to improve welfare beyond the minimum necessary to enable it to earn money. The study demonstrates a high prevalence of abnormalities of teeth, eyes and limbs across all three species. Dental rasping is not carried out in most working equines; this may be due to sparse availability, cost or quality of veterinary services and lack of knowledge on behalf of owners. Eye abnormalities seen ranged from mild discharge to signs of ocular pain, keratitis, uveitis and blindness. Determination of aetiology and risk factors for eye abnormalities in working equines is one potential area for future study. The prevalence of increased skin tent duration was 37.1% of donkeys, 45.9% of mules and 50.2% of horses. Observations of skin tent duration, a clinical indicator of dehydration, were made in the winter and spring periods. In summer, the prevalence of both increased skin tent duration and signs of heat stress would be expected to be much higher. There was a highly significant difference (P < 0.001) between the species. Assuming that the clinical indicators of dehydration are equally appropriate across all three species, it is unsurprising that donkeys, which are adapted to arid environments and conserve body water in conditions of water deprivation (Maloiy and Boarer, 1971; Yousef, 1991), show the lowest

16 J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 prevalence of increased skin tent duration. The amount of suffering caused by dehydration is unknown, but where animals are working in ambient temperatures of up to 48 8C this is an area of serious concern for their welfare. Only one foot of each animal was lifted to minimise disturbance through handling and to maximise the safety of the observer. For uniformity, the sole of the right fore foot was examined in every case. Swollen tendons and joints, angular limb deformities and abnormalities of the hoof horn and sole surface were present in over 75% of equines observed and have serious welfare implications for animals which may be required to work at trot or canter on roads for several hours per day. Ninety percent of working equines showed a gait abnormality over 12 paces at walk; these varied from uneven gait in animals with poor limb conformation through mild to severe lameness. A standard lameness examination includes a 20 m trot away and then back towards the observer. This was not possible under field conditions due to time constraints, crowding in observation areas and animals being untrained to trot in-hand. In addition, some of the equines were examined while harnessed to loaded carts. A standard lameness examination would be expected to identify more lame animals than the method used. In dairy heifers, lameness was found to be associated with hyperalgesia, which is a component of pain (Whay et al., 1997). Conditions of the foot causing lameness in cattle and sheep were given pain scores of 6 8 (scale: 0, no pain; 10, the worst pain imaginable) by the majority of a group of experts from the UK Cattle Veterinary Society and Sheep Veterinary Society, respectively (Scott et al., 2003). It is reasonable to assume that a high proportion of equines showing lameness and swollen tendons/joints will be suffering pain throughout their working day. Tables 4 and 5 illustrate associations between different welfare measures and the type of work carried out by the horse or donkey. Some observations are easier to explain than others. For example, in many draught equines a bit is used for steering and braking, so lesions at the commissures of the lips are expected to be more common than in pack animals which are usually not bitted. The cause of lesions on the body is not understood in some cases, although the site of some lesions is likely to be related to the type and position of the girth, breast strap and saddle. Draught animals may also be more likely to fall on roads at high speed than those doing other types of work, causing lesions on the cranial aspect of the carpus. Pack donkeys commonly work in brick kilns or on construction sites and are driven using a stick to beat the hindquarters and tail base, which may explain the greater prevalence of lesions in these body areas. However the proposed behavioural marker for beating (tail tuck) is less prevalent in pack donkeys than in draught or ridden donkeys. This could be due to the pack animal s ability to attempt to escape a raised stick by running forward, while a harnessed or ridden donkey cannot avoid contact by moving away so responds by tucking in the hindquarters and tail. The results of the correlations between aggregated health and behaviour parameters are shown in Table 6. Although no correlation coefficients were greater than r = 0.38, they indicate the presence of some complex interactions between groups of parameters. The higher correlation coefficients were those between expected parameters; for example, those correlating limb problems with abnormal gait (r = 0.337) and correlating low body condition score with systemic health abnormalities (r = 0.316). In addition, the highest correlation coefficient was that between low body condition score and lesions of skin and deeper tissues (r = 0.371). Low body condition score is an indicator of reduced body fat

J.C. Pritchard et al. / Preventive Veterinary Medicine xxx (2005) xxx xxx 17 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 (Henneke et al., 1983); consequently, thin animals may have less natural padding protecting them from pressure, friction and shear lesions caused by harness. Interestingly, lack of behavioural responsiveness showed correlation coefficients that were always below r = 0.15. This may be surprising, as depression and behavioural unresponsiveness could be expected to result from lesion pain, or the negative energy balance causing low body condition score; however, stimulation by the approach of an observer or from a noisy environment may over-ride chronic physical states. Although the protocol gives a snapshot of the animals welfare at a single point in time, most of the observations were chosen as indicators of problems that have developed over the previous weeks or months. A limitation is that most observations (with the exception of some behavioural tests, such as alertness and friendly approach) reflected adverse welfare rather than looking for positive measures of good welfare. Identifying good aspects of welfare is particularly important if information is to be fed back to owners. Positive behaviour patterns, such as play or grooming, are event behaviours and as such cannot be induced by a simple test or interaction, so are not suitable for this type of snapshot assessment. However, in the context of this study, the absence of health problems and observations of negative behaviour can be presented to owners as a positive outcome. Where the prevalence of a welfare issue was found to be high, such as with eye abnormalities or gait abnormalities, a more detailed set of observations could be formulated in order to identify specific treatable or avoidable conditions. The welfare significance of some animal-based observations, such as abnormal gait, is beginning to be understood. In others, such as where an animal carries a pack overlying large skin lesions or pulls a cart for 8 h without water, the welfare significance is as yet unquantified and it may be necessary to extrapolate from analogous situations in humans or other animal species in order to inform priorities for intervention. When the most urgent welfare problems have been identified and ranked, risk factor analysis is the next complex area to be addressed, for example, using resource examination and detailed questionnaires for working equine owners and users. Repetition of the welfare assessment can then be used to measure the effect of targeted resource or management interventions. Acknowledgements This study was supported and funded by the Brooke Hospital for Animals. The authors would like to thank Drs. Amro Hassan Abdullah, Ahmad Raza Khan, Mohammed Hammad El-Shahat, Shabir Ahmed Mir, Ali Ahmed Mohammed Twaissi and Mr. Kartick Satyanarayan for their assistance with data collection. We would also like to thank all the owners who kindly permitted examination of their horses, mules and donkeys. References Bartussek, H., 1999. A review of the animal needs index (ANI) for the assessment of animals well-being in the housing systems for Austrian proprietary products and legislation. Livest. Prod. Sci. 61 (2 3), 179 192.