A welfare assessment scoring system for working equids—A method for identifying at risk populations and for monitoring progress of welfare enhancement strategies (trialed in Egypt)

Applied Animal Behaviour Science 176 (2016) 52–62

Contents lists available at ScienceDirect

Applied Animal Behaviour Science journal homepage: www.elsevier.com/locate/applanim

A welfare assessment scoring system for working equids—A method for identifying at risk populations and for monitoring progress of welfare enhancement strategies (trialed in Egypt) Ahmed B.A. Ali a,c,∗ , Mohammed A. El Sayed b , Mohamed Y. Matoock c , Manal A. Fouad c , Camie R. Heleski a a

Animal Behavior and Welfare Group, Michigan State University, East Lansing, MI 48824, United States Cairo Clinic, The Brooke Hospital for Animals, Cairo, Egypt c Animal Behavior and Management, Veterinary Medicine, Cairo University, Egypt b

a r t i c l e

i n f o

Article history: Received 28 January 2015 Received in revised form 17 December 2015 Accepted 23 December 2015 Available online 29 December 2015 Keywords: Working equids Welfare assessment Equine behavior Equine welfare

a b s t r a c t There are an estimated 112 million horses, donkeys and mules (i.e., working equids) in developing regions of the world. Though their roles are often fundamental to the well-being of the families they work for, their welfare is often severely compromised due to the limited resources and/or limited knowledge base of owners. The main objective of this study was to develop a multifactorial welfare assessment score for accurate, comprehensive, quick and reliable assessment of these equids. A total of 5248 working equids (n = 2198 horses, 2640 donkeys, 410 mules) were assessed between February 2012–January 2014. Equids were divided into categories based on the three species involved, as well as the four work types involved (transporting goods or people by cart, ridden (e.g., in tourist locations), or working in brick kilns). Analysis of variance “ANOVA” was used to compare differences between groups with ␣ set at 0.05. In terms of behavioral measures, the most at-risk equids appeared to be horses who pulled goods by cart with 20.7% showing a depressed attitude and 22.6% being unresponsive to an observer’s approach (significantly greater than the other species and the other work types, P < 0.05). Mules who pulled goods by cart showed 30.8% avoiding an observer’s approach, 42.7% avoiding chin contact and 14.2% showing an aggressive response to observer (significantly greater than the other species and the other work types, P < 0.05). In terms of physical measures, 21.6% of donkeys who pulled goods by cart had harness-induced lesions and 21.9% showed evidence of firing-type lesions (significantly greater than the other species and other work types, P < 0.05). Mules who pulled goods by cart had the highest prevalence of mistreatmentinduced lesions at 36.7% (significantly greater than the other species and other work types, P < 0.05). From a positive perspective, horses used for riding or transporting people by cart (e.g., most often animals working in tourist areas) were most likely to be in a healthy physical state (over 85% for both categories; significantly greater than other species and other work types, P < 0.05). To conclude, this welfare assessment scoring system met our initial objective of being a useful tool in identifying which equids had the most significant welfare problems (i.e., which species, work type, age and sex). This, in turn will help in selecting appropriate interventions, and in targeting interventions toward the most vulnerable equids. © 2015 Elsevier B.V. All rights reserved.

1. Introduction According to the World Health Organization (WHO), working animals are defined as those animals that provide an essential resource for their owners who live in poverty (Perry, 2002;

∗ Corresponding author at: Department of Animal Science, Michigan State University, 474 S. Shaw Lane 1250 Anthony Hall East Lansing MI 48824, United States. E-mail addresses: [email protected], [email protected] (A.B.A. Ali). http://dx.doi.org/10.1016/j.applanim.2015.12.001 0168-1591/© 2015 Elsevier B.V. All rights reserved.

Thornton et al., 2002). Pritchard (2010) concluded that transport animals perform a wide variety of economic, labour reducing and social roles in a broad range of developing environments. After cattle, working equids are considered the main working animals worldwide (Blench and MacDonald, 2006); there are an estimated 112 million working horses, donkeys and mules in the developing world (FAOSTAT, 2011). They make a substantial socioeconomic contribution to their communities according to non-governmental organizations (NGOs), yet there is a scarcity of published scientific evidence (Chang et al., 2010; Velázquez-Beltrán et al., 2011).

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

Traction animals are often neglected in the allotment of resources such as shelter, food, and appropriate harnessing as they are usually owned by the poorest section of the society (Mekuria et al., 2013). Veterinary care, adequate feed and a good quality harness or even harness adjustments are mainly supplied by NGOs. Fortunately, there are a few UK-based NGOs that specifically address health and welfare issues relating to working equids. In addition, there are some smaller, independently funded organizations that work in Egypt. Usually these organizations provide veterinary care from fixed or mobile clinics. Services include preventive care (e.g., anthelmintic administration), services for acute problems (e.g., wounds and colic), and supportive care for chronic problems such as lameness, as well as improving social awareness about the best management practices for these animals (http://www.thebrooke.org, 2013). Usually NGOs operating in the working equid sector define their own objectives with regard to interventions (Upjohn et al., 2014). In addition, it is imperative that their objectives and activities make a positive difference to the community in which they operate. Thus, an effective monitoring and evaluation (M&E) process is required to ensure that these organizations are able to investigate whether their objectives are being achieved, and also to monitor the efficiency of their activities in improving equid welfare. Such processes should be done on a regular basis, and involve a continuous process of data gathering and analysis (Poate, 1993; Martin, 2001). Due to the scarcity of NGO resources, M&E is considered essential as it enables the organization to identify and evaluate the most efficient allocation of these resources in addition to following up on the fulfillment of the objectives granted by such resources. Thus, formal application of precise M&E has risen on the NGO development agenda in recent years (Wallace, 2010) and numerous guidelines and general literature on effective M&E activities are publicly available (Wood et al., 1998; Roche, 1999; Davies, 2001). For those concerned about working equid welfare, there is an additional ethical necessity for the assessment of the working equids’ welfare status in order to identify the existing welfare problems, as well as risks for other potential problems that may occur. M&E exercises help to identify the optimum solutions that can be applied to welfare problems. Several comprehensive studies on the lives of working equids, and the influence of their lifestyles on their health, behavior and in turn their welfare status have been conducted over the past years (de Aluja, 1998; Pritchard et al., 2005; Swann, 2006; Tadich et al., 2008; Burn et al., 2010b; Popescu and Diugan, 2013). Working equids are managed differently from most stabled equids (e.g., leisure horses and horses engaged in competitive events) as they are neither kept in fully equipped group housing stables, nor receive the same type of husbandry, nutrition, veterinary care and handling as many of their counterparts. Moreover they are obliged to work long periods, pull or carry heavy loads, and are often exposed to harsh environmental conditions. As such, reliance on resource based or input based methods for evaluating the welfare status of working equids may not be the best methodology (Wood et al., 1998; Bartussek, 1999). Although use of resource based methods is usually objective and repeatable, it mainly indicates the risk of welfare problems, rather than actual and existing ones (Rousing et al., 2001). Moreover, a positive score does not guarantee good welfare status (Winckler and Willen, 2001; Whay et al., 2003). For working equids in developing parts of the world, animal based measurements or direct observations are considered to be more reliable, relevant and suitable for assessing the status of these animals. Resource based measurement is rarely practical in many cases, and there are almost never records kept for animals describing their past treatment, vaccination/deworming history or even disease history.

53

Direct observation has shown to be highly effective for assessing the immediate welfare status of the working equid; however, one obstacle to this methodology is that the fast working daily rhythm of working equids means that it is necessary to perform the assessment very quickly. Another potential disadvantage may be low intra- and/or inter-observer reliability of the assessments. Few studies have been conducted to examine this obstacle. However, Main et al. (2000) demonstrated that a high level of observer repeatability could be achieved in locomotion score of pigs, and Hansen and Møller (2001) concluded that after 30 min of training, six farmers were able to achieve 74–100% agreement when performing temperament tests on farmed mink. The main objectives of this study were (1) to build up a multifactorial welfare assessment system for accurate, comprehensive and reliable welfare assessment of working equids, which NGOs, practitioners, veterinarians and researchers can use to evaluate the welfare status of working equids, and (2) monitor the progress of NGO welfare intervention strategies over time, and identify high need areas for implementation of welfare improvement strategies. It is important that such a welfare scoring protocol can identify which type of equid, in terms of species, age, sex and type of work, has the most at risk welfare status, as well as to identify types of welfare impairment (e.g., behavior, body condition score (BCS), body lesions etc.). This will help the assessor to decide upon necessary interventions. It has been repeatedly shown (Burn et al., 2010b; Popescu and Diugan, 2013) that when a working equid’s welfare status is improved, that animal’s wellbeing and longevity are typically enhanced. This, subsequently, has an economic benefit to the family it works for, as less money needs to be spent on health interventions, fewer days are lost to the poor health of the equid, and fewer resources go toward animal replacement. 2. Methods All research protocols were approved by the Veterinary Medicine College, Cairo University, Animal Behavior and Welfare Committee prior to the start of data collection. 2.1. Welfare assessment scoring system As a starting point for this study, we used published literature investigating the welfare of working equids. These included the welfare assessment protocol developed by Pritchard et al. (2005), the behavioral welfare indicators described in detail by Burn et al. (2010a) and also modifications made by Popescu and Diugan (2013). The welfare assessment scoring system was refined using the experience in the field of the first and last authors, and consultation with thirteen working equid experts via a written questionnaire. 2.2. Working equids welfare assessment Based on our literature reviews, there are no official data available about the total population of working equids in the visited regions in Egypt. Estimated numbers are based on local knowledge and estimates done by Brooke (http://www.thebrooke.org, 2013). Both were used to make the best estimates for the total number for equids working in each region. The aim of our sampling technique was to assess at least 10% of the population among visited sites, while appropriately representing the actual proportion of each species and working equid type within each region. Multiple regions inside the Cairo and Giza governorates that were visited regularly for this purpose include: Nazzlet Elsemman, Abo Seir, Kasr Elnile St. and Elkanater Elkhyria. These were visited to assess riding horses (R) and equids that are used to transport people by carts (TPC). Nahya, Elbragil, Tersa, Rod Elfarag and Mansheyt

54

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

Table 1 Description of work type, age and sex of 5248 working equids assessed in Egypt. Speciesa (n) Horse (2198)

Donkeys (2640)

Work type (n) TGC (1142)

Mules (410)

Work type (n)

Work type (n)

Total

TPC (570)

R (486)

Total

TGC (1500)

TBC (1140)

Total

TGC (160)

TBC (250)

Sex groupsb M 797 103 G 242 F

375 70 125

114 87 285

1286 260 652

1161 98 241

1140 0 0

2301 98 241

94 0 66

171 0 79

265 0 145

Age groupsc Y 230 MA 735 177 O

105 363 102

77 244 165

412 1342 444

400 797 303

287 781 72

687 1578 375

48 80 32

61 138 51

109 218 83

a b c

Equid’s species (Horses (H), Mules/Hinnies (M) or Donkeys (D)). Equid’s sex (Male, intact (M), Gelding (G) or Female (F)). Age group (Young (Y) = five years of age and younger, Medium (MA) = more than five, and up to and including fifteen years, and Old (O) = over fifteen years).

Nasser were visited for assessment of equids transporting goods by cart (TGC). Helwan brick kilns were visited to assess equids transporting bricks by cart (TBC). A total number of 5248 equids (Table 1, gives a description of sex, age group and work type of the observed equids) were assessed during the period from February 2012 to January 2014. All of the assessments were done by the first author using the previously mentioned checklist showed in Table 4. The assessments were carried out in the actual working sites for the equids that were assessed, including streets, markets and brick kilns. Assessments were conducted after gaining the owner’s permission and with a minimal interruption of the equid’s work (only a minimum rest period could be used). Equids were individually identified within working sites by asking the owner to identify the animal and by individual body markings, which included noting the base color of the equid, any natural markings (e.g., a star or blaze or sock or stocking) and any acquired marking (e.g., old scars, brands or decorative markings). 2.3. Animal based welfare indices (ABI) In addition to a 6 month period of preliminary work at sites relying upon working equids, a range of descriptors and parameters were selected based on work in published literature, and after with consultation with working equid experts. This preliminary work helped to identify the optimum selection and sequence of observations to address the study objectives, suit the nature of the equids’ working style and the owners’ culture, and to reach optimum speed and accuracy of recording, as they had to be done without interrupting the animals’ work (only a minimal rest period could be used). During this period the first author also thoroughly studied and investigated the 102 page photographic guide for working equine welfare assessment of Pritchard (Pritchard and Whay, 2004) and Whay (Whay et al., 2003). Measures were carefully selected and modified for collection in the current study. Practical field training for the measures to be used in the study, and verification of the observer reliability, was carried out with an experienced Brooke Egypt veterinarian (second author), and a highly experienced horse behaviorist (third author). Additionally, more than 25% of the acquired data were fully verified, post hoc, for observer reliability with a working equine expert (fifth author). Six descriptors for each equid were recorded: date, geographical region and the governorate, equid species (Horses (H), Mules/Hinnies (M) or Donkeys (D)), and equid’s sex (Male, intact (M), Gelding (G) or Female (F)). The equid’s work type was classified as described in Table 2. The age of the selected equid was determined by history from the owner and confirmed by denti-

Table 2 Components of different work types. Work type Transport goods by cart (TGC) Transport bricks by cart (TBC) Transport people by cart (TPC) Ridden (R)

Agricultural goods being taken to markets, furniture and manufactured goods in rural areas. Transport of bricks in brick kilns, from excavation sites to chimneys and then to stores. Transporting people, mostly tourists or natives in tourist sites. Riding for tourists or natives in tourist sites.

tion (Crane, 1997), then grouped into one of three age categories (Pritchard et al., 2006): Young (Y) = five years of age and younger, Medium (MA) = more than five, and up to and including fifteen years, and Old (O) = over fifteen years. Six measurement clusters were identified, which incorporated 27 parameters that were chosen to represent all critical aspects of working equid welfare. These included behavior measures which assessed the equid’s attitude and response to human approach and handling, to investigate the nature of that equid’s relationship to humans (either as friendly, avoiding or even aggressive). Health parameters focused heavily on the importance of distinguishing healthy, diseased and/or heat stressed equids (of particular importance in hot and arid climate zones like Egypt). Body lesions were recorded and categorized according to their severity and the most likely initial cause (e.g., caused by ill-fitting harness). Body condition score (BCS) was assessed and categorized to be indicative of the nutritive status of the working equids. Coat health, presence of ectoparasites (mainly ticks) and grooming occurrence were also assessed, as well as hoof health, including wall and sole shape and horn quality. 2.4. Measuring of the selected animal-based indicators (ABI) A specific sequence for measuring the parameters was developed, to optimize efficiency while maintaining accuracy, and to ensure that that the order did not influence the outcomes. The following sequence was followed: 1) The observer stood a distance of 3 m away from the equid at an angle of 45◦ from the sagittal plane of the animal’s body and maintained this position for 10 s without disturbing the animal, to observe the animal’s attitude. This was followed by counting of the respiratory rate, and recording any change of the animal’s attitude for a further 60 s. The animal’s state of alertness

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

2)

3)

4)

5)

6)

7)

8)

9)

10)

and respiratory rate were assessed after stopping their work to minimize uncertainty between depression and relaxation, as well as to prevent any increase in respiratory rate due to fear of observer approach and/or manipulation. At the same distance, but at an angle of 20◦ from the sagittal plane of the animal’s head, the observer approached the animal with slow and calm steps, stopping at about 30 cm from the head. The animal’s response to the observer was recorded, while any lesions on the animal’s nasal bridge and mouth were recorded. From the same position, the observer slowly raised his open hand toward the animal’s chin and touched it, then recorded the animal’s response. Following this the observer held the animal’s head and counted its pulse from the external maxillary artery. The observer assessed the color of the oral mucus membranes according to (Mekuria et al., 2013), then, applied the capillary refill time test according to the methods of (Popescu et al., 2014). The equids age was simultaneously determined (Crane, 1997). The observer recorded any neck lesions and used the tips of his fingers against the direction of the equids hair to detect if the equid was recently groomed. This was followed by applying the skin tent test following the method of Pritchard et al. (2006). Lesions at the anterior aspect of the knee, firing lesions and tethering lesions along the inner aspects of the thighs were recorded, followed by gentle manual compression to detect pain and/or swelling of the forelimb tendons and fetlock joints. The same steps were followed for the hocks and hind tendons, and fetlock joints. Manual compression was performed to detect swelling of tendons and/or fetlock joints rather than visual observation, since swelling was not always visible, as in the case of donkeys with hairy legs. The animal’s hooves were observed on the ground from both front and side view, and any abnormalities of either the shape or the horn of the hoof wall on any foot recorded. The animal’s right foreleg was then lifted to observe the surface of the sole of the foot. A hoof-pick was used to clean the sole if necessary. The presence of any abnormality of the shape or structure of the sole, bars or frog was recorded (Upjohn et al., 2013). The equid’s BCS was recorded, as per Pearson and Ouassat (2000), firstly from the rear view while recording lesions at the hind quarters, withers and spine, followed by two side views while also recording lesions of the chest area/rib cage area. In addition to the lesions listed above, any other lesions were also recorded, and any area of matted, scabby, scurfy skin throughout the equid’s body, as per Upjohn et al. (2013). Presence of ectoparasites (ticks) was also recorded (Pritchard et al., 2005). Finally the observer measured the animal’s rectal temperature using a digital thermometer.

2.5. Interpreting the selected (Animal based indicators) ABIs 2.5.1. Behavioral parameters The following methods, developed by Burn et al. (2010b), were used to assess the animal’s attitude, response to observer approach, and chin contact. 2.5.1.1. Animal’s attitude (AA). Equids were considered as being ‘Alert’ (A) when they had an active interest in their surroundings, with active movement of the ears toward an existing stimulus. Eyes were usually wide open and head up unless sniffing or eating. Animals also might exhibit sniffing, vocalizing, head movement, tail swishing and moving feet (Pritchard et al., 2006).

55

Equids were considered to be ‘Depressed’ (D) if they showed a passive interest in surroundings. The following postures indicated this mental state • Ears could be back, or lowered, with minimal or no movement in response to stimuli. • Eyes could be open, half or fully closed. • Head could be up, level with withers, or lowered, but rose in response to surrounding activity. • The animal shows a minimal or absent response to self-need, e.g., flapping ears, moving feet or swishing tail to remove flies (Pritchard et al., 2006). Severe depression was excluded in this study as it was most likely to occur in animals suffering from serious diseases, with a deteriorated health state. Such animals were usually kept stabled for vet care with no or minimum work and thus were excluded from the study. 2.5.1.2. Animal’s response to observer approach (ROA). The following responses to observer approach were recorded • Friendly response (F): movement of the head toward the observer with relaxed face and the eyes opened but not overly wide, forward turning of the ears (Waring, 2003). • Ignoring observer approach (IOA): immovable without attempts to approach or avoid observer with depressed or relaxed body position and facial expressions. • Avoidance (AV): the animal was either immovable, retaining a tensed body position including upward holding of the head with tensed muscles and facial expression, the latter showing widely opened eyes and tightly held lips, or the animal tried to move away from the approaching observer (Popescu and Diugan, 2013). • Aggressive response (AG): the animal attempts to bite. This can include behaviors such as, the animal’s ears are laid back against the skull, eyes are fully opened and oriented toward observer, nostrils are dilated with or without wrinkles along the upper posterior edge, and mouth may be open, exposing teeth. It could also include attempts to kick, including pawing or stomping of one or more legs on the ground. (Waring, 2003). 2.1.5.3. Animal’s response to chin contact (CC). The animal either accepted (Acc) chin contact, or tried to avoid chin contact (Avc). 2.5.2. General health parameters Respiratory rate (RR), pulse rate (PR) and rectal temperature (RT), were classified as belonging to one of three categories: normal, moderately abnormal or severely abnormal, as per Lemma and Moges (2009) and Pritchard et al. (2006). The limits of each category are shown in Table 3. The skin tent test (STT) was considered to be normal when the skin returned to normal immediately after it was released, and abnormal when skin stayed elevated for more than 1/2 s. Capillary refill time (CRT) was considered to be normal when the return of normal pink color of oral mucous membranes occurred in less than 2 s in the bleached area which was produced by finger pressure, just above the incisor teeth. 2.5.2.1. Aggregated health parameter scores. Animals were then classified according to their aggregated health parameters into one of three categories which relate to possible heat stress (Ali et al., 2015): Average aggregated healthy score (AAHS): This was assigned to animals showing normal RR, PR, RT, MM, STT and CRT. Aggregated heat stress score (AAHSS): This was assigned to animals showing moderately abnormal RR, PR, RT and MM, and abnormal STT and CRT.

56

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

Table 3 The normal, moderately abnormal and severely abnormal values for health parameters including respiratory rate, pulse rate, rectal temperature and the color of oral mucous membranes.

Normal (N) Moderately abnormal (MA) Severely abnormal (SA)

Respiratory rate (RR) (breath/min)

Pulse rate (PR) (beat/min)

Rectal temperature (RT)

Mucous membrane (MM)

(14–26) Up to 80 (>80)

(38–48) Up to 60 (>60)

(36–38.3 ◦ C) Up to 39.5 (>39. 5◦ C)

Pink Moderately congested Congested and/or with petechial hemorrhage spots

Aggregated diseased score (AADS): This was assigned to animals showing abnormal RR, PR, RT, and MM. The assessment included only equids that appeared healthy, and were eating and working properly. Thus equids with AAHSS were affected by moderate rather than severe heat stress, as they were able to continue working. 2.5.3. Coat health parameters Coat health (CH) was examined and considered to be unhealthy or abnormal (Ab) when the hair was matted, scabby, and scruffy, or when alopecia was detected due to mange and/or ring worm. Equids were examined for presence of ectoparasites (EP) and considered to be free (F) when no ectoparasites (primarily ticks) could be detected during examination. Equids were examined for evidence of grooming (G) and considered to be recently groomed (RG) if no dirt were found on the tips of the observer’s fingers. 2.5.4. Body lesions Lesions were initially classified according to surface area. The affected area had to be at least 2 × 2 cm2 (quadratic lesion), 1 × 4 cm2 (rectangular lesion) or 2.3 cm in diameter (circular lesion) in order to be recorded as a lesion. The entire region being scored was assessed using the following scoring system: No lesion (NL): the region of the animal being assessed had no lesions, or only had lesions that were smaller than the qualifying area. Superficial or healed (S/H): the lesion included a qualifying area of hairless skin. The skin was unbroken, partially broken (pale pink) or scabbed. Scars might be healed with hairless skin or might be covered in white hairs once healed. The score was given for the most severe part of the lesion. Skin and immediate subcutaneous layers broken (Sc): the lesion included a qualifying area of broken skin with visible pink or red tissue present. Granulation tissue could also be present. Deep lesion (DL): the qualifying area included a visible muscle, tendon or bone. Lesions on the commissures of the lips or nasal bridge, and lesions on the front of the knee (carpus) or point of the hock were assessed for either being present (Pr) or absent (Ab) according to Dennison et al. (2007). As true lesion might include scars of old wounds, broken skin and/or subcutaneous tissue, visible muscle and/or bone or tendons. Finally, where possible, lesions were categorized according to their most likely initial cause, based on work by Popescu and Diugan (2013). Some lesions were either omitted or assigned to the following extra categories: Harness-induced body lesions (HIL): This included neck/point of shoulder lesions (LNS) due to poor quality collars, lesions at the sides of chest (LC) due to faulty fitted cart shafts, and lesions at the withers and spine (LW) due to using the pack saddle without proper padding. Mistreatment-induced body lesions (MIL): This included lesions at the nasal bridge due to tightly fitted cloth or metal nose bands, lesions at commissures of the lips (LNB/LC) due to a tightly or poorly fitted bit, and lesions at the hind quarters (LHq) due (probably) to frequent and/or aggressive beating of the equids Lesions at Points

of Hocks (LPH): These were categorized as due to the hocks hitting the cart when kicking back as a response to being beaten. Overwork-induced body lesions (OIL): These were lesions where there was an indication of pain and/or swelling of the forelimbs and/or hind limb tendons and/or fetlock joints (PSTF). This also included lesions at the anterior aspect of the knee (Lkn), which were classified as due to falling down while pulling overly heavy loads up and down steep terrain. Lesions due to firing (FL): Firing is the practice of burning the skin with hot irons. In traditional medicine it is thought to enhance healing. Lesions from firing were recorded whether they were fresh or healed, or scars on any body part which formed linear, geometric or dot patterns. The most common locations were the fore limbs and hind limbs over joints and tendons, but they could be found anywhere on the body. We also noted whether firing was used for decoration or identification. Lesions due to tethering (TL): These lesions or scars appeared at any point along the length of the limb and were always horizontal. Other body lesions (OBL): These include other lesions that were recorded in any area other than previously mentioned. They were classified as mostly accidental, animal-induced lesions or parasiteinduced (e.g. habronemiasis). 2.5.5. Body condition score (BCS) Body condition score was initially assigned based upon a fivepoint scale, from 1 (poor) to 5 (obese). Following this, categories were compiled and recorded as follows “Thin animals” (scores 1–2), “Moderate BCS animals” (≥2.5–≤3.5) and “Obese animals” (scores 4–5). Donkeys with BCS >2.5 tend to have deceptively large bellies and store fat into uneven fat pads, which can be in the neck, hind quarters or even the belly. Thus it is recommended to use other descriptors to thoroughly assess BCS in these cases. 2.5.6. Hoof observation Wall shape (WS): the front view of a normal hoof should show two symmetrical sides of equal length. The coronary band should be parallel to the ground. In horses the medial wall might be slightly more upright than the lateral one, whereas with donkeys both sides are straight upright. When viewed from the side, the front of both wall and pastern should run parallel with the hoof pastern axis (An imaginary line drawn from the center of the fetlock joint along the axis of the limb to the ground). Finally the wall length should be 2–3 times longer at the toe than at the heel (Pritchard and Whay, 2004). Horn quality (HQ): Normal hoof horn quality was considered when no broken hoof wall, cracks, ridges, or rings were observed. Sole shape (SS): the sole was considered normal when it appeared round in horses or approximately keyhole-shaped in donkeys. Furthermore, the sole should have two symmetrical halves and have a large, rubbery, intact frog, distinct bars and a slightly cupped sole. 2.6. Welfare assessment score checklist The checklist was categorized into the following subcategories: behavior score, general health score, body lesions and coat scores,

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

body condition score and hoof score. Each category was given an assigned weight, which was then subdivided either equally or unequally based on the significance of the subcategories (Table 4). The weighting and the subdivisions for each category were defined depending on the significance of the category with regard to its perceived influence on the welfare status of working equids. These assignments were based on the earlier input from previous literature, the working equids expert consultations, and the preexperimental preliminary work. Each measure was given a precise score depending on whether the output was normal or abnormal, or depending on the degree of severity. The welfare scoring checklist was then field tested during the pre-experimental preliminary work. This was carried out at various sites, with different types of working equids, and among equids with estimated differences in welfare status, to determine its applicability, efficacy and suitability in a fast paced scoring environment. Moreover, the accessibility and practicability of assessing each individual parameter was tested under these conditions. The scoring of a single equid took approximately 10–13 min. For calculation of partial (e.g., for behavior only) or total welfare score either for a single or a little number of working equids. Scores for each parameter should be given as explained in the welfare scoring checklist (Table 4), and then by adding the corresponding values to these scores in the formulas as shown in (Table 5), both partial scores for behavior, health, body lesions, BCS, body coat, hoof and total welfare score can be obtained. While for a large number of working equids, a predesigned data sheet and SAS code for calculation of partial and total welfare scores are available upon request from author. 2.7. Data handling Data handling and statistical analysis were carried out using SAS 9.2 (SAS 9.2, SAS Institute Inc., Cary, NC, USA). Cut off values for classifying either partial (e.g., behavior, health etc.) or total welfare scores as Poor (Pr), Inadequate (IAd) and Adequate (Ad) were based primarily on qualitative assessment of the animals. The upper and lower quartile values for the measures of behavior, health, BCS, body lesions, hoof and coat observations were also used in determining cut off values. As the data set contains 5248 randomly selected working equids assessed throughout the study, representative of the various work types, age groups, gender groups and species of working equids in Egypt, these statistically calculated quartile values can be used alongside the previously mentioned qualitative measures in determining the cut off values. The cut off values for different categories of partial and total welfare scores are shown in Table 6. The cut off limits were qualitatively assigned through pictures and records of the equids. For example, limits for BS were designated as follow: all equids with a depressed attitude and/or an aggressive response would never fall within the “Adequate” category, and those with an alert attitude and/or a friendly response would never fall within either “Poor” or “Inadequate” categories for BS. With regard to GHS, the cut off values were designed such that equids with five or more parameters scoring normal would fall within the “Adequate” category, while all equids with not more than three categories scoring as normal would fall within the “Inadequate” category. Finally equids with more than three parameters scored either moderately or severely abnormal would fall within the “Poor” category. For BCS, only equids with moderate BCS (≥2.5–≤3.5) would be ranked within the “Adequate” category. Data were recorded by hand and each single welfare scoring checklist included 10 equids. Due to the large sample size, the data were entered into the spread sheet as an average for every 10 animals of the same species, work type, age group and the gender. Though this has reduced the degrees of freedom, the method has

57

Table 4 The welfare scoring checklist showing the corresponding score for each parameter. Date I- Behavior score (BS) 1- Animal’s attitude (AA) 2- Response to observer approach (ROA)

3- Chin Contact (CC)

II- General health score (GHS) 1- Pulse rate (PR) 2- Respiratory rate (RR) 3- Rectal temperature (RT) 4- Mucous membrane (MM) 5- Skin tent tests (STT) 6- Capillary refill time CRT III- Body coat score (BCoS) 1- Coat Health (CH) 2- Ectoparasites (EP) 3- Grooming (G) IV- Body lesions score (BLS) 1- Harness Induced body Lesions (HIL)a A- Lesions at Neck and point of Shoulders (LNS)a B- Lesions at sides of Chest (LC)a C- Lesions at Withers and spine (LW)a 2- Mistreatment induced body Lesions (MIL)b A- Lesions at points of hocks (LPH)b B- Lesions at nasal bridge and/or commissures of the lips (LNB/LC)b C- Lesions at hind quarters (LHq)b 3- Overwork induced body lesions (OIL)c A- Pain and/or Swelling of the forelimbs and/or hind limbs Tendons and/or Fetlock joint (PSTF)c B- Lesions at anterior aspect of the knee (LKn)c 4- Other body lesions (OBL) 5- Firing lesions (FL) 6-Tethering lesions (TL) V- Body condition score (BCS)

VI- Hoof Score (HS) 1- Wall shape (WS) 2- Horn quality (HQ) 3- Sole shape (SS) VII- Other observations

Alert (A) = 1, Depressed (D) = 0 Ignore Observer Approach (IOA) give score (0.5), Aggressive (Ag) give score (0), Avoid observer (Av) give score (0.25) and if Friendly response (F) give score (1) (Acc = Accept chin contact) give score (1) and if (Avc = Avoid chin contact) give score (0) (N = Normal) give score (1), (MA = Moderately abnormal) give score (0.25) and if (SA = Severely abnormal) give score (0) (N = Normal) give score (1) and if (Ab = abnormal) give score (0) (N) give score (1) and if (Ab) give score (0) Free (F) give score (1), Present (Pr) give score (0) Recently Groomed (RG) give score (1), Un-Groomed (UG) give score (0)

No Lesions (NL) give score (1), Superficial or Healed lesions (S/H) give score (0.5), broken Subcutaneous or Deep Lesions (Sc or DL) give score (0).

Absent (Ab) give score (1), Present (Pr) give score (0) (NL) give score (1), (S/H) give score (0.5) and if (Sc or DL) give score (0).

(Ab) give score (1) and if (Pr) give score (0)

(NL) give score (1), (S/H) give score (0.5) and if (Sc or DL) give score (0). (BCS ≥ 2.5–≤ 3.5) give score (1), (BCS = 4–5) give score (0.75) and if (BCS = 1–2) give score (0.25) (N = Normal) give score (1) and if (Ab = Abnormal) give score (0)

a Harness induced body lesions either due to poor quality collars, faulty fitted cart shafts or due to using the pack saddle without proper padding. b Mistreatment induced body lesions due to tightly fitted cloth or metal nose bands, tightly fitted bit or due to continuous and aggressive beating of the equids and lesions at points of hocks due to hitting the cart when kicking back as a response for being beaten. c Overwork induced body lesions include pain and/or swelling of the forelimb and/or hind limb tendons and/or fetlock joints, and also lesions at anterior aspects of the knee due to falling down while pulling overly heavy loads up and down steep terrain.

58

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

Table 5 Formulas for calculating partial and total welfare scores. Welfare scorea

Calculations (Scorec × statistical parameter weight)

Behavior Score (BS) General health score (GHS) Body condition score (BCS) Body coat score (BCoS) Body lesions score (BLS)

=((AA × 6) + (ROA × 8) + (CC × 4))

Hoof score (HS) Total welfare score (TWS)b

due to either small sample size (TBC gelding donkeys, n = 6 and TBC female donkeys, n = 4) or due to never having been observed during assessment (TGC and TBC gelding mules). 3. Results

=((PR × 5) + (RR × 5) + (RT × 5) + (MM × 3) + (STT × 1) + (CRT × 3))

3.1. The influence of work type on behavior =(BCS × 20)

=((CH × 3) + (EP × 2) + (G × 1))

=((HIL (LNS × 2) + (LC × 2) + (LW × 3)) + (MIL (LNB/LC × 2) + (LHq ×3) + (LPH ×1)) + (OIL (PSTF × 5) + (LKn × 2)) + (FL × 2) + (TL × 2) + (OBL × 1)) =((WS × 4) + (HQ × 3) + (SS × 2)) =(BS + GHS + BCS + BCoS + BLS + HS)

a Partial welfare score calculations were made by multiplying the precise score, which was given for each assessed parameter depending on either its presence as normal or abnormal or depending on the degree of severity as mentioned in checklist by the statistical weights, which were determined based on the data collected from previous literature, working equids experts’ consultation and the preliminary work period. b The Total Welfare Score can be calculated by summing the partial scores for different categories. c AA: Animal’s Attitude, ROA: Response to Observer Approach, CC: Chin Contact, PR: Pulse Rate, RR: Respiratory Rate, RT: Rectal Temperature, MM: Mucous Membrane, STT: Skin Tent Test, CRT: Capillary Refill Time, CH: Coat Health, EP: Ectoparasites, G: Grooming, HIL: Harness Induced body Lesions (LNS: Lesions at Neck and point of Shoulders, LC: Lesions at sides of Chest: Lesions at Withers and spine), MIL: Mistreatment Induced body Lesions (LNB/LC: Lesions at Nasal Bridge and/or Commissures of the Lips, LHq: Lesions at Hind quarters, LPH: Lesions at Points of Hocks), OIL: Overwork Induced body Lesions (PSTF: Pain and/or Swelling of the forelimbs and/or hind limbs Tendons and/or Fetlock joint, LKn: Lesions at anterior aspect of the Knee, FL: Firing Lesions, TL: Tethering Lesions, OBL: Other Body Lesions), WS: Wall Shape, HQ: Horn Quality, SS: Sole Shape.

Table 6 The cut off values for different categories of partial and total welfare scores. Cut off Scoresa

Poor (Pr)

Inadequate (IAd)

Adequate (Ad)

Behavior Score (BS) General Health Score (GHS) Body Condition Score (BCS) Body Coat Score (BCoS) Body Lesions Score (BLS) Hoof Score (HS) Total Welfare Score (TWS)

0 to ≤10 0 to ≤10 0 to ≤5 0 to ≤3 0 to ≤17 0 to ≤5 0 to ≤51

>10 to <14 >10 to <18 >5 to ≤15 >3 to <5 >17 to ≤23 >5 to <7 >51 to <82

≥14 to ≤18 ≥18 to ≤22 >15 to ≤20 ≥5 to ≤ 6 >23 to ≤25 ≥7 to ≤ 9 ≥82 to ≤100

a Cut off scores were assigned for describing both the total welfare status and its subcategories. A score of “Poor” was given to the lowest levels of welfare status, which were considered at the highest risk and needing maximum attention and the fastest interventions. A score of “Inadequate” was given to a deteriorated status, which ordinarily represents the majority of working equids and also needs proper intervention and strategic planning to reduce equids’ suffering within this category. Finally a score of “Adequate” was given to the equids that were assessed as having at least a fair or satisfactory welfare status.

been approved by CANR Statistical Consulting Center at Michigan State University. Descriptive statistics were used to obtain means of all parameters, and analysis of variance “ANOVA” was used to compare differences between groups, with the statistical significance level set at p ≤ 0.05. Nine work type groups of equids were eliminated either due to small sample size (Ceremonial horses, n = 7, Pack donkeys, n = 11, TBC horses, n = 7, R donkeys, n = 8, TPC donkeys, n = 6) or due to never having been observed during assessment (R and TPC mules, pack horses or mules). Four sex groups were eliminated

The prevalence of each of the behavior measures investigated (AA, ROA and CC) within each work type group and species can be seen in Table 7. The TGC horses showed the highest prevalence of depressed attitude and unresponsiveness toward observer approach, both when compared with equids of the same species (depressed attitude: F2,2195 ≥ 598.25; p ≤ 0.05, unresponsiveness: F2,2195 ≥ 369.58; p ≤ 0.05, in both cases versus TPC and R horses respectively), and of the same work type group, but a different species (depressed attitude: F2,2799 ≥ 652.36; p ≤ 0.05, unresponsiveness: F2,2799 ≥ 782.25; p ≤ 0.05, in both cases versus TGC donkeys and mules respectively). The TGC mules showed the highest prevalence of avoiding both observer approach and chin contact (avoid observer approach: F2,2799 ≥ 698.02; p ≤ 0.05, avoid chin contact: F2,2799 ≥ 589.89; p ≤ 0.05, in both cases versus TGC horses and donkeys respectively), and also had the highest prevalence of an aggressive response towards the observer (F2,2799 ≥ 452.25; p ≤ 0.05, versus TGC horses and donkeys respectively). In contrast, the TPC and R horses had the highest prevalence of alert attitude, friendly response and accepting chin contact (TPC: F2,2195 ≥ 250.96; p ≤ 0.05, R: F2,2195 ≥ 352.69; p ≤ 0.05, in all cases versus TGC horses respectively). 3.2. The influence of work type on the health, body lesions, BCS, coat and hoof observations The TPC and R horses had the highest prevalence of healthy animals, moderate BCS and obese BCS (TPC: F2,2195 ≥ 485.69; p ≤ 0.05, R: F2,2195 ≥ 520.69; p ≤ 0.05, in all cases versus TGC horses respectively, Table 8). The TGC donkeys had a higher prevalence of harness induced lesions, other body lesions, firing lesions, an unhealthy coat, and presence of ectoparasites (F2,2799 ≥ 125.96; p ≤ 0.05, versus TGC horses, F2,2799 ≥ 185.69; p ≤ 0.05, versus TGC mules respectively). TGC mules had the highest prevalence of mistreatment lesions, tethering lesions and lack of grooming (F2,2799 ≥ 325.52; p ≤ 0.05, versus TGC horses, F2,2799 ≥ 385.69; p ≤ 0.05, versus TGC donkeys respectively). Finally, the animals were most affected by heat stress, diseases, overwork induced body lesions (swelling of tendons and or fetlock joints and lesions on the anterior aspects of knee), abnormal hoof shape and horn quality, and having the highest prevalence of thin equids, were TGC horses (F2,2799 ≥ 250.23; p ≤ 0.05, versus TGC donkeys, F2,2799 ≥ 333.85; p ≤ 0.05, versus TGC mules respectively). 3.3. Total and partial welfare scores of 5210 working horses, donkeys and mules according to their type of work The total scores and subcategory scores using our numerical assignments are shown in Table 9. With regard to total welfare scores, brick kiln donkeys are at the greatest risk for poor welfare status (F6,5241 ≥ 236.58; p ≤ 0.05, versus TGC, TPC and R horses, TGC donkeys, TGC and TBC mules respectively). Riding horses and TPC horses had generally adequate welfare status, at least in the populations examined in this study. From a behavior sub-score standpoint, both brick kiln donkeys and brick kiln mules were at considerable risk with regard to their welfare status. In terms of body lesions, it appears that both donkeys and mules are at higher risk than horses.

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

59

Table 7 The prevalence of behavior observations for different work types within each species of working equids. Observations of behavior (% of equids)

% Within each work type for different species Horse (n = 2198)

Donkeys (n = 2640)

Mules (n = 410)

TGC (n = 1142)

TPC (n = 570)

R(n = 486)

TGC(n = 1500)

TBC(n = 1140)

TGC(n = 160)

TBC(n = 250)

1-Equid attitude Alert Depressed

79.33Aa 20.67Aa

90.17B 9.83B

92.07B 7.93B

80.98Aa 19.02Aa

55.99Ba 44.01Ba

87.79Ab 12.21Ab

72.33Bb 27.67Bb

2- Response to observer approach Friendly response Ignore approach Avoid approach Aggressive response

42.96Aa 22.56Aa 25.77Aa 8.71Aa

62.19B 12.95B 20.63A 4.23B

68.37B 9.79B 19.85A 1.99B

46.42Aa 20.14Aa 27.22Aa 6.22Aa

30.66Ba 49.96Ba 16.37Ba 3.01Ba

35.2Ab 19.82Aa 30.76Aa 14.22Ab

13.55Bb 27.34Bb 39.5Bb 19.61Bb

3- Chin contact Accept contact Avoid contact

69.29Aa 30.71Aa

83.04B 16.96B

88.98B 11.02B

67.44Aa 32.56Aa

53.5Ba 46.5Ba

57.34Ab 42.66Ab

44.16Bb 55.84Bb

All parameters are presented as mean percentage of equids demonstrating behavior. ABC different superscripts within the same row indicate statistically significant difference at (P < 0.05) between different work types within the same species of equids. a,b,c different superscripts within the same row indicate statistically significant difference at (P < 0.05) between different species of the same work type except for R and TPC in horses, as the two groups were excluded in donkeys and mules due to either small sample size or never having been observed.

Table 8 The influence work type health, body lesions, BCS, coat and hoof observation of working equids expressed as proportions within each type of work for each species. Observations of parameters (% of equids)

% Eithin each work type for different species Horse

Donkeys

Mules

TGC

TPC

R

TGC

TBC

TGC

TBC

1- General health parameters AAHS1 AAHSS2 AADS3

56.86Aa 39.53Aa 3.61Aa

88.63B 10.23B 1.14B

95.12C 4.23C 0.65B

72.25Ab 24.76Ab 2.99Aa

40.15Ba 48.33Ba 11.52Ba

73.43Ab 23.96Ab 2.61Aa

68.56Ab 26.86Ab 4.58Ab

2- Body lesions Harness induced lesions Mistreatment induced lesions Overwork induced lesions Other body lesions Firing lesions Tethering lesions

12.61Aa 19.86Aa 22.67Aa 3.65Aa 13.71Aa 3.15Aa

8.02B 12.42B 9.82B 1.25B 8.71B 3.13A

2.02C 4.36C 3.12C 0 3.03C 2.00A

21.61Ab 20.63Aa 20.71Aa 10.63Ab 21.86Ab 4.27Aa

47.52Ba 25.16Aa 32.91Ba 6.12Aa 27.78Ba 6.13Aa

17.76Ac 36.71Ab 13.03Ab 5.21Ac 17.51Ac 11.90Ab

24.41Bb 41.72Bb 19.08Ba 4.96Aa 24.97Ba 13.57Ab

3- BCS Thin animals Moderate BCS animals Obese animals

27.63Aa 71.34Aa 1.03A

5.63B 92.52B 1.85A

4.02B 92.03B 3.95A

23.96Aa 76.04Aa 0

53.50Ba 49.5Ba 0

17.89Ab 82.11Ab 0

32.62Bb 67.38Bb 0

4- Coat observations Coat health (unhealthy) Ectoparasites presence Grooming (ungroom)

26.87Aa 32.40Aa 29.85Aa

17.12B 16.74B 8.91B

10.70C 11.95B 4.86B

51.20Ab 48.92Ab 36.52Aa

62.43Ba 54.96Ba 58.66B

28.60Aa 41.23Aa 41.48Ab

46.88Bb 39.13Ab 54.23B

5- Hoof observations Hoof shape (abnormal) Hoof horn quality (abnormal) Sole shape (abnormal)

44.77Aa 34.79Aa 17.88Aa

24.15B 23.56B 30.02B

12.54C 22.95B 2.76C

29.54Ab 24.08Ab 23.64Ab

38.18B 35.74B 30.52Ba

8.06Ac 13.23Ac 12.22Ac

17.17B 21.73B 17.35Ab

All parameters are presented as mean percentage of equids demonstrating different parameters. The average aggregated health score: for animals showed normal (RR, PR, RT, MM, ST, and CRT). The average aggregated heat stress score: for animals showed Moderate abnormal (RR, PR, RT, MM) and abnormal (STT, CRT). 3 The average aggregated diseased score: for animals showed abnormal (RR, PR, RT, and MM). A,B,C superscripts within the same row indicate a statistical significant difference at (P < 0.05) between different work types within the same species of equids. a,b,c superscripts within the same row indicate a statistical significant difference at (P < 0.05) between different species of the same work type except for R and TPC in horses, as the two groups were excluded in donkeys and mules due to either small sample size or never been observed. 1 2

4. Discussion When considering whether to begin welfare enhancement activity in a certain location, set goals with regard to prioritization of desired improvements, or whether to end health support activities in a certain location, many decisions need to be made by NGOs, such as working equid welfare charities. These decisions should be based primarily on the welfare status of working equids in the region in question. Thus a highly applicable, reliable welfare scoring system is crucial for this process. Hence, our objective for this

study was to develop a comprehensive welfare assessment score for accurate, multifactorial and reliable welfare assessments of working equids. This can be used by NGOs, practitioners, veterinarians and researchers to evaluate the welfare status of working equids and monitor the progress of their welfare status over time. It is important that such welfare scoring protocol could identify which equids, in terms of species, age, sex and type of work, demonstrate the most at risk welfare status, as well as being able to identify which aspects of welfare are impaired (e.g., behavior, BCS, body lesions etc.).

60

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

Table 9 Scores for partial and total welfare and their corresponding cut off categories for different work types within each species of working equids.

TGC horses TPC horses R horses TGC donkeys TBC donkeys TGC mules TBC mules

Total behavior score

Total health score

Total BCS

Total body lesions score

Total body coat score

Total hoof score

Total welfare score

12.3 (IAd)a 14.5 (Ad) b 15.5 (Ad)b 12.6 (IAd)a 9.7 (Pr) c 12 (IAd)a 8.6 (Pr)c

14.3 (IAd)a 20.2 (Ad)b 22 (Ad)b 16.2 (IAd)a 10.1 (IAd)c 16.2 (IAd)a 14.5 (IAd)a

14.9 (IAd)a 16.5 (Ad)b 18.3 (Ad)b 15.6 (Ad)a 10.3 (IAd)c 16.2 (Ad)b 12.4 (IAd)c

19.7 (IAd)a 21.2 (Ad)b 23.7 (Ad)b 16.1 (Pr)c 13.7 (Pr)c 16 (Pr)c 14.9 (Pr)c

3.4 (IAd)a 4.6 (IAd)b 5.1 (Ad)b 2.9 (Pr)a 1.6 (Pr)c 3.1 (IAd)a 2.1 (Pr)c

5.8 (IAd)a 6.5 (IAd)b 7.2 (Ad)b 6.4 (IAd)b 5.2 (IAd)a 8.7 (Ad)b 5.9 (IAd)a

70.4 (IAd)a 83.5 (Ad)b 91.8 (Ad)b 69.3 (IAd)a 50.6 (Pr)c 72.2 (IAd)b 58.4 (IAd)c

All parameters are presented as the given scores and the corresponding cut off categories. A score of “Poor” (Pr) was given to the lowest levels of welfare status, which were considered at the highest risk and needing maximum attention and the fastest and most critical interventions. A score of “Inadequate” (IAd) was given to a deteriorated status, which ordinarily represented the majority of working equids in this study and also in need of proper intervention and strategic planning to reduce equids’ suffering within this category. Finally a score of “Adequate” (Ad) was given to the equids that were assessed as having at least a fair or satisfactory welfare status. abc different superscripts within the same column indicate statistically significant difference at (P < 0.05) between different groups of working equids.

The protocol was developed to address the welfare indicators mentioned by Sørensen et al. (2001), and to reliably describe aspects of welfare that may matter from the point of view of the animals involved. It was also important that it could describe changes in welfare status over time, be easily influenced by decisions taken by the equids’ handlers or owners, and finally, be useable in a relatively cheap, easy and practical manner. It should be noted that results for brick kiln donkeys and mules were published previously and discussed more thoroughly (Ali et al., 2015). This was to investigate whether donkeys or mules fare better within the harsh environmental and working conditions of the brick kilns. The brick kiln data is used here only for comparative purposes. Behavioral tests are considered a major component in welfare assessment of working equids. These tests indicate the responsiveness of the equid to the surrounding environment, and help identify fearfulness or aggression toward humans. They can help to demonstrate the nature of the human-equid interaction, and its implication on the equid’s psychological status. The higher prevalence of a depressed attitude, and the unresponsiveness toward the observer approach test in TGC horses compared with TBC and R horses and TGC donkeys and mules could be attributed to the higher prevalence of low BCS, heat stress and the numerous lesions caused by overwork in these animals. This is in agreement with the findings of Popescu and Diugan (2013). These findings are also consistent with those of Pritchard et al. (2005), who added that working equids may be unresponsive due to disease, exhaustion, over stimulation by a crowded and noisy city environment, or to avoid soliciting harsh handling. It could also be the case that these behavioral findings could be related to the high thermal load experienced by TGC horses; they perform one of the hardest jobs of working horses. Alternatively, it may be that they are not as able to withstand high temperatures as working donkeys and mules performing the same type of work. In fact Singh et al. (2005) concluded that donkeys are sturdy working animals, and can be considered better than other draught animals because of their inherent tolerance for dehydration, low sweating rate and good thermoregulatory mechanisms. However, that study did not include mules. The current findings are also consistent with those reported by Burn et al. (2010a) who found that equids with frequent and severe physical problems usually demonstrate depressed attitude and unresponsiveness. This implies that the equids’ resources are being stretched beyond their limits, and their fitness is compromised. It could be the case that the equids are conserving their energy by not responding to potentially threating stimuli, such as flies or approach by a human. This behavior could be indicative of several negative welfare states, such as exhaustion due to overwork (Tadich et al., 1997; Pritchard et al., 2009), chronic pain (Ashley et al., 2005), and general malaise (Millman, 2007). Much

of the literature has linked these states with sickness behavior, which can be described as a ubiquitous, non-specific proinflammatory cytokine response to a range of problems (Hart, 1988; Kelley et al., 2003; Dantzer and Kelley, 2007). Sickness behavior redirects the body’s resources towards the immune system, and as a consequence leads to a reduction of the animal’s wakefulness and maintenance behaviors. As Hart (1988) stated, “The sick individual is viewed as being at a life or death juncture and its behavior is an all-out effort to overcome the disease”. Further examination as to how closely ‘exhaustion’ behavior mimics sickness behavior would be valuable to the study of working equids. Millman (2007) suggested that animals with sickness behavior should be rested in a quiet area: while this might be suitable for farm animals, it presents a challenge for working equids in developing regions due to the extreme dependency of their owners on the daily income provided by their equids. In defining the relationship between low BCS and unresponsiveness, we found that the causes of low BCS are multifactorial and likely to include malnutrition, overwork, parasitism and diseases. Thus, low BCS could either cause behavioral unresponsiveness, or be a result of it (Hart, 1988; Dantzer and Kelley, 2007; Weary et al., 2009). Finally the partial and total welfare scores complement the results from the separate analysis of the behavior and health measures. The TPC and R horses had the highest and the most desirable behavior scores, indicating the highest prevalence of an alert attitude, a friendly response and acceptance of contact by an observer. These horses typically work at a lower intensity than most other categories in this study. This type of work did not induce either a depressed attitude or a lack of interest in responding to an approaching person. This also indicates responsible human handling that does not induce avoidance of an approaching human and/or contact, or an aggressive response. Also, because these horses are most likely in the public eye, there is additional pressure to their owners to present them in a positive light. The highest health, BCS, body coat and hoof scores were recorded also in TPC and R horses, which also highlights these as the equids with the lowest prevalence of heat stress, disease, low BCS, an unhealthy coat, ectoparasitic infestation, and abnormal hoof shape and horn quality. This in turn indicates an acceptable level of workload, adequate health care, nutrition, coat and hoof care for this type of animal. Similarly, TPC and R horses also had the highest body lesions score, meaning they had the lowest prevalence of lesions due to harnesses, mistreatment, overwork, firing, tethering and accidental or parasite induced lesions. Contrarily, TBC donkeys and mules showed the lowest levels among all scores. This indicates the lowest scores for almost all behavior, health, body lesion, coat and hoof measures, and consequently, a deteriorated life in terms of workloads, nutrition, harness quality, health care, coat and hoof care, and probably improper handling.

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

TGC horses, donkeys and mules had similar scores, except that TGC donkeys and mules were healthier and had higher hoof scores than horses. This could be because of overloading of the horses due to their bigger size and heavier weight, or because of the horses’ reduced ability to tolerate strenuous draft work, especially in the high temperatures. Finally it should be mentioned that TPC and R horses were the only categories that had an “adequate” total welfare score. Except for TBC donkeys, which had a “Poor” total welfare score, all other categories were recorded as “Inadequate”. Thus applicable and thorough plans for intervention should be developed to improve the welfare status of working donkeys in brick kilns, TGC donkeys, mules and horses. These plans should focus on the highest risk areas (e.g., behavior, health, BCS, body lesions, coat or hoof) for each species, gender, age and work type of working equids, and consider their integrated effects on behavior and health. 5. Conclusion This welfare assessment scoring system met our initial objective to be a useful tool in identifying which equids have the most welfare problems (i.e., which species, type of work, age and sex). This, in turn will help in the selection of appropriate interventions or enhancement strategies, and in targeting such interventions toward the most vulnerable working equids. The idea of describing the welfare status of working equids by a number, which can then be assigned a corresponding welfare category of “Adequate, Inadequate or Poor” was designed to be used easily either for many, or just a few, equid welfare assessments. Due to the large number of equids that were assessed in this study, representing a wide array of Egypt‘s working equids, the results serve as a valuable bench mark for future assessments and for comparing intervention and enhancement strategies. It should be viewed as a valuable tool to assist with outcomes assessments amongst working equids. Conflict of interest None. Acknowledgments This study was supported by the Egyptian Ministry of Higher Education. The authors would like to thank The Brooke Egypt and the Society for the Protection and Welfare of Donkeys and Mules in Egypt (SWDPME) members, in addition to Amy McLean, for their participation in the questionnaire and their assistance with data allocation. Full thanks and appreciation to the ISAE and Dr. Keelin O’Driscoll for her generous help in editing the English language of the manuscript, also to the MSU CANR Statistical Consulting Centre and Dr. Luis Montiel for their great help in building up and designing the statistical part of the welfare score, and we would also like to thank all the owners who kindly permitted examination of their horses, mules and donkeys. References Ali, A.B., Matoock, M.Y., Fouad, M.A., Heleski, C.R., 2015. Are mules or donkeys better adapted for Egyptian brick kiln work? (Until we can change the kilns). J. Vet. Behav.: Clin. Appl. Res. 10, 158–165. Ashley, F., Waterman-Pearson, A., Whay, H., 2005. Behavioural assessment of pain in horses and donkeys: application to clinical practice and future studies. Equine Vet. J. 37, 565–575. 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, 179–192. Blench, R., MacDonald, K., 2006. The Origins and Development of African Livestock: Archaeology, Genetics, Linguistics and Ethnography. Routledge.

61

Burn, C.C., Dennison, T.L., Whay, H.R., 2010a. Environmental and demographic risk factors for poor welfare in working horses, donkeys and mules in developing countries. Vet. J. 186, 385–392. Burn, C.C., Dennison, T.L., Whay, H.R., 2010b. Relationships between behaviour and health in working horses, donkeys, and mules in developing countries. Appl. Anim. Behav. Sci. 126, 109–118. Chang, C.R., Sapón, M., Rodríguez, D., 2010. Economic valuation of the impact of the working equine in the Peten and Chimaltenango communities in Guatemala, The 6th International Colloquium on Working Equids: Learning from Others. In: Proceedings of an International Colloquium held at the India Habitat Centre, New Delhi, India, pp. 106–110. Crane, M., 1997. Medical care of donkeys. The Professional Handbook of the Donkey, In: Whittet Books LTD, W.B. (Ed.), Medical. In: Svendsen E (Ed):, London, pp. 19–36. Dantzer, R., Kelley, K.W., 2007. Twenty years of research on cytokine-induced sickness behavior. Brain Behav. Immun. 21, 153–160. Davies, R., 2001. Monitoring and evaluating NGO achievements. In: Potter, R.B., Desaid, V. (Eds.), The Companion to Development Studies. Arnold Publishing, London, p. 28. de Aluja, A.S., 1998. The welfare of working equids in Mexico. Appl. Anim. Behav. Sci. 59, 19–29. Dennison, T., Khan, G., Khan, A., Pritchard, J., Whay, H., 2007. A comparative study of the welfare of equines working in the brick kilns of Multan and Peshawar, Pakistan. In: Fifth International Colloquium on Working Equines The Future for Working Equines, Addis Ababa, Ethiopia 30 2007 October 2 November, 2006., The Donkey Sanctuary, pp. 153–160. FAOSTAT, 2011. FAO Statistical Year Book. Food and Agricultural Organization of the United Nations, http://faostat.fao.org/site/573/DesktopDefault. aspx?PageID. Biological basis of the behavior of sick animals., 1988. Neuroscience & Biobehavioral Reviews 12, 123–137. Hansen, S.W., Møller, S.H., 2001. The application of a temperament test to on-farm selection of mink. Acta Agric. Scand., Sect. A-Anim. Sci. 51 (Suppl. 30), 93–98. http://www.thebrooke.org, 2013. data/assets/pdf file/0008/96524/AR2012 AW7. pdf. Kelley, K.W., Bluthé, R.-M., Dantzer, R., Zhou, J.-H., Shen, W.-H., Johnson, R.W., Broussard, S.R., 2003. Cytokine-induced sickness behavior. Brain Behav. Immun. 17, 112–118. Lemma, A., Moges, M., 2009. Clinical, haematological and serum biochemical reference values of working donkeys (Equus asinus) owned by transport operators in Addis Ababa, Ethiopia. Res. Rural Dev. 21, 8. Main, D.C.J., Clegg, J., Spatz, A., Green, L.E., 2000. Repeatability of a lameness scoring system for finishing pigs. Vet. Rec. 147, 574–576. Martin, M., 2001. The impact of community animal health services on farmers in low-income countries: a literature review. Mekuria, S., Mulachew, M., Abebe, R., 2013. Management practices and welfare problems encountered on working equids in Hawassa town, Southern Ethiopia. J. Vet. Med. Anim. Health 5, 243–250. Millman, S., 2007. Sickness behaviour and its relevance to animal welfare assessment at the group level. Anim.Welf. 16, 123–125. Perry, B.D., 2002. Investing in Animal Health Research to Alleviate Poverty. ILRI (aka ILCA and ILRAD). Poate, D., 1993. More’m’and less’e’: changing ideas on information for development. The Rural. Popescu, S., Diugan, E.-A., 2013. The relationship between behavioral and other welfare indicators of working horses. J. Equine Vet. Sci. 33, 1–12. Popescu, S., Diugan, E., Spinu, M., 2014. The interrelations of good welfare indicators assessed in working horses and their relationships with the type of work. Res. Vet. Sci. 96, 406–414. Pritchard, J., Barr, A., Whay, H., 2006. Validity of a behavioural measure of heat stress and a skin tent test for dehydration in working horses and donkeys. Equine Vet. J. 38, 433–438. Pritchard, J., Lindberg, A., Main, D., Whay, H., 2005. Assessment of the welfare of working horses, mules and donkeys, using health and behaviour parameters. Prev. Vet. Med. 69, 265–283. Pritchard, J., Whay, H., 2004. Guidance Notes to Accompany Working Equine Welfare Assessment. University of Bristol, 1–101. Pritchard, J.C., 2010. Animal traction and transport in the 21st century: Getting the priorities right. Vet. J. 186, 271–274. Pritchard, J.C., Burn, C.C., Barr, A.R., Whay, H.R., 2009. Haematological and serum biochemical reference values for apparently healthy working horses in Pakistan. Res. Vet. Sci. 87, 389–395. Roche, C.J., 1999. Impact Assessment for Development Agencies: Learning to Value Change. Oxfam, Oxford. Rousing, T., Bonde, M., Sørensen, J.T., 2001. Aggregating welfare indicators into an operational welfare assessment system: a bottom-up approach. Acta Agric. Scand., Sect. A-Anim. Sci. 51, 53–57. Singh, M., Gupta, A., Yadav, M., 2005. The donkey: its role and the scope for better management. Livest. Int. 9, 18–20. Sørensen, J.T., Sandøe, P., Halberg, N., 2001. Animal welfare as one among several values to be considered at farm level: the idea of an ethical account for livestock farming. Acta Agric. Scand. Sec. A-Anim. Sci. 51, 11–16. Swann, W.J., 2006. Improving the welfare of working equine animals in developing countries. Appl. Anim. Behav. Sci. 100, 148–151.

62

A.B.A. Ali et al. / Applied Animal Behaviour Science 176 (2016) 52–62

Tadich, N., Mendez, G., Wittwer, F., Meyer, K., 1997. Blood biochemical values of loadcart draught horses in the city of Valdivia (Chile). Arch. Med. Vet. 2981, 45–61. Tadich, T., Escobar, A., Pearson, R., 2008. Husbandry and welfare aspects of urban draught horses in the south of Chile. Arch. Med. Vet. 40, 267–273. Thornton, P., Kruska, R., Henninger, N., Kristjanson, P., Reid, R., Atieno, F., Odero, A., Ndegwa, T., 2002. Mapping Poverty and Livestock in Developing Countries. International Livestock Research Institute, Nairobi. Upjohn, M., Attwood, G., Lerotholi, T., Pfeiffer, D., Verheyen, K., 2013. Quantitative versus qualitative approaches: A comparison of two research methods applied to identification of key health issues for working horses in Lesotho. Prev. Vet. Med. 108, 313–320. Upjohn, M.M., Pfeiffer, D.U., Verheyen, K.L., 2014. Helping working Equidae and their owners in developing countries: monitoring and evaluation of evidence-based interventions. Vet. J. 199, 210–216. Velázquez-Beltrán, L.G., Sánchez-Vera, E., Nava-Bernal, E.G., Arriaga-Jordán, C.M., 2011. The role of working equines to livelihoods in current day campesino hill-slope communities in central Mexico. Trop. Anim. Health Prod. 43, 1623–1632.

Wallace, T., 2010. Monitoring and evaluation in development work: why involving local people matters and how it can be done. In: The 6th International Colloquium on Working Equids: learning from others Proceedings of an International Colloquium, New Delhi, India 29 2010 November–2 December 2010, The Brooke, pp. 1–4. Waring, G.H., 2003. Horse Behavior, Norwich, Library of Congress. Noyes Publications/William Andrew Publishing, NY, pp. 273–281. Weary, D., Huzzey, J., Von Keyserlingk, M., 2009. Board-invited review: using behavior to predict and identify ill health in animals. J. Anim. Sci. 87, 770–777. Whay, H., Main, D., Green, L., Webster, A., 2003. Observations and investigation of farm. Vet. Rec. 153, 197–202. Winckler, C., Willen, S., 2001. The reliability and repeatability of a lameness scoring system for use as an indicator of welfare in dairy cattle. Acta Agric. Scand. Sect. A-Anim. Sci. 51, 103–107. Wood, J., Holder, J., Main, D., 1998. Quality assurance schemes. Meat Sci. 49, S191–S203.