Hostname: page-component-78c5997874-xbtfd Total loading time: 0 Render date: 2024-10-31T08:30:08.718Z Has data issue: false hasContentIssue false

Equipment-related wounds and associated risk factors in working equids of the Oromia national regional state in Ethiopia

Published online by Cambridge University Press:  31 October 2024

Mathilde S Merridale-Punter*
Affiliation:
Melbourne Veterinary School, Faculty of Science, University of Melbourne, Parkville, VIC 3010, Australia
Abel L Wodajo
Affiliation:
College of Veterinary Medicine and Agriculture, Addis Ababa University, PO Box 34, Bishoftu, Ethiopia
Belay Elias
Affiliation:
College of Veterinary Medicine and Agriculture, Addis Ababa University, PO Box 34, Bishoftu, Ethiopia Faculty of Veterinary Medicine, Hawassa University, PO Box 05, Hawassa, Ethiopia
Anna-Marie Bakos
Affiliation:
Melbourne Veterinary School, Faculty of Science, University of Melbourne, Parkville, VIC 3010, Australia
Hanna Zewdu
Affiliation:
College of Veterinary Medicine and Agriculture, Addis Ababa University, PO Box 34, Bishoftu, Ethiopia
Reta Tesfaye
Affiliation:
College of Veterinary Medicine and Agriculture, Addis Ababa University, PO Box 34, Bishoftu, Ethiopia
Gizachew Hailegebreal
Affiliation:
Faculty of Veterinary Medicine, Hawassa University, PO Box 05, Hawassa, Ethiopia
Teshale Sori
Affiliation:
College of Veterinary Medicine and Agriculture, Addis Ababa University, PO Box 34, Bishoftu, Ethiopia
Charles M El-Hage
Affiliation:
Melbourne Veterinary School, Faculty of Science, University of Melbourne, Parkville, VIC 3010, Australia
Anke K Wiethoelter
Affiliation:
Melbourne Veterinary School, Faculty of Science, University of Melbourne, Parkville, VIC 3010, Australia
Peta L Hitchens
Affiliation:
Equine Centre, Melbourne Veterinary School, University of Melbourne, 250 Princes Hwy, Werribee, VIC 3030, Australia
*
Corresponding author: Mathilde Seabra Merridale-Punter; Email: mseabramerri@student.unimelb.edu.au
Rights & Permissions [Opens in a new window]

Abstract

Working equids support the livelihoods of millions of low-income households worldwide and face several welfare challenges. Although equipment-related wounds are common, little is known about specific risk factors. This cross-sectional study surveyed equids used for cart-work in three Ethiopian towns. Number, size, severity and location of wounds were recorded for each animal, as well as work equipment characteristics and indicators of equipment fit and assembly. Questionnaires were conducted with each cart-driver focusing on equipment practices and attitudes. Logistic regression models were generated to investigate associations between equipment-related wounds (outcome) and equipment, work, driver and animal-related factors (predictors). In total, 369 equids and cart-drivers were surveyed. The prevalence of equipment-related wounds was 72.6% (268/369) with girth wounds being the predominant wound type in horses (50%; 122/244) while donkeys had predominantly shaft-related wounds (59%; 72/122). Donkeys were two times more likely to have equipment-related wounds than horses. The presence of equipment-related wounds was associated with factors such as previously having wounds, inadequate collar positioning and using purchased equipment compared to partly home-made equipment. Characteristics of specific equipment components were also associated with certain wound types, such as narrow saddle pressure points and saddle wounds. Equipment-related wounds are highly prevalent in working equids, representing a serious welfare concern. Factors relating to equipment design, fit and assembly were associated with the presence of wounds. Importantly, species differences require tailoring of preventive approaches amongst working equids. A better understanding of work equipment should therefore be promoted as part of wound prevention and animal welfare strategies.

Type
Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2024. Published by Cambridge University Press on behalf of The Universities Federation for Animal Welfare

Introduction

Working equids support millions of low- and middle-income households worldwide, often being a central source of income and playing a key role in the wider community through their use in transport, agriculture, domestic support and supply chains (Fernando & Starkey Reference Fernando and Starkey2004; Sturgeon Reference Sturgeon2021; Grace et al. Reference Grace, Diall, Saville, Warboys, Ward, Wild and Perry2022). Of all the African nations, Ethiopia has the highest number of equids (Food and Agriculture Organisation [FAO] 2021; Norris et al. Reference Norris, Little, Ryding and Raw2021), where over 12 million are estimated to be used for transport and other draught or pack work (Central Statistical Agency [CSA] 2020; Geiger et al. Reference Geiger, Hockenhull, Buller, Tefera Engida, Getachew, Burden and Whay2020).

With working equids frequently working every day of the week (Bazezew et al. Reference Bazezew, Chanie, Tesfaye, Kassa, Mekonnen and Wagaw2014; Ali et al. Reference Ali, Orion, Tesfaye and Zambriski2016), work equipment is an essential part of their lives, and can impact their welfare (Ali et al. Reference Ali, Orion, Tesfaye and Zambriski2016; Teferi et al. Reference Teferi, Tesfaye, Zewdu, Gemechu, Tefera and Ashine2020; Eriso et al. Reference Eriso, Mekuriya and Abose2023a). Equipment-related wounds are one of the most visible manifestations of welfare compromise attributable to equipment and are widely prevalent in working equids around the world (Farhat et al. Reference Farhat, McLean and Mahmoud2020; Robledo-Reyes et al. Reference Robledo-Reyes, Hernández-Gil, Rojas-Hernández, Camacho-Díaz, Cipriano-Salazar, Villa-Mancera and Olivares-Pérez2020; Rodrigues et al. Reference Rodrigues, Sullivan, Judge, Norris and Burden2020). In Ethiopia, wounds in the region of the harness have been commonly reported (Teferi et al. Reference Teferi, Tesfaye, Zewdu, Gemechu, Tefera and Ashine2020; Aliye et al. Reference Aliye, Nigusie, Fesseha and Mathewos2022), with prevalence varying from 13% (Eriso et al. Reference Eriso, Mekuriya and Abose2023b) to as high as 60% (Fikru et al. Reference Fikru, Tadese and Gebreegziabher2015) or 80% (Usman et al. Reference Usman, Disassa, Kabeta, Zenebe and Kebede2015) in different regions. Further, the presence of wounds has also been linked to other conditions such as lameness (Ali et al. Reference Ali, Orion, Tesfaye and Zambriski2016), suggesting that inadequate harnessing can not only impact the animal directly but also carries broader indirect welfare implications.

Work equipment characteristics such as harness materials and fitting are described in harnessing guidelines as a potential risk for injury (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Heleski et al. Reference Heleski, McLean, Swanson and Grandin2009; Garrett Reference Garrett2017), and the type of work and harness design have been shown to affect the presence of wounds (Daneil & Asmare Reference Daneil and Asmare2013; Ali et al. Reference Ali, Orion, Tesfaye and Zambriski2016). Some studies have reported harness cleanliness (Farhat et al. Reference Farhat, McLean and Mahmoud2020), loads carried (Fikru et al. Reference Fikru, Tadese and Gebreegziabher2015; Ali et al. Reference Ali, Orion, Tesfaye and Zambriski2016), and poor general harness quality causing rubbing or friction to be associated with wounds (Ali et al. Reference Ali, Orion, Tesfaye and Zambriski2016; Bereket & Addis Reference Bereket and Addis2019). However, previous surveys have classified harnessing equipment broadly as appropriate or inappropriate, with few studies examining the specific equipment attributes that represent risk factors for the development of wounds. Equid harnesses and carts are complex work equipment with multiple interacting components that act as a unit and depend upon other elements to be adequately functional, efficient and comfortable (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Oudman Reference Oudman2004). Additionally, the way in which this equipment fits the animal and is hitched, as well as the way in which it is used and kept by the driver likely impacts the risk of developing equipment-related wounds (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Garrett Reference Garrett2017; Rodrigues et al. Reference Rodrigues2022).

Guidelines are already available for affordable, efficient and welfare-friendly equipment designs (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Heleski et al. Reference Heleski, McLean, Swanson and Grandin2009; Garrett Reference Garrett2017) but implementation of such guidelines remains challenging with some reports suggesting targeted improvement and modification of existing work equipment may be more successful in effecting change than radical modifications or introduction of new equipment designs (Chanie et al. Reference Chanie, Fentahun, Mitiku and Berhan2012). Therefore, a better understanding of the specific equipment attributes and practices associated with the presence of equipment-related wounds would provide valuable insight into preventing this welfare risk to working equids.

This study therefore aimed to: (1) determine the prevalence of equipment-related wounds in working equids at three Ethiopian locations; and (2) investigate the association between equipment-related wounds (outcome) and work practices, cart-driver demographics and animal and equipment characteristics (predictors). A greater understanding of which work equipment and driver characteristics or practices are associated with the presence of wounds will help devise recommendations that may be implemented by cart and harness makers, as well as informed wound-prevention strategies for working equid owners and users.

Materials and methods

Study design and data collection

Full methodology for data collection and sample size calculation are described in detail in Merridale-Punter et al. (Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024). In brief, this was a cross-sectional study conducted in three locations in the Oromia national regional state of Ethiopia, within the East and North Shewa zones (Bishoftu and Fiche towns, respectively) as well as the West Arsi Zone (Shashamene town). Data were collected between February and April 2022. The study population consisted of working equids (defined in this study as donkeys, horses or mules used for work that involves wearing a harness hitched to a cart) and their corresponding cart-drivers. Recruitment of participants took place via random systematic sampling frame and occurred at dedicated cart-holding stations selected on the basis of convenience within each of the study regions.

A sample size of 369 working equids and their drivers was determined, based on a 20% expected prevalence of equipment-related wounds in the study regions, estimated from previous reports of 17% prevalence in Shashamene (Teferi et al. Reference Teferi, Tesfaye, Zewdu, Gemechu, Tefera and Ashine2020) and 18% in Bishoftu (Chala et al. Reference Chala, Ayele and Tariku2017). Desired absolute precision was set at 5% and adjustments for clustering variations using a 1.5 design effect inflation (Stevenson Reference Stevenson and Sergeant2015). Data collection sheets used for the survey of animal welfare and work equipment indicators, as well as a structured questionnaire with the corresponding cart-drivers are presented in Supplement 1 in the Supplementary material. For each animal, signalment, work type and work intensity indicators were recorded. Animal welfare indicators were selected from and collected in accordance with the validated Standardised Equid-Based Welfare Assessment Tool (SEBWAT) (Sommerville et al. Reference Sommerville, Brown and Upjohn2018). The number, size and severity of wounds (Sommerville et al. Reference Sommerville, Brown and Upjohn2018) were recorded for each animal based on wound location. For each surveyed animal, selected cart and harness characteristics, their fit, materials, design and presence of standard equipment components (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Oudman Reference Oudman2004; Davis Reference Davis2014; Garrett Reference Garrett2017) were recorded (Merridale-Punter et al. Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024). Data collection was carried out by two assessors formally trained for the purpose (ALW and BE).

Ethical approval

The study received ethical approval from the Human and Animal Research Ethics Committees of the University of Melbourne (Reference ID: 2021-22261-24285-4 and 2021-22886-21887-1, respectively), and from the Animal Research Ethics Review Committee of the Addis Ababa University (Reference: VM/ERC/36-37/07/13/2021). Each cart-driver provided informed consent to participate in the study and was compensated for their time with a pre-paid bag of animal feed.

Data processing and analysis

Data collected were entered into REDCap (Harris et al. Reference Harris, Taylor, Thielke, Payne, Gonzalez and Conde2009, Reference Harris, Taylor, Minor, Elliott, Fernandez, O’Neal, McLeod, Delacqua, Delacqua and Kirby2019), a secured electronic data capture software hosted at Melbourne Veterinary School. R software (R Core Team Reference Team2021) was used to analyse the data descriptively and perform inferential statistical analysis. All equid surveys were complete and included in the study (n = 369); one cart-driver questionnaire was incomplete and removed from analysis (n = 368). Wounds located in the region of the bit, blinkers, chest-band, girth, saddle and shafts, as well as limb and tail wounds in contact with equipment were classified as ‘equipment-related wounds’. Wounds located in other body parts were classified as ‘other wounds’. Work type was recorded as the approximate percentage of time the equid spent performing each given type of work. A categorical variable ‘main type of work’ was then created to reflect the predominant work type, where ‘taxi’, ‘transport of goods’ and ‘transport of water’ were considered when the percentage of time performing this work was higher than any other work types, and with ‘mixed work’ referring to equids performing different types of work in similar amounts. Additional variables were created to reflect the functionality of interacting harness components, as well as to classify harness contact surface materials as ‘breathable’ or not according to whether they were made from natural fibres or synthetic, non-breathable materials, respectively (Heleski et al. Reference Heleski, McLean, Swanson and Grandin2009; Merridale-Punter et al. Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024).

Associations between the presence of equipment-related wounds as a binary categorical variable (outcome) and individual equipment characteristics, work practices, species, location, driver demographic and attitudes (predictors) were tested using univariable logistic regression models (see Supplement 2 in the Supplementary material). Due to the low number of mules surveyed (n = 3), only horses and donkeys were considered when investigating associations between species and equipment-related wounds. Collinearity between predictors were tested by pair-wise correlations and Chi-squared tests for associations between categorical variables. Predictor variables from the univariable logistic models with a significance level of P < 0.25 were retained and used to generate a directed acyclic graph (DAG) using DAGitty (available on www.dagitty.net; Textor et al. Reference Textor, Van der Zander, Gilthorpe, Liśkiewicz and Ellison2016) to investigate potential causal pathways between predictors and equipment-related wounds. To generate the DAG, retained predictors were grouped into categories of shared qualities and the predictor that had the highest significance level or was the best biological representation of each category was included in the analysis. Multivariable logistic regression models were then built to estimate the total effect of each predictor category on equipment-related wounds, with minimal sufficient covariate adjustments for the other predictors in the DAG.

Further univariable logistic regression models were used to investigate specific predictors of individual equipment wound types as the outcome (saddle, girth, chest-band, bit, blinker and shaft wounds, and limb and tail wounds attributable to equipment) (see Supplement 2; Supplementary material). Regression coefficients (coef), Odds ratios (OR) and 95% confidence intervals (95% CI) are reported. Statistical significance was set at P < 0.05.

Results

Animal and equipment survey

Three-hundred and sixty-nine working equids were surveyed between the three study locations, of which 66.1% (244/369) were horses, 33.1% (122/369) were donkeys and 0.8% (3/369) were mules. All equids were male (100%; 369/369) with an overall mean (± SD) body condition score (BCS) out of 5 of 2 (± 0.6). For 63.1% (233/369) of equids, taxi work was the main type of work performed, whereas 28.7% (106/369) spent most of their work time transporting goods, doing mixed work (5.1%; 19/369) or transporting water (3.0%; 11/369). Animals worked a median of six days per week (inter-quartile range [IQR] 6-6 days) and 10 h of work per day (IQR 8–10 h). The median number of breaks taken during the workday was two (IQR 1–2 breaks) lasting a median of 60 min per break (IQR 30–60 min). Three percent of drivers (11/368) only removed the harness during breaks sometimes and 4.9% (18/368) did not remove it. Welfare parameters by species and location are summarised in Supplement 3 in the Supplementary material.

The overall prevalence of any type of wound was 75.9% (280/369; 95% CI 71.1–80.1%), with a prevalence of equipment-related wounds of 72.6% (268/369; 95% CI 67.7–77.1%). Eighty-two percent of donkeys (82%; 100/122) and 67.6% (165/244) of horses had equipment-related wounds. Overall, girth wounds were the most prevalent type of equipment-related wound among surveyed equids (43.9%; 162/369), followed by breast collar (23.3%; 86/369) wounds, saddle (21.7%; 80/369), shaft (20.3%; 75/369), bit (12.7%; 47/369), limb (6.5%; 24/369), blinker (3.3%; 12/369) and tail wounds (1.1%; 4/369). When stratifying by species, girth wounds (50%; 122/244) were the most common wound type in horses, while in donkeys shaft (59%; 72/122) and bit (31.1%; 38/122) wounds were the most prevalent (Figure 1). Median scores and IQRs for the size, severity and number of each wound type are summarised in Supplement 3 (Supplementary material).

Figure 1. Overall prevalence of different types of equipment-related wounds in a cross-sectional study investigating equipment-related wounds and associated factors in working equids at three Ethiopian locations in 2022.

Cart-driver questionnaires

Findings of the cart-driver questionnaires are detailed in Merridale-Punter et al. (Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024). Briefly, four drivers were female (1.1%; 4/368) and median cart-driver age was 28 years (IQR 22–35 years). The median cart-driving experience was four years (IQR 3–6 years), and 6.8% of respondents (25/368) were exclusively cart-drivers while 93.2% (343/368) both owned and drove the equid. The majority of respondents purchased their work equipment (84.7%; 310/366), while some (8.5%; 31/366) reported partly purchasing and partly home-making it. A total of 76.6% (282/368) of cart-drivers reported their equid had a previous history of wounds, and 28.8% (106/368) actively took measures to prevent wounds. The most common measure to prevent wounds was taking the animal to a veterinary professional (42.5%; 45/106) (Table 1). Questionnaire responses are described in Supplement 4 in the Supplementary material.

Table 1. Equid wound history and wound prevention approaches of cart-drivers in a cross-sectional study investigating equipment-related wounds in Ethiopian working equids in 2022. Overall equids include horses, donkeys and mules, although due to the low frequency of mules (n = 3) species comparison is presented for horses and donkeys only

Factors associated with equipment-related wounds

Overall presence of equipment-related wounds

Table 2 presents significant predictors of equipment-related wounds found on univariable analysis.

Table 2. Univariably significant predictors of equipment-related wounds in a study investigating the prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Comparison of species was done between horses (n = 244) and donkeys (n = 122) only

Collinearity was present between several predictor variables. Town and species were strongly associated, where the surveyed population in Fiche and Bishoftu was made up almost entirely of horses (98.8%; 242/245) while equids surveyed in Shashamene were almost exclusively donkeys (96.8%; 120/124; χ2 = 348.3; P < 0.001).

Nine predictor categories were created in the DAG based on retained predictors with a significance level of P < 0.25: ‘town’, ‘species’, ‘main work type’, ‘driver education’, ‘source of equipment’, ‘training source’ (including source of training and source of information on hitching and assembly; represented by training source), ‘equipment quality’ (including cost of cart, presence of traces, neck-strap, swingle, breeching, and equipment receiving maintenance; represented by cost of cart), ‘hitching’ (including collar traction through point of draught and presence of functional breeching; represented through traction through point of draught) and ‘welfare standard’ (including BCS, history of wounds and measures to prevent wounds; represented by history of wounds) (Table 3; Figure 2).

Table 3. Results of the directed acyclic graph (DAG) -based statistical modelling of the total effects of exposure variables on the likelihood of equipment-related wounds, in a cross-sectional study investigating prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Exposure variables within DAG category were P < 0.25 in univariable analysis

Figure 2. Directed acyclic graph (DAG) representing possible causal pathways between predictors of equipment-related wounds, in a cross-sectional study investigating prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Only predictors with a significance level of P < 0.25 in univariable logistic regression analysis were retained and represented in the DAG. Collinearity between predictors was present and town was associated with all other predictors in the DAG. Predictors coloured in red represent significant predictors of equipment-related wounds after adjustments (P < 0.05) and those in light blue represent non-significant variables. The model outcome (equipment-related wounds) is represented in darker blue.

Assessing total effects, only town, equipment source and prior history of wounds were significant predictors of equipment-related wounds after minimal sufficient adjustments (Table 3; Figure 2). Animals from Shashamene were more likely to have wounds than those from Fiche (OR 2.56; 95% CI 1.40–4.63) or Bishoftu (OR 2.22; 95% CI 1.20–4.07; P = 0.006), purchased equipment was more likely to be associated with wounds than partly purchased and partly home-made equipment (OR 3.36; 1.36–8.29; P = 0.009), and animals with a history of wounds were more likely to have equipment-related wounds than those with no previous history (OR 9.71; 95% CI 5.47–17.20; P < 0.001) (Table 3).

Presence of individual equipment-related wound types

Significant associations between population and equipment characteristics and the presence of specific types of equipment-related wounds are presented in Table 4. Specific equipment characteristics were associated with certain wound types: for example, narrow saddle pressure points (OR 2.83; 95% CI 1.62–4.95; P < 0.001) and insufficient (OR 2.55; 95% CI 1.41–4.55) or excessive (OR 3.24; 95% CI 1.35–7.78; P = 0.004) saddle padding were associated with the presence of saddle wounds (Table 4). Poor bit fit (OR 0.24; 95% CI 0.08–0.68; P = 0.007) and condition of the bit (OR 0.18; 95% CI 0.04–0.76; P = 0.02) were associated with the presence of bit wounds.

Table 4. Significant predictors of specific types of equipment-related wounds through univariable logistic regression analysis in a study investigating the prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Comparison of species was done between horses and donkeys only. Due to high collinearity, associations with town are not represented in this table

Discussion

This study investigated the prevalence and factors associated with equipment-related wounds in working equids in three Ethiopian locations. A high overall prevalence of equipment-related wounds was found affecting 7 out of 10 working equids, thus representing a serious welfare concern for these animals. This study suggests that the likelihood of equipment-related wounds is influenced by species, town, work-type and equipment source, and that animals with a low welfare standard are more likely to have equipment-related wounds. Additionally, specific equipment characteristics and practices influenced the likelihood of having equipment-related wounds.

More than two-thirds of the equids in this study had at least one type of equipment-related wound. Although reporting of the causes and classification of lesions are not generally standardised, and the aetiology of wounds is multifactorial and could be influenced by several factors external to the equipment, such as age, work-type or low body condition (Pritchard et al. Reference Pritchard, Lindberg, Main and Whay2005; Burn et al. Reference Burn, Dennison and Whay2010; Sánchez-Casanova et al. Reference Sánchez-Casanova, Masri-Daba, Alonso-Díaz, Méndez-Bernal, Hernández-Gil and Fernando-Martínez2014); this is among the highest prevalence reported for the Oromia region (Usman et al. Reference Usman, Disassa, Kabeta, Zenebe and Kebede2015). Donkeys had a higher prevalence of equipment-related wounds than horses affecting 8 out of 10 donkeys, which is consistent with other studies (Demelash Biffa Reference Demelash Biffa2006; Usman et al. Reference Usman, Disassa, Kabeta, Zenebe and Kebede2015). Other authors suggest a higher wound prevalence in horses (Kompi et al. Reference Kompi, Molapo and Mosebo2023) or no significant species differences (Fikru et al. Reference Fikru, Tadese and Gebreegziabher2015), which may be due to different study regions and designs, or differing risk factors and environmental contexts. However, donkeys have frequently been reported to have a lower welfare standard than horses (Sells et al. Reference Sells, Pinchbeck, Mezzane, Ibourki and Crane2010; Haddy et al. Reference Haddy, Burden, Prado‐Ortiz, Zappi, Raw and Proops2021) and a lower perceived value (Fernando & Starkey Reference Fernando and Starkey2004; Geiger et al. Reference Geiger, Hockenhull, Buller, Tefera Engida, Getachew, Burden and Whay2020). Additionally, previous history of wounds was strongly associated with equipment-related wounds, suggesting poor welfare sustains further welfare compromise. This could be explained by a failure to adequately address the root causes of the problem, such as equipment design, work or husbandry practices. Similarly, the association between town and wounds may reflect regional variations in the standard of welfare, species, work type or quality of equipment among other factors. As such, Shashamene may have a higher prevalence of wounds because donkeys are more common than horses in that town, but potentially also due to other environmental, population or contextual characteristics within the region. Town, work-type, history of wounds and species were not independent of each other, and thus it could not be determined whether it was the individual factor, or a combination of them that increased the risk of wounds. Nonetheless, this suggests the issue varies according to location and population of animals and targeted advice and solutions are therefore required.

The absence of certain equipment components was associated with a greater likelihood of having equipment-related wounds, which can be expected given the role and described function of these components within a harnessing system. For example, the presence of a neck strap, traces and swingle tree within a breast collar harness system is designed to enable efficient traction through the collar, maintain it in an appropriate position and reduce its friction against the animal’s shoulders during movement (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Oudman Reference Oudman2004). Therefore, the absence of these components would not only decrease efficiency of traction but also reasonably increase the likelihood of friction wounds in the region of the collar as observed in this study. When considering specific types of equipment-related wounds, lesions in the region of the girth were the overall most common type of wound, followed by collar and saddle wounds. However, this differed between species and while girth, collar and saddle wounds were most prevalent in horses, which is similar to reports from other countries (Pritchard et al. Reference Pritchard, Lindberg, Main and Whay2005; Teferi et al. Reference Teferi, Tesfaye, Zewdu, Gemechu, Tefera and Ashine2020), wounds in the region of the shafts and bit were the most prevalent in donkeys. This suggests equipment characteristics vary between horses and donkeys with equipment used for donkeys generally being of a lower standard (Merridale-Punter et al. Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024).

Several equipment characteristics were also associated with specific types of equipment-related wounds. For example, saddle sores were associated with saddles that had narrow pressure points or excessive or insufficient levels of padding, and bit wounds were associated with a poor bit fit and condition. Some associations were less intuitive and potentially contrary to recommendations (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Heleski et al. Reference Heleski, McLean, Swanson and Grandin2009; Garrett Reference Garrett2017): for example, girths made from synthetic materials were less associated with wounds than those made from fabrics or other natural, breathable materials. This is a similar finding to what has been previously found for strap lesions in working donkeys (Burn et al. Reference Burn, Pritchard, Farajat, Twaissi and Whay2008), and could reflect the fact that more porous materials can easily absorb sweat and dirt which, if not adequately cleaned and maintained, may make them stiffer, dirtier and more abrasive. Similarly, an excessive level of saddle padding would increase under-saddle heat, pressure and sweating, increasing friction and the likelihood of wounds (Garrett Reference Garrett2017), particularly if the girth is loose. Draught dynamics are complex, and equipment components cannot be viewed in isolation but rather considered as part of a dynamic interacting unit which is also influenced by factors such as work practices, fitting, loading and weight distribution. As such, equipment practices such as maintenance must also be considered in the context of wound prevention. Nonetheless, these findings provide guidance for specific adjustments in the design or use of individual components. Prioritising targeted improvement of existing equipment, rather than replacing entire harness designs, could help increase user uptake of preventive strategies in a more cost-effective way (Chanie et al. Reference Chanie, Fentahun, Mitiku and Berhan2012).

Equipment source was a significant predictor of equipment-related wounds, with purchased and particularly inherited equipment being more associated with the presence of wounds than partially home-made equipment. As suggested in our previous study, equipment may be being purchased from sources where the design or materials are not fit for purpose, potentially due to a lack of technical knowledge or capacity among equipment makers (Merridale-Punter et al. Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024). As the cost of the harness was not associated with the presence of wounds, this could also suggest that simple, home-made or inexpensive harness components may be efficient and fit for purpose if designed appropriately, as reported by other authors for different full-collar designs (Rodrigues et al. Reference Rodrigues, Garrett, Norris, Albuquerque, Queijo, Cooke and Judge2021). This becomes particularly relevant in the context of the economic pressures faced by most working equid users (Geiger et al. Reference Geiger, Hockenhull, Buller, Tefera Engida, Getachew, Burden and Whay2020) and given that cost is often the main motivator and barrier to obtaining the ideal work equipment (Merridale-Punter et al. Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024). Partial home-made production or adaptations may also reflect a higher standard of care for the equid, and may allow for equipment adjustments that provide a better fit to the individual animal or species, improving its comfort and efficiency where a range of equipment models and sizes may not be commercially available or affordable (Rodrigues et al. Reference Rodrigues, Garrett, Norris, Albuquerque, Queijo, Cooke and Judge2021; Merridale-Punter et al. Reference Merridale-Punter, Elias, Wodajo, El-Hage, Zewdu, Tesfaye, Hailegebreal, Sori, Wiethoelter and Hitchens2024). Conversely, carts that cost less were associated with presence of equipment-related wounds. Carts are more complex items to produce at home, and it can be reasonably assumed that most carts are purchased from manufacturers. This suggests there may be price-related differences in cart design and quality as well as knowledge gaps among manufacturers, and highlights the importance of considering the vehicle itself and the way in which it is designed, hitched, loaded and used as an extension of the work equipment (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Oudman Reference Oudman2004; Garrett Reference Garrett2017). As such, understanding the principles of harnessing and equipment design and function is not only relevant for cart and harness makers but also to working equid users and could be incorporated in community and professional training initiatives aimed at improving equipment standards. It should be noted that most participants purchased their equipment with no differentiation between individual producers or markets considered in this study, and equipment source was not independent of town and other predictors of wounds. There may therefore be region-specific variations in the standard of equipment sources which requires further investigation.

In this study, demographic characteristics of cart drivers were not associated with equipment-related wounds, unlike other reports from Ethiopia where lower driver education was associated with greater prevalence of wounds and other indicators of poor welfare (Molla et al. Reference Molla, Dembela, Megersa and Mekuria2017; Teferi et al. Reference Teferi, Tesfaye, Zewdu, Gemechu, Tefera and Ashine2020). Similarly, income comfort level was not associated with equipment-related wounds (and, in fact, exclusively purchased equipment was associated with wounds), which supports findings from other reports that suggest working equid welfare does not necessarily correlate with the owner’s socioeconomic status and level of education (Bazezew et al. Reference Bazezew, Chanie, Tesfaye, Kassa, Mekonnen and Wagaw2014; Lanas et al. Reference Lanas, Luna and Tadich2018). Conversely, the prevalence of equipment-related wounds was associated with hitching and harnessing practices, where not fitting the breast collar over the animal’s point of draught (POD) was almost twice as likely to be associated with wounds compared to a collar positioned over the POD. This is in line with existing harnessing recommendations, increasing comfort and efficiency of traction as well as decreasing the likelihood of injury (Pearson et al. Reference Pearson, Simalenga and Krecek2003; Oudman Reference Oudman2004; Davis Reference Davis2014). Once more, this also suggests that in addition to the equipment itself, the practices and ways in which the equipment is used must also be considered part of prevention strategies. As such, there is an opportunity for education on equipment design and practices, where human behaviour change could lead to a decrease in equipment-related wounds regardless of demographic or socioeconomic constraints.

There are several limitations to this study. Equipment-related wounds were recorded based on their location in relation to the equipment, but we did not have sufficient information to confirm their cause. Further longitudinal studies would be required to confirm the causal factors of equipment-related wounds. Additionally, the effects of equipment components on wounds and other welfare indicators cannot be looked at in isolation as they are part of an interconnected animal-harness-cart apparatus with complex interacting traction dynamics, which would require investigation in motion and under controlled conditions of work intensity, environment, load distribution, among other factors. Nonetheless, important insights into specific equipment characteristics associated with wounds were found in a non-invasive observational manner, identifying key areas for improvement and further investigation. We classified wounds in accordance with the SEBWAT tool definition of lesions (Sommerville et al. Reference Sommerville, Brown and Upjohn2018), ranging from superficial or healed skin lesions to deep open wounds, which may influence prevalence results in relation to other studies using alternative classifications. However, descriptive results of wound size and severity are provided for transparency, and this allows for result comparison with other literature using the same standardised assessment tool. The temperament and behaviour of equids may differ between castrated and entire males and should be considered in future research as it may affect the prevalence of wounds. Some equipment classifications may be influenced by the point within the workday when the animal was surveyed among other factors. However, limiting all data collection to two trained researchers reduced misclassification bias. Finally, regional differences may mean these findings are not applicable to other sectors of the population, and the multicollinearity obtained between town, species, work type and other predictor variables prevents accurate determination of the influence of each individual predictor on the likelihood of equipment-related wounds. Nonetheless, this work offers further characterisation of potentially problematic equipment attributes than previous literature focusing on overall adequacy or inadequacy of equipment; and highlights equipment-related wounds as a complex and multifactorial welfare issue requiring context-specific solutions.

Animal welfare implications

Equipment-related wounds are highly prevalent in Ethiopian working equids thereby posing an important welfare concern. Although a better understanding of work-related risk factors for welfare is still needed; species, work-type and regional differences in both the prevalence and the type of equipment-related wounds suggest that tailored preventive strategies and equipment adaptations are required. Driver demographics and economic comfort-level did not influence the presence of equipment-related wounds, and several equipment attributes associated with wounds could be tangibly modifiable without significant cost implications. Equipment source, harnessing and hitching practices also affected the likelihood of having equipment-related wounds, highlighting the importance of including human behaviour change as part of wound prevention strategies. A better understanding of equipment function, assembly and use among different stakeholders is therefore needed to safeguard the welfare of working equids, and training and education efforts should be targeted not only at cart drivers for improved harnessing, hitching and equipment care practices, but also at harness and cart makers for improved equipment design.

Supplementary material

The supplementary material for this article can be found at http://doi.org/10.1017/awf.2024.52.

Acknowledgements

The authors are grateful to all working equid drivers who participated in this study, as well as to the College of Veterinary Medicine and Agriculture of the Addis Ababa University and the international non-governmental organisation Society for the Protection of Animals Abroad (SPANA) for facilitating contacts and logistic support for this study.

Competing interest

None.

Footnotes

Author contributions: Conceptualisation: MMP, HZ, CME, AKW, PLH; Formal analysis: MMP, ALW, BE, AMB, AKW, PLH; Investigation: MMP, ALW, BE, AMB, HZ, RT, TS, CME, AKW, PLH; Methodology: MMP, ALW, BE, HZ, RT, TS, CME, AKW, PLH; Resources: ALW; Project administration: MMP, HZ, RT, GH, TS, AKW, PLH; Supervision: PLH, HZ, RT, GH, TS, CME, AKW; Visualisation: MMP; Writing – original draft: MMP; Writing – review & editing: MMP, ALW, BE, AMB, HZ, RT, GH, TS, CME, AKW, PLH

References

Ali, A, Orion, S, Tesfaye, T and Zambriski, JA 2016 The prevalence of lameness and associated risk factors in cart mules in Bahir Dar, Ethiopia. Tropical Animal Health and Production 48: 14831489. https://doi.org/10.1007/s11250-016-1121-7CrossRefGoogle ScholarPubMed
Aliye, S, Nigusie, K, Fesseha, H and Mathewos, M 2022 Study on welfare and health status of working equines in and around shashamene town, Ethiopia. Emerging Animal Species 3: 100004. https://doi.org/10.1016/j.eas.2022.100004CrossRefGoogle Scholar
Bazezew, M, Chanie, M, Tesfaye, T, Kassa, A, Mekonnen, B and Wagaw, N 2014 Lameness and associated risk factors in Cart Mules in Northwestern Ethiopia. Global Veterinaria 12(6): 869877. https://doi.org/10.5829/idosi.gv.2014.12.06.84289Google Scholar
Bereket, MT and Addis, KG 2019 The neglected welfare statue of working donkeys in Ethiopia: The case of Dale district in southern Ethiopia. Journal of Veterinary Medicine and Animal Health 11(1): 611. https://doi.org/10.5897/JVMAH2018.0717CrossRefGoogle Scholar
Burn, CC, Dennison, TL and Whay, HR 2010 Environmental and demographic risk factors for poor welfare in working horses, donkeys and mules in developing countries. The Veterinary Journal 186(3): 385392. https://doi.org/10.1016/j.tvjl.2009.09.016CrossRefGoogle ScholarPubMed
Burn, CC, Pritchard, JC, Farajat, M, Twaissi, AA and Whay, HR 2008 Risk factors for strap-related lesions in working donkeys at the World Heritage Site of Petra in Jordan. The Veterinary Journal 178(2): 263271. https://doi.org/10.1016/j.tvjl.2007.07.014CrossRefGoogle ScholarPubMed
Chala, F, Ayele, B and Tariku, T 2017 Prevalence of work related wound and the associated risk factors in cart horses in Bishoftu Town, Central Ethiopia. Prevalence 7(17).Google Scholar
Chanie, M, Fentahun, T, Mitiku, T and Berhan, M 2012 Strategies for improvement of draft animal power supply for cultivation in Ethiopia: a review. European Journal of Biological Sciences 4(3): 96104. https://doi.org/10.5829/idosi.ejbs.2012.4.3.6648Google Scholar
CSA 2020 Report on livestock and livestock characteristics (private peasant holdings). Federal Democratic Republic of Ethiopia Central Statistical Agency: Addis Ababa, Ethiopa. https://www.statsethiopia.gov.et/wp-content/uploads/2020/12/Livestock-and-Livestock-Characteristics-Private-Peasant-Holdings-Meher-Season-2019-20-2012-E.C..pdf (accessed 13 August 2024).Google Scholar
Daneil, G and Asmare, AA 2013 Assesment on the welfare and its influence on the health of cart horses in Combolcha District, Ethiopia. Folia Vet 57: 183189.Google Scholar
Davis, T 2014 The Four Principles of Animal Draught. Practical Action. https://www.ctc-n.org/sites/default/files/resources/the_four_principles_of_animal_draught.pdf (accessed 18 September 2024).Google Scholar
Demelash Biffa, D 2006 Causes and factors associated with occurrence of external injuries in working equines in Ethiopia. International Journal of Applied Research in Veterinary Medicine 4: 17.Google Scholar
Eriso, M, Mekuriya, M and Abose, Y 2023a Prevalence, causes, and risk factors of wounds in working donkeys at Duna Woreda, Hadiya Zone, Ethiopia. Journal of Cardiology Research Reviews & Reports 184: 23. https://doi.org/10.47363/JCRRR/2023(4)184Google Scholar
Eriso, M, Mekuriya, M and Abose, Y 2023b Wounds in working donkeys: Prevalence, causes, and risk factors at Duna Woreda, Hadiya Zone, Ethiopia. Studies in Social Science & Humanities 2(7): 3540. https://doi.org/10.56397/SSSH.2023.07.06CrossRefGoogle Scholar
FAO 2021 FAOSTAT, Production Data, Live Animals. https://www.fao.org/faostat/en/#data/QCL (accessed 26 July 2023).Google Scholar
Farhat, SF, McLean, AK and Mahmoud, HF 2020 Welfare assessment and identification of the associated risk factors compromising the welfare of working donkeys (Equus asinus) in Egyptian brick kilns. Animals 10(9): 1611. https://doi.org/10.3390/ani10091611CrossRefGoogle ScholarPubMed
Fernando, P and Starkey, P 2004 Donkeys and development: Socio-economic aspects of donkey use in Africa. Donkeys, People and Development. A Resource Book in the Animal Traction Network for Eastern and Southern Africa (ATNESA) 15.Google Scholar
Fikru, A, Tadese, A and Gebreegziabher, Z 2015 Prevalence of equine wound and associated risk factors in and around Kombolcha town, North Ethiopia. Journal of Veterinary Science and Technology 6(4). https://doi.org/10.4172/2157-7579.1000234Google Scholar
Garrett, C 2017 The good harness guide. The Donkey Sanctuary: UK.Google Scholar
Geiger, M, Hockenhull, J, Buller, H, Tefera Engida, G, Getachew, M, Burden, FA and Whay, HR 2020 Understanding the attitudes of communities to the social, economic, and cultural importance of working donkeys in rural, peri-urban, and urban areas of Ethiopia. Frontiers in Veterinary Science 7: 60. https://doi.org/10.3389/fvets.2020.00060CrossRefGoogle Scholar
Grace, D, Diall, O, Saville, K, Warboys, D, Ward, P, Wild, I and Perry, BD 2022 The global contributions of working equids to sustainable agriculture and livelihoods in Agenda 2030. EcoHealth 19(3): 342353. https://doi.org/10.1007/s10393-022-01613-8CrossRefGoogle ScholarPubMed
Haddy, E, Burden, F, Prado‐Ortiz, O, Zappi, H, Raw, Z and Proops, L 2021 Comparison of working equid welfare across three regions of Mexico. Equine Veterinary Journal 53(4): 763770. https://doi.org/10.1111/evj.13349CrossRefGoogle ScholarPubMed
Harris, PA, Taylor, R, Minor, BL, Elliott, V, Fernandez, M, O’Neal, L, McLeod, L, Delacqua, G, Delacqua, F and Kirby, J 2019 The REDCap consortium: building an international community of software platform partners. Journal of Biomedical Informatics 95: 103208.CrossRefGoogle ScholarPubMed
Harris, PA, Taylor, R, Thielke, R, Payne, J, Gonzalez, N and Conde, JG 2009 Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics 42(2): 377381.CrossRefGoogle ScholarPubMed
Heleski, C, McLean, A, Swanson, J and Grandin, T 2009 Practical methods for improving the welfare of horses, donkeys and other working draught animals in developing areas. Improving Welfare: A Practical Approach pp 252273.Google Scholar
Kompi, PP, Molapo, S and Mosebo, KA 2023 Prevalence of equine wounds and associated risk factors in different agro-ecological zones of Lesotho. Journal of Advanced Veterinary Research 13(8): 16641667.Google Scholar
Lanas, R, Luna, D and Tadich, T 2018 The relationship between working horse welfare and their owners’ socioeconomic status. Animal Welfare 27(1): 4754. https://doi.org/10.7120/09627286.27.1.047CrossRefGoogle Scholar
Merridale-Punter, MS, Elias, B, Wodajo, AL, El-Hage, CM, Zewdu, H, Tesfaye, R, Hailegebreal, G, Sori, T, Wiethoelter, AK and Hitchens, PL 2024 Putting the cart before the horse: mixed-methods participatory investigation of working equid harnessing practices in three selected towns of the Oromia national regional state in Ethiopia. BMC Veterinary Research 20(1): 113. https://doi.org/10.1186/s12917-024-03967-3CrossRefGoogle ScholarPubMed
Molla, B, Dembela, S, Megersa, B and Mekuria, W 2017 The welfare, watering, housing, feeding and working management of working donkeys in and around Hawassa City, Southern Ethiopia. Journal of Veterinary Research and Animal Husbandry 2(1): 106.Google Scholar
Norris, SL, Little, HA, Ryding, J and Raw, Z 2021 Global donkey and mule populations: Figures and trends. PLoS One 16(2): e0247830. https://doi.org/10.1371/journal.pone.0247830CrossRefGoogle ScholarPubMed
Pearson, RA, Simalenga, TE and Krecek, RC 2003 Harnessing and hitching donkeys, mules and horses for work. Centre for Tropical Veterinary Medicine, University of Edinburgh, UK.Google Scholar
Pritchard, J, Lindberg, A, Main, D and Whay, H 2005 Assessment of the welfare of working horses, mules and donkeys, using health and behaviour parameters. Preventive Veterinary Medicine 69(3–4): 265283. https://doi.org/10.1016/j.prevetmed.2005.02.002CrossRefGoogle ScholarPubMed
Team, R Core 2021 R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing: Vienna, Austria.Google Scholar
Robledo-Reyes, EE, Hernández-Gil, M, Rojas-Hernández, S, Camacho-Díaz, LM, Cipriano-Salazar, M, Villa-Mancera, A and Olivares-Pérez, J 2020 Management and welfare of working equids in the Guerrero state. Ecosistemas y Recursos Agropecuarios 7(2). https://doi.org/10.19136/era.a7n2.2333Google Scholar
Rodrigues, JB et al. 2022 The Donkey Academy - Harness and Working Equids (WRKE 102). https://www.thedonkeysanctuary.org.uk/form/donkey-academy (accessed 13 August 2024).Google Scholar
Rodrigues, JB, Garrett, C, Norris, SL, Albuquerque, F, Queijo, L, Cooke, F and Judge, A 2021 Collar pressure mapping: An evaluation of seven collar types used on working donkeys in Europe. Veterinary Record 189(8): e772. https://doi.org/10.1002/vetr.772CrossRefGoogle ScholarPubMed
Rodrigues, JB, Sullivan, RJ, Judge, A, Norris, SL and Burden, FA 2020 Quantifying poor working equid welfare in Nepalese brick kilns using a welfare assessment tool. Veterinary Record 187(11): 445. https://doi.org/10.1136/vr.106135CrossRefGoogle ScholarPubMed
Sánchez-Casanova, RE, Masri-Daba, M, Alonso-Díaz, , Méndez-Bernal, A, Hernández-Gil, M and Fernando-Martínez, JA 2014 Prevalence of cutaneous pathological conditions and factors associated with the presence of skin wounds in working equids in tropical regions of Veracruz, Mexico. Tropical Animal Health and Production 46: 555561. https://doi.org/10.1007/s11250-013-0529-6CrossRefGoogle ScholarPubMed
Sells, P, Pinchbeck, G, Mezzane, H, Ibourki, J and Crane, M 2010 Pack wounds of donkeys and mules in the Northern High Atlas and lowlands of Morocco. Equine Veterinary Journal 42(3): 219226. https://doi.org/10.2746/042516409X478532CrossRefGoogle ScholarPubMed
Sommerville, R, Brown, AF and Upjohn, M 2018 A standardised equine-based welfare assessment tool used for six years in low and middle income countries. PLoS One 13(2): e0192354. https://doi.org/10.1371/journal.pone.0192354CrossRefGoogle ScholarPubMed
Stevenson, M and Sergeant, E 2015 Package ‘epiR’. https://cran.r-project.org/web/packages/epiR/index.html (accessed 18 September 2024).Google Scholar
Sturgeon, B 2021 Working Animals–One Health, One Welfare. One Welfare in Practice pp 279317. CRC Press: Boca Raton, Florida, USA.CrossRefGoogle Scholar
Teferi, M, Tesfaye, R, Zewdu, H, Gemechu, G, Tefera, G and Ashine, T 2020 Prevalence of wound, its associated risk factors and wound management practices in carthorses of ten selected towns of Ethiopia. Ethiopian Veterinary Journal 24(2): 115. https://doi.org/10.4314/evj.v24i2.1CrossRefGoogle Scholar
Textor, J, Van der Zander, B, Gilthorpe, MS, Liśkiewicz, M and Ellison, GT 2016 Robust causal inference using directed acyclic graphs: the R package ‘dagitty’. International Journal of Epidemiology 45(6): 18871894. https://doi.org/10.1093/ije/dyw341Google ScholarPubMed
Usman, S, Disassa, H, Kabeta, T, Zenebe, T and Kebede, G 2015 Health and welfare related assessment of working equine in and Around Batu Town, East Shoa, Central Ethiopia. National Science 13(10): 18.Google Scholar
Figure 0

Figure 1. Overall prevalence of different types of equipment-related wounds in a cross-sectional study investigating equipment-related wounds and associated factors in working equids at three Ethiopian locations in 2022.

Figure 1

Table 1. Equid wound history and wound prevention approaches of cart-drivers in a cross-sectional study investigating equipment-related wounds in Ethiopian working equids in 2022. Overall equids include horses, donkeys and mules, although due to the low frequency of mules (n = 3) species comparison is presented for horses and donkeys only

Figure 2

Table 2. Univariably significant predictors of equipment-related wounds in a study investigating the prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Comparison of species was done between horses (n = 244) and donkeys (n = 122) only

Figure 3

Table 3. Results of the directed acyclic graph (DAG) -based statistical modelling of the total effects of exposure variables on the likelihood of equipment-related wounds, in a cross-sectional study investigating prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Exposure variables within DAG category were P < 0.25 in univariable analysis

Figure 4

Figure 2. Directed acyclic graph (DAG) representing possible causal pathways between predictors of equipment-related wounds, in a cross-sectional study investigating prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Only predictors with a significance level of P < 0.25 in univariable logistic regression analysis were retained and represented in the DAG. Collinearity between predictors was present and town was associated with all other predictors in the DAG. Predictors coloured in red represent significant predictors of equipment-related wounds after adjustments (P < 0.05) and those in light blue represent non-significant variables. The model outcome (equipment-related wounds) is represented in darker blue.

Figure 5

Table 4. Significant predictors of specific types of equipment-related wounds through univariable logistic regression analysis in a study investigating the prevalence and factors associated with equipment-related wounds in Ethiopian working equids in 2022. Comparison of species was done between horses and donkeys only. Due to high collinearity, associations with town are not represented in this table

Supplementary material: File

Merridale-Punter et al. supplementary material 1

Merridale-Punter et al. supplementary material
Download Merridale-Punter et al. supplementary material 1(File)
File 868.1 KB
Supplementary material: File

Merridale-Punter et al. supplementary material 2

Merridale-Punter et al. supplementary material
Download Merridale-Punter et al. supplementary material 2(File)
File 165.8 KB
Supplementary material: File

Merridale-Punter et al. supplementary material 3

Merridale-Punter et al. supplementary material
Download Merridale-Punter et al. supplementary material 3(File)
File 165.3 KB
Supplementary material: File

Merridale-Punter et al. supplementary material 4

Merridale-Punter et al. supplementary material
Download Merridale-Punter et al. supplementary material 4(File)
File 207.7 KB