1 Epidemiology and preclinical management of dog bites 2 among humans in Wakiso and Kampala districts, Uganda: 3 Implications for prevention of dog bites and rabies. 4 Stevens Kisaka1,2,3*, Fredrick E. Makumbi2, Samuel Majalija3, Alexander Bangirana4, SM 5 Thumbi1,5,6 6 1 University of Nairobi Institute of Tropical and Infectious Diseases, Nairobi, Kenya 7 2 School of Public Health, Makerere University, Kampala, Uganda 8 3 College of Veterinary Medicine, Animal Resources and Biosecurity, Makerere University 9 4 Department of Emergence Medicine, Mulago National Referral Hospital, Kampala, Uganda 10 5 Rabies Free Africa, Washington State University, Pullman, US 11 6 Paul G Allen School for Global Animal Health, Washington State University, Pullman, US 12 *Corresponding author: skisaka@chs.mak.ac.ug (SK) 13 Author Contributions 14 Conceived and designed the study: SK, FEM, SMT. Conceived the data collection protocols: SK, 15 FEM, SM, AB, SMT. Collected and analyzed the data: SK, SM, AB, FEM. Wrote the paper: SK, 16 FEM, SM, SMT. 17 Author summary 18 Dog-mediate rabies is on the rise, especially in sub Saharan Africa. Though the disease is fatal 19 upon exposure, it can be effectively prevented through appropriate post-exposure management. It 1 20 is recommended that dog bite victims wash bite wounds for 15 minutes with soap, water and 21 disinfectant 22 recommendations are not usually followed in many societies, including in Uganda. There are 23 numerous reports of victims not seeking or delaying to seek healthcare. Additionally, victims have 24 been reported not to wash their wounds and applying traditional herbal concoctions before 25 presenting at health facilities. Such divergence from the recommended standards has negative 26 implications on the effectiveness of post-exposure prophylaxis that is given when victims report 27 to health facilities. Our study investigated the epidemiology of dog bites and preclinical practices 28 for the victims in the context of dog bite prevention and rabies prevention respectively. We call 29 for targeted health education programs to improve pre-clinical behavior, regulation of herbalist 30 activities and interventions that minimize human-dog interactions. immediately before seeking medical 31 2 care. However, such pre-clinical 32 Abstract 33 In rabies endemic areas, appropriate management of dog bites is critical in human rabies 34 prevention. Victims must wash bite wounds for 15 minutes with soap, water and disinfectant 35 immediately before seeking medical care. This study investigated the epidemiology of dog bites 36 and determinants of compliance to these pre-clinical guidelines requirements among dog bite 37 victims from high rabies-burden areas of Wakiso and Kampala, Uganda. An explanatory 38 sequential mixed-methods study design was used. Quantitative data were collected from 376 dog- 39 bite patients at two healthcare facilities. Qualitative data as also collected through 13 in-depth 40 interviews with patients, healthcare workers, herbalists and veterinarians. Qualitative data were 41 analyzed using a deductive thematic approach. Generalized lineal models were used to determine 42 factors associated with compliance. Nearly half (190, 51%) of the patients were from Wakiso 43 District and 293 (77.9%) had grade II wounds. Most of the wounds (171, 45.5%) were on the legs. 44 Two-thirds of the bites occurred in public places. Only 70 (19%) of the bite patients had complied 45 with pre-clinical guidelines. Nearly half of the patients had applied un-recommended substances 46 such as herbs (47/193), antiseptics (46/193), “black stone” (25/193) and unknown creams (10/193) 47 on the wounds. Factors negatively associated with compliance included: being aged 15 years or 48 older, adjPR = 0.70 (0.47 - 0.92) and knowing the dog owner, adjPR=0.65 (0.36 - 0.93). However, 49 attainment of secondary or higher education, adjPR= 1.76 (1.24 – 3.79), being in employment, 50 adjPR = 1.48 (1.09 – 2.31), perception that the dog was sick, adjPR = 1.47 (1.02 – 2.72) and 51 knowledge about the dog’s subsequent victim(s) adjPR=0.35 (0.17 - 0.70) were positively 52 associated with compliance. High occurrence of dog bites in public places by free-roaming dogs 53 suggests the need for promotion of responsible dog ownership. Additionally, targeted health 54 education may be required to improve the low compliance to pre-clinical guidelines. 3 55 Introduction 56 Rabies, a neglected tropical disease, is estimated to cause 59,000 human deaths, over 3.7 million 57 disability-adjusted life years (DALYs) and USD 8.6 billion in economic losses worldwide annually 58 [1]. Although the rabies virus can infect all mammals, over 99% of all human rabies cases are 59 transmitted through dog bites [2]. Consequently, in addition to mass dog vaccination that breaks 60 rabies transmission cycles, strategies for prevention of rabies in humans include prevention of dog 61 bites and appropriate post-exposure treatment (PET) [2, 3]. 62 World Health Organization (WHO) has developed guidelines for dog bite victims before 63 presenting to a healthcare facility (preclinical management) as well as how the cases must be 64 managed in the healthcare facility (clinical guidelines) [4]. These preclinical guidelines are 65 summarized as: wash the bite wound with running water for 15 minutes; disinfect the wound with 66 substances with capacity to kill the rabies virus (soap, disinfectant); and seek medical care 67 immediately to receive post-exposure prophylaxis vaccines. Appropriate washing and disinfection 68 of wounds can prevent one-third of rabies infections [3, 5]. Inadequate dog bite wound care has 69 been associated with increased likelihood of PET failure and progression to rabies [5, 6]. 70 Pre-clinical practices that deviate from recommendations have been reported to include not 71 seeking medical care following dog-bites [7], delay in seeking treatment [3, 8, 9], lack of wound 72 washing or treatment of wounds with chilies, salt, turmeric powder, lime, snuff powder, paste of 73 leaves, acid and ash provided by traditional healers and magicians [10, 11]. Non-compliance to the 74 preclinical guidelines has been attributed to geographical, social, economic, cultural, 75 organizational, dog and wound factors [12-15]. 4 76 In Uganda, there are an estimated over 30,000 animal bites reported to healthcare facilities 77 annually and the burden keeps on rising despite ongoing interventions like health education [16]. 78 The country had approximately 486 suspected human rabies deaths between 2001 and 2015 [17] 79 although some authors have estimated the per capita annual death rate from rabies to be at 80 0.39/100,000 [18]. Despite such a high burden of bites and rabies, the reports of delays in seeking 81 medical care and victims treating dog bite wounds with traditional herbal concoctions remain 82 largely anecdotal. There is barely any published data on pre-clinical management of dog-bites in 83 Uganda. In this study, we investigated the epidemiology and preclinical management of dog bites 84 in Wakiso and Kampala districts, Uganda. We include data on circumstances of dog bites and what 85 influences people’s responses dog-bites with a view of identifying opportunities for prevention of 86 dog bites and rabies. 87 88 Methods 89 Study design and area 90 We used an explanatory sequential study with a mixed methods approach. This included collection 91 and analysis of quantitative data followed by collection and analysis of qualitative data. The study 92 was conducted in two referral healthcare facilities; Mulago National Referral Hospital (Kampala 93 City Authority) and Entebbe General Referral Hospital (Wakiso district) in Uganda. Both facilities 94 routinely provide dog bite post-exposure treatment in the two rabies endemic districts. 95 Approximately 14% and 8% of households own an average number of 1.9 and 1.7 dogs per 96 household in Wakiso and Kampala respectively [19]. From the perspective of interventions, the 5 97 districts have the highest number of dogs vaccinated against rabies [20]. Details of the study sites 98 are shown in S1 Fig. 99 100 Study population and data collection 101 Quantitative data 102 All patients presenting with dog bites at the two study health facilities for first-time PET between 103 April 2019 and October 2019 were enrolled upon providing informed consent to enroll in the study. 104 Based on severity of the wounds, patients were classified in one of 3 categories: Category I 105 (unbroken skin); Category II (superficial scratches without bleeding) and Category III (bites / 106 scratches which penetrate the skin with bleeding). Patients with category I bite exposure (44/420, 107 11%) who were assessed as not requiring PET, were excluded from the study. Quantitative data 108 including pre-clinical practices, socio-demographic factors, patient and biting dog factors and 109 circumstances surrounding the bite incidence were collected as shown in S1 Table. All data 110 collection tools were in English and translated into Luganda languages. Pre-testing of the 111 questionnaires was completed on animal bite patients in Mukono Health Center IV, Mukono 112 district, Uganda. 113 Qualitative data 114 In-depth interview (IDI) guides were used to collect qualitative data on dog bite circumstances and 115 preclinical practices. In total, 13 IDIs were conducted with 7 patients, 3 health care workers, 1 116 herbalist and 2 local veterinary officers to understand different perspectives of health seeking by 117 dog bite victims. Selection of patients for in-depth interviews was purposeful and based on their 118 reported outstanding compliance or non-compliance to preclinical guidelines. This approach is 119 generally used for collecting qualitative data [21]. IDIs were recorded using a digital audio 6 120 recorder device (SONY ICD PX333 Digital Voice Recorder®). Key points brought up during 121 interviews were also written down. Data saturation was determined to have been reached when no 122 new or / and relevant information materialized from the additional interviews conducted. 123 Data Management and analysis 124 Quantitative data 125 The outcome variable, “compliance” was recorded and categorized as “compliant” (if the patient 126 had washed wound with water and soap in addition to seeking medical care within 24 hours) and 127 “non-compliant” (if one of the former was missed or patient applied non-recommended substances 128 to the wound). 129 Data were double-entered by independent data entrants into Epi-info version 7.1.4.0, cleaned and 130 exported to STATA14 (StataCorp.; College Station, TX, USA) for analysis. Exploratory data 131 analyses were conducted and generated descriptive statistics for the continuous and categorical 132 variables. Median (IQR) for the continuous variables and percentages for categorical variables 133 were computed. Compliance to the pre-clinical guidelines was coded as 1 if patient was 134 “compliant” and 0 if “non-compliant”, to form a binary outcome variable. In the bivariate analyses, 135 categorical variables of importance were tabulated against compliance. The association was based 136 on chi-square and determined to be statistically significant if p < 0.05. In the multivariable analysis, 137 prevalence ratios (PRs) were computed using a generalized linear model (GLM) analysis with 138 Poisson family and a log link with robust standard errors. The model included variables with p < 139 0.25 at bivariate analysis or variables found to be potential or known to be associated with the 140 outcome from the literature. Both the unadjusted and adjusted prevalence ratios and corresponding 141 95% confidence intervals are presented. 7 142 Qualitative data 143 Independent individuals transcribed the recorded data into written text. Each transcript was given 144 to the respective data collectors to verify the transcripts. NVivo 11.4.1® software (QSR 145 International, 2017) was used to organize these data according to pre-set categories. The transcripts 146 were reviewed to identify the information that is related to the pre-set categories and themes were 147 developed. Under each theme, the information was inductively coded into sub-themes and then 148 patterns identified to form the explanatory points of what is being observed. Key statements 149 corresponding to the themes were presented together with quantitative findings. 150 Ethical considerations 151 The study protocol was approved by University of Nairobi - Kenyatta National Hospital Ethics 152 Review Committee (Kenya) REF: P687/09/2018; Mulago National Referral Hospital Research and 153 Ethics Committee (Uganda) REF: MREC 1518; and the Uganda National Council of Science and 154 Technology (Uganda) REF: SS4911. Written permission was obtained from hospitals before 155 commencement of the study. Informed assent was obtained from participants as well as caretakers 156 of minors prior to the study. For minors, assent was obtained after giving them an explanation of 157 study purpose, procedure and their rights. All data were anonymized and handled confidentially. 158 159 Results 160 The total number of dog-bite patients enrolled in the study was 376. Of these, 201(54%) were 161 male, and the median (IQR) age was 18 (22.75)18 years. One hundred and ninety (50.5%) of the 162 patients were from Wakiso district. Eleven percent of the bite-patients reported to own at least one 163 dog while only 5.1% had ever been vaccinated against rabies. Nearly three-quarters (72%) had 8 164 ever received some information about dogs and dog bites from sources including friends (46%), 165 family (14%), school (10%), and books (4%). Some victims (8%) reported to have suffered dog- 166 bites previously. A summary of the socio-demographic characteristics of the dog-bite patients, 167 dog-ownership and sources of information on dog-bites for the study participants is provided in 168 Table 1. 169 Table 1: Characteristics of the 376 dog bite study participants stratified by district of bite event. Characteristics Sex Male Female Age ≤15 years ˃15 years Hospital Entebbe (Wakiso) Mulago (Kampala) Religion Christian Non-Christian Marital status Not in union In union Highest education level No formal education Primary Secondary and above Household size ≤4 5-8 ≤9 Teens at home No Yes Employment status No Yes Current dog ownership No Yes Immunised against rabies No Frequency Wakiso N=190 (50.5%) Kampala N=186 (49.5%) 201 (53.5) 175 (46.5) 97 (51.1) 93 (48.9) 104 (55.9) 82 (44.1) 0.345 173 (46.0) 203 (54.0) 85 (44.7) 105 (55.3) 88 (47.3) 98 (52.7) 0.616 110 (29.3) 266 (70.7) 72 (37.9) 118 (62.1) 38 (20.4) 148 (79.6) ≤0.001 301 (80.1) 75 (19.9) 159 (83.7) 31 (16.3) 142 (76.3) 44 (23.7) 0.145 285 (75.8) 91 (24.2) 137 (72.1) 53 (27.9) 148 (79.6) 38 (20.4) 0.091 52 (13.8) 160 (42.7) 163 (43.5) 26 (13.8) 76 (40.2) 87 (46.0) 26 (13.9) 84 (45.2) 76 (40.9) 0.572 176 (46.7) 161 (44.6) 24 (6.7) 81 (45.3) 84 (46.9) 14 (7.8) 95 (52.2) 77 (42.3) 10 (5.5) 0.357 188 (50.0) 188 (50.0) 97 (51.1) 93 (48.9) 91 (48.9) 95 (51.1) 0.680 181 (48.1) 195 (51.9) 89 (46.8) 101 (53.2) 92 (49.5) 94 (50.5) 0.611 334 (88.8) 42 (11.2) 165 (86.8) 169 (90.9) 25 (13.2) 17 (9.1) 0.216 357 (94.9) 183 (96.3) 174 (93.6) 9 p-value Yes Get dog information No Yes Socio-economic status Lower Middle Upper Believed a dog could bite them No Yes 19 (5.1) 7 (3.7) 12 (6.4) 0.221 105 (27.9) 271 (72.1) 50 (26.3) 140 (73.7) 55 (29.6) 131 (70.4) 0.482 197 (52.5) 62 (16.5) 116 (31.0) 95 (50.2) 33 (17.5) 61 (32.3) 102 (54.8) 29 (15.6) 55 (29.6) 0.673 313 (83.2) 63 (16.8) 150 (78.9) 40 (21.1) 163 (87.6) 23 (12.4) 0.024 170 171 Characteristics of dog bite injuries 172 Nearly two-thirds of the dog bite wounds (239/376, 63.7%) were single bites. Three-quarters 173 (293/376, 77.9%) of the wounds were grade II and the rest were grade III. Forty-six percent of the 174 bite patients had wounds on their legs, 14% on the head, 3% on the face and 3% several bite sites. 175 The dog-bite distribution by body part and age of bite-patients are summarized in Table 2. 176 Table 2: Age-specific dog bite distribution by body part among the 376 participants Age (yrs) ≤15 years Percentage ˃15 years Percentage Total Percentage 177 Leg 62 35.8 109 53.7 171 45.5 Thigh 31 17.9 38 18.7 69 18.4 Arm 17 9.8 7 3.5 24 6.4 Abdomen 3 1.7 0 0.0 3 0.8 Back 16 9.3 10 4.9 26 6.9 Head 25 14.5 29 14.3 54 14.4 Face 7 4.1 4 1.9 11 2.9 Other 3 1.7 3 1.5 6 1.6 Combination 9 5.2 3 1.5 12 3.2 178 Characteristics of the biting dogs 179 Seventeen percent of the dog-bite patients had been bitten by their own dogs while 46% of the 180 victims knew the owner of the dog that bit them. Nearly a third (30%) of the bite patients could 181 identify the offending dog. Of these 113 biting dogs known to the victim, 21% had been vaccinated 182 against rabies, 26% had not been vaccinated, and 53% were of unknown vaccination status. The 10 Total 173 100.0 203 100.0 376 100.0 183 victims described the dog as being male in 35% of the cases, 19% female and the rest were of 184 unknown sex. The details on the characteristics of the biting dogs are presented in S2 Table. 185 Circumstances of dog bites 186 Most of the dog bites (166/376, 44.2%) occurred in the afternoons (12 noon – 6pm) and the least 187 (58/376, 15.4%) happened at night (7pm – 5am). Majority of the bites (339, 90%) were 188 unprovoked and 137 (37%) of the bites occurred when the persons bitten were walking on the 189 road. Nearly all the biting dogs (324, 86%) were unrestrained without a leash. Table 3 summarizes 190 data on circumstances surrounding the bites as reported by the bite patients. 191 192 Table 3: Circumstances of dog bite events among the 376 dog bite victims seeking post-exposure prophylaxis in the 2 selected hospitals in Uganda. Circumstances /contextual factor What time of day did the dog bite event happen? Morning Afternoon Evening / night Was it raining when the dog bite event happened? No Yes If the dog bite happened at night, was there a visible moon? No Yes Was the owner around when the bite happened? No Yes Did victim previously know the biting dog? No Yes Where did the event happen (place of event)? Own home Premises of person known to victim Premises of person not known to victim On the road Other (e.g. market, classroom) Was the victim in company of other people when dog bite occurred? No Yes What was the victim doing just before the dog bite? Walking 11 Frequency (n) Percentage (%) 152 166 58 40.4 44.2 15.4 347 29 92.3 7.7 27 31 46.5 53.5 255 121 67.8 32.2 262 113 69.9 30.1 124 86 4 137 25 33.0 22.9 1.1 36.4 6.6 211 165 56.1 43.9 209 55.6 Seated Chasing dog away Feeding dog Other Was it the victim that approached the biting dog? No Yes Was the biting dog on the leash? No Yes Did the victim attempt to fend off the biting dog? No Yes Did the victim think or feel that the dog intended to bite them? No Yes Does the victim blame anyone for the bite? No Yes What immediate action was taken against biting dog? Chased it away Killed it Nothing Ran away by itself Other 46 8 8 105 12.2 2.1 2.1 27.9 37 339 9.8 90.2 324 52 86.2 13.8 218 158 58.0 42.0 124 252 33.0 67.0 286 90 76.1 23.9 91 19 177 83 6 24.1 5.1 47.1 22.1 1.6 193 194 Circumstances of dog bites 195 Routine activities bringing dogs and humans into close proximity 196 Additional insights into dog bite circumstances were grouped as shown in S3 Table. A common 197 view was that victims were bitten while undertaking everyday activities. Respondents spoke about 198 holding something that drew the interest of the dog. Additionally, they talked about activities that 199 brought dogs into close proximity with people as some explained: 200 “On my way back from the abattoir to buy meat, I didn’t know that there is a dog nearby, 201 I only realized when it was holding onto my leg…… the dog continued biting me until a 202 man came and hit it. By this time, it had even bitten my buttocks.” (Adult patient, female). 12 203 “We were playing with other children, running in circles in the compound. Our dog joined 204 us and we ran with it. When I stopped, it jumped and bit me without warning.” (Patient, 205 male child). 206 Disturbing dogs and threatening owner 207 However, some respondents explained that the biting dog had been deliberately disturbed either 208 by themselves or by others. In addition, some thought that dogs also bit them when they acted in 209 a way that threatened the dogs’ masters. Notably, such dogs had been on the loose in presence of 210 strangers. One of the participants explained as follows; 211 “That Saturday morning, I went to visit my friend. We talked right in the compound, 212 standing. However, when we laughed loud, I remember the dog barked. When we gave 213 each other a ‘high-five’ and hugged, all I remember is the owner shouting at the dog to let 214 go of my shirt. In the struggle, it bit me two times on the back and leg.” (Male adult patient). 215 Unusual behavior and protective tendencies 216 Some dog owners explained unusual behavior of the dog e.g. biting every living thing in the 217 homestead, whether it posed a danger to it or not. They interpreted this as potentially rabid 218 behavior. In addition, others were bitten by dogs protecting each other in a pack or with young 219 ones as one explains; 220 “…. since our dog produced it does not want to interact with us. It no longer sits in front 221 of the kitchen door as it used to do. I was with this boy in the kitchen and when I left to go 222 to the house, he says he went behind the kitchen to see the dog and its babies. He said that 13 223 is when it jumped and bit him on the shoulder. When I checked on the dog, it also wanted 224 to bite me.” (caretaker / mother to a child patient). 225 Deviant handling practices 226 A number of respondents bitten by their own dogs explained circumstances that pointed to 227 deviation from routine practices of handling dogs. They tended to inflict pain on the dogs during 228 handling. In retaliation, the dogs bit them as one of them elaborates; 229 “Normally, I call them to follow me to their kernel and they do. But this time one of them 230 refused and after taking in others, I went back and dragged it by the front leg. When it 231 resisted, I lifted it and tried to push it into the house. This is when it bit my hand…...” 232 (Adult male patient). 233 Seasons 234 For some, there were conditions like rain that caused interaction with the dog in open shelters. 235 However, some described circumstances of having been bitten by dogs left unattended to, even 236 without sharing shelter with them. On the other hand, some practitioners described bites as a 237 seasonal issue linking them to late night activities especially during festive days as one explains; 238 “I get most of the people during big days [festive] like Christmas and Easter. This is when 239 my house [serves as the care facility] is always full. Do you know why? People drink yet 240 most of the dogs without owners also move at night. So they meet themselves and in most 241 cases people harass these dogs first because they are scared of them. This is when they get 242 bitten and come here for treatment.” (Herbalist attending to dog bite victims). 14 243 Immediate actions taken by dog bite victims 244 Reporting to local leaders and area veterinarians 245 When we inquired into what victims did immediately after the bite, the key actions included 246 seeking medical care and legal action as summarized in S3 Table. Reporting to local authorities 247 was quite common especially when victims wanted local leaders to put owners of the biting dog 248 to task of owning up the responsibility of treating them. However, local veterinarians explained 249 that some victims immediately call them because they know that it is their responsibility to ensure 250 that dogs do not bite them. In other circumstances, the victims call veterinarians to seek treatment 251 advice as one explains; 252 “They can call to be advised, others rush to the nearest health center and that is where 253 they refer them to Entebbe hospital…... Many of them ask if my office has anti-rabies 254 vaccines thinking such vaccines are kept with the area vet. They even get annoyed when I 255 tell them I don’t have the vaccine.” (Local veterinarian). 256 Presenting to healthcare facility 257 Notably, there are some who immediately went to a healthcare facility. In comparison, some 258 victims spent time regretting and filled with fear of bite consequences, especially death. Those that 259 experienced this state related to the previous events that they had heard or witnessed in their lives 260 as one narrates below; 261 “I cried, I just sat there and cried. I thought I was going to die. In our place [of origin], a 262 dog bit a man and after 3 months he started barking like a dog, yes. All my thoughts ran to 263 that man who died thinking like I was going to be like him. Besides, I was also in too much 264 pain. You see this finger, I still feel paralysis and sharp pain in it.” (Adult female patient). 15 265 Compliance to pre-clinical guidelines 266 Only 70 participants (19%) complied to the guidelines and reported that they washed the wounds 267 with water and soap and presented to a healthcare facility within 48 hours. Of these, 45% (32/70) 268 applied an antiseptic in addition to washing. However, 19/376 (5%) washed with water only and 269 183/376 (48.7%) neither washed the wound not applied anything. Overall, the commonest material 270 applied on the wound by the 193 victims conducting pre-clinical care were antiseptic (46), herbs 271 (25) black stone (10) unknown creams or other materials such as beans, dog urine, dust, tobacco, 272 coins, brake fluid, acid, powder made out of dog hair and salt. Notably only 8 out of 29 study 273 participants who have had previous dog bite episode complied with pre-clinical guidelines. 274 Presentation within 48 hours was mentioned by three-quarters (74.7%) of the victims. The median 275 (IQR) time to presentation at a health facility was 18 (41) hours. Table 4 shows that compliance 276 differed by education status (p<0.001), employment status (p = 0.01) and accessing information 277 about dogs (p = 0.005). 278 Table 4: Distribution of selected characteristics of 376 respondents by compliance Characteristics District Wakiso Kampala Sex Male Female Age ≤15 years ˃15 years Religion Christian Non-Christian Marital status Not in union In union Highest education level No formal education Primary Frequency, n (%) Comply, n (%) p-value 190 (50.5) 186 (49.5) 38 (20.0) 32 (17.2) 0.486 201 (53.5) 175 (46.5) 34 (19.9) 36 (20.6) 0.364 173 (46.0) 203 (54.0) 36 (20.8) 34 (16.8) 0.313 301 (80.1) 75 (19.9) 54 (17.9) 16 (21.3) 0.499 285 (75.8) 91 (24.2) 56 (19.7) 14 (15.4) 0.363 52 (13.8) 160 (42.7) 7 (13.5) 15 (9.4) 16 Secondary and above Household size ≤4 5-8 ≤9 Employment status No Yes Current dog ownership No Yes Patient vaccinated against rabies No Yes Get dog information No Yes Socio-economic status Lower Middle Upper Dog looked sick No Yes Don’t know Exhibited fear of people No Yes Don’t know Vaccination status No Yes Don’t know Bitten someone after No Yes Don’t know Dog owner known No Yes 163 (43.5) 48 (29.5) <0.001* 176 (46.7) 161 (44.6) 24 (6.7) 30 (17.1) 35 (21.7) 4 (16.7) 0.541 181 (48.1) 195 (51.9) 24 (13.3) 46 (23.6) 0.010* 334 (88.8) 42 (11.2) 66 (19.8) 4 (9.5) 0.140 357 (94.9) 19 (5.1) 64 (17.9) 6 (31.6) 0.136 105 (27.9) 271 (72.1) 10 (9.5) 60 (22.1) 0.005* 197 (52.5) 62 (16.5) 116 (31.0) 27 (13.7) 21 (33.9) 22 (18.9) 0.002* 250 (66.5) 73 (19.4) 53 (14.1) 25 (10.0) 35 (48.0) 10 (18.9) <0.001* 253 (67.3) 102 (27.1) 21 (5.6) 24 (9.5) 36 (35.3) 10 (47.6) <0.001 * 50 (13.3) 41 (10.9) 285 (75.8) 7 (14) 5 (12.2) 58 (20.4) 0.303 104 (27.7) 76 (20.2) 196 (52.1) 19 (18.3) 39 (51.3) 12 (6.1) <0.001* 201 (53.5) 175 (46.5) 51 (25.4) 19 (10.9) <0.001* 279 *Significance at p≤0.05 280 Explanations for application of non-recommended substances 281 To kill micro-organisms 17 282 On deeper inquiry, some respondents thought that by applying substances of unusual pH or 283 temperature, they would kill the rabies virus. This came out as one of the reasons why some applied 284 hot water, salt and battery acid as one explains; 285 “When the dog came and bit me, many of my colleagues in the garage where I work told 286 me to first put battery acid to kill the germs [virus] that cause dog madness [rabies] before 287 they could go very far inside the meat [flesh]. So, they removed the battery from the car 288 and drained its acid into the wound here [shows hand].” (Male, adult patient). 289 Routine management of wounds 290 Some respondents had witnessed routine wound management with certain substances or by some 291 procedures. It was the reason they managed the dog bite in a similar way without the specifics of 292 a dog bite as one explains: 293 “At times you find people with a bandage. When you ask them why, they tell you they do 294 not want the blood to move to the brain carrying dog poison. They think rabies is like snake 295 poison that travels in the bloodstream.” (local veterinarian). 296 Knowledgeable caretakers and trust in herbalist 297 Additionally, some victims did not apply herbs out of choice but relied on the knowledge, skills 298 and practices of first responders who they thought were more knowledgeable in managing dog 299 bite. This was more pronounced when the caretaker also doubled as the decision-maker on which 300 line of treatment to take. Similarly, a number of respondents applied herbs because they trusted 301 the herbalist. This trust extended to the treatment which they took without questioning as one 302 recounts: 18 303 “My mother sent me to the traditional doctor [herbalist]. There is some powdered medicine 304 he tells you to put under the tongue then he cuts you on the leg here like this [shows around 305 the ankle] then he puts black stone…... He told me to go home and not to bathe using cold 306 water drink cold drinks ……... I did not ask, I just followed instructions, it is my mother 307 who had sent me to him”. (Female adult patient). 308 Pedigree of herbalist 309 The pedigree of a particular herbalist also played a key role in informing the decisions of victims. 310 Some respondents based their decisions on success stories they had heard as one herbalist explains: 311 “they come because I have a history of healing them since the 70s. Even when they go to Mulago 312 [hospital], some pass here. People believe in me. My treatment is cheap because over time, I have 313 found out that dogs bite the poor. They should thank God, not me, for He has kept me longer.” 314 (Herbalist for dog bite victims). 315 Perceived high cost of conventional treatment 316 However, some patients sought herbalist assistance because they thought they could not afford 317 conventional treatment. These only went to hospital when they learnt that treatment was free as 318 one elaborates; 319 “I sent my girl [daughter] to the herbalist, and I did not go because I did not have money 320 for both of us. I first felt much pity for this young one [smiles]. Me I stayed and put tobacco. 321 But when the dog had died, I was worried, I went to Mulago [hospital] after a week where 322 I learnt that the treatment was free. I went back home and brought my daughter too. She 323 didn't go back to the herbalist again.” (Adult female patient and mother to a patient). 324 Conflicting information on efficacy of both herbs and modern treatment 19 325 When we investigated why some of the patients used conventional and non-conventional medicine 326 at the same time, they pointed to information from fellow patients they found in hospital. Another 327 reason they gave for simultaneous resort was the conflicting information proving and disproving 328 efficacy of herbs. Therefore, they chose to use two lines as one elaborates: 329 “I went to the herbalist because our family knew very well that he works well on dog 330 bites…...one of my daughters healed well, so I was sure that his medicine [herbs] heal those 331 bitten by dogs. But when our LC [local leader] told me that in Mulago treatment was more 332 effective and free, I also decided to come this side [hospital].” (Adult female patient). 333 Explanations for seeking medical care from hospital 334 Mistrust in herbalists 335 Some patients talked about the mistrust they had in herbalists, even when some of them patients 336 first went to them. They indicated dissatisfaction with the herbalist’s procedures. Some of them 337 deliberately refused the processes and left for hospital without applying any herbs as one narrates; 338 “Now to go to Mulago [hospital], it has professional doctors but the one they had directed 339 me to is a herbalist. He even wanted to cut my leg and put black stone. He did not wear 340 gloves, so I refused. That is why I stopped him from him adding more things on my wound. 341 I went away” (Adult male patient). 342 Knowledge and experiences on dangers of dog bites 343 Knowledge of someone who had suffered negative consequences of dog bites attributed to 344 inadequate medical care also came out as one of the reasons why some people immediately went 345 to hospital. Such experiences were common among the victims as one recounts; 20 346 “People talk. There is also a time we were in Kikandwa [place of birth] and a child passed 347 on. A dog bit him and he was taken to a [herbalist] and received treatment. After a period 348 of some months that I can’t recall, a child started barking and passed on. This was last 349 year. So I could not risk going to that man [herbalist].” (Female adult patient). 350 Community advice 351 However, other respondents attributed their action of seeking medical care paradoxically to both 352 mistrust and trust in community advice. Those who mistrusted community advice questioned the 353 efficacy of different applications that were suggested to them. However, those who trusted 354 community advice heeded and went to the hospital. 355 Factors associated with compliance to standard preclinical management 356 guidelines for victims seeking post-exposure prophylaxis 357 In the adjusted analysis, factors significantly associated with higher likelihood of compliance to 358 pre-clinical guidelines were having a formal education (adjPR = 1.76, 95% CI: 1.24 – 3.79, p= 359 0.024), being in employment (adjPR = 1.48, 95% CI: 1.09 – 2.31, p = 0.047), perceiving the dog 360 as being sickly (adjPR = 1.47, 95% CI: 1.02 – 2.72, p = 0.042) and knowing that the dog went on 361 to bite another person (adjPR = 1.69, 95% CI: 1.01 – 2.86, p = 0.048). Lower likelihood of 362 compliance was associated with being older than 15 years of age (adjPR = 0.70, 95% CI: 0.47 - 363 0.92, p = 0.045), not being certain whether the dog went to bite another person or not (adjPR = 364 0.35, 95% CI: 0.17 - 0.70, p = 0.003) and knowing the owner of the biting dog (adjPR = 0.65, 95% 365 CI: 0.36 - 0.93, p = 0.034). Important to note is that sex and rabies immunization status of the 366 victim did not have any bearing on the compliance as shown in Table 5. Notably, the interaction 367 effects between sex and age as well as sex and marital status on compliance were not significant. 21 368 369 370 Table 5: Multivariable analysis of factors associated with compliance to standard preclinical management guidelines for 376 victims seeking post-exposure prophylaxis in the 2 selected hospitals in Uganda. Characteristics District Wakiso Kampala Sex Male Female Age ≤15 years ˃15 years Religion Christian Non-Christian Marital status Not in union In union Highest education level No formal education Primary Secondary and above Employment status No Yes Current dog ownership No Yes Immunised against rabies No Yes Get dog information No Yes Socio-economic status Lower Middle Upper Perceived health status of dog Healthy Sickly Don’t know Exhibited fear of people No Yes Don’t know Unadjusted PR (95% CI) p-value Adjusted PR (95% CI) p-value 1 0.86 (0.56 - 1.32) 0.488 1 1.22 (0.79 – 1.86) 0.365 1 1.04 (0.73 – 1.49) 0.798 1 0.81 (0.53 - 1.23) 0.315 1 0.70 (0.47 - 0.92) 0.045* 1 1.19 (0.72- 1.96) 0.495 1 0.74 (0.39 - 1.41) 0.364 1 0.70 (0.30 - 1.62) 2.19 (1.05 – 4.54) 0.400 0.036 1 0.89 (0.81 – 2.05) 1.76 (1.24 – 3.79) 0.783 0.024* 1 1.78 (1.13 – 2.79) 0.012 1.48 (1.09 – 2.31) 0.047* 1 0.48 (0.18 - 1.25) 0.135 1 1.76 (0.88 – 3.54) 0.112 1 1.48 (0.81 - 2.74) 0.203 1 2.32 (1.23 – 4.37) 0.009 1 1.40 (0.74 - 2.66) 0.295 1 2.47 (1.51 – 4.05) 1.38 (0.83 - 2.31) <0.001 0.216 1.29 (0.82 - 2.05) 1.01 (0.63 – 1.62) 0.269 0.292 1 4.79 (3.08 - 7.46) 1.89 (0.96 - 3.69) <0.001 0.064 1.47 (1.02 – 2.72) 1.29 (0.63 – 2.45) 0.042* 0.430 1 3.72 (2.34 - 5.91) 5.01 (2.79 – 9.05) <0.001 <0.001 1.53 (0.88 – 2.67) 1.52 (0.79 – 2.91) 0.132 0.931 22 371 Rabies vaccination status of dog No 1 1 Yes 0.87 (0.30 - 2.54) 0.801 0.72 (0.28 - 1.83) Don’t know 1.45 (0.70 - 3.00) 0.312 0.96 (0.38 – 2.45) Bitten someone after No 1 Yes 2.81 (1.77 – 4.46) <0.001 1.69 (1.01 – 2.86) Don’t know / not certain 0.34 (0.17 - 0.66) 0.002 0.35 (0.17 - 0.70) Dog owner known No 1 Yes 0.43 (0.26 - 0.69) 0.001 0.65 (0.36 - 0.93) *Significance at p-value ≤0.05 and 95% confidence interval (95% CI). 0.491 0.931 0.048* 0.003* 0.034* 372 373 Discussion 374 The study investigated the epidemiology of dog bites and preclinical practices for the victims in 375 the context of dog bite prevention and rabies prevention respectively. The finding that there were 376 more males than females is in concurrence with majority of studies that have reported a 377 preponderance for males [9, 22]. Some authors attribute this to personality variation between 378 genders with more males being subject to dog bites [23]. Others have attributes it to males being 379 frequently involved in day and night activities [9]. However, our findings contradict some studies 380 which reported that females are more likely to be bitten [24]. 381 Regardless of age, the leg was the most affected part of the body followed by hands and arms. 382 Limbs have been documented to be the most bitten parts [25-27]. This may be attributed to 383 accessibility, especially for the legs, and the struggles that usually ensue during the bite. Such 384 scuffles usually involve use of arms and legs to ward off the dog. However, bites on the head were 385 among children only and this may be explained by their height which puts the head near the mouth 23 386 of the dog. Likewise, some authors have attributed this to the small physique of children, their 387 inclination to put their faces close to animals, and limited motor skills to provide defense [22]. 388 Majority of wounds were Category II involving skin scratches. This is expected especially when 389 the majority of wounds were singular in number and extremities were the most affected parts. 390 These parts are not only accessible by dogs but they are easily movable in self-defense. Given that 391 most of the victims were walking, it was unlikely that biting dogs got a firm grasp of the victim 392 before disentanglement. Besides, dog attacks usually last a very brief duration, which explains 393 why very severe and fatal bites are not a common finding in literature, just like in our study. Such 394 findings on severity are consistent with other studies [28] though they conflict with some [29]. 395 The owner of the biting dogs was not known in most cases (53.5%). This is perhaps because 396 majority of victims were bitten on the road or in public places like markets. Notably, the study 397 area is mostly urban, and characterized by rapid urbanization, high population of people, 398 abundance of garbage heaps that serve as a source food for dogs and un-owned animals especially 399 dogs. In addition, it might be due to some dog owners not chaining their dogs and leaving them to 400 wander hence posing a risk of bites to strangers. Some authors have attributed it to weak legislation 401 on responsible dog ownership [30]. Just like in our study, the increased risk of bite events by such 402 dogs compared to those with known owners has been reported in India [31] and Nigeria [26] 403 though in Mozambique [32], they play a minor role. This shows that the role of wandering dogs in 404 the bites may vary with each setting. 405 For majority of bites, it is the victims that approached the dogs rather than the other way round. 406 Territorial invasion easily forces dogs to bite out of self-defense. Such a risk increases when dogs 407 are in a pack or nursing young ones as explained in the in-depth interviews for our study. Studies 24 408 have widely reported increased dog aggression due to territorial invasion especially by children 409 [33, 34]. Our findings on this are consistent with that of related study in the United Kingdom which 410 reported 50% of the victims as having approached the biting dog [35]. 411 Before presentation to hospital, only 18.6% of the patients had complied with recommended 412 preclinical guidelines. The WHO recommends meticulously flushing of wounds with water and 413 soap and application of an antiseptic like povidone iodine if available [2]. The low level of 414 compliance in our study may be due to inadequate knowledge on the guidelines. Moreover, many 415 respondents expressed not knowing what to do immediately the dog bit them. However, our 416 prevalence is comparable with that reported in India which varied between 2 -21% depending on 417 the township [36]. Nonetheless, in India still, another study reported a higher rate (58%) than ours 418 though there was a significant rural-urban divide with the former performing worse [37]. This, 419 combined with an 7% - 45% prevalence of wound washing with soap and water in Kenya and 420 Ethiopia respectively [38, 39], is evidence of how the practice varies across communities. It may 421 also indicate variations in the coverage and uptake of health education interventions across 422 societies. 423 Of those who applied some substances before reporting for PET, only 23.8% applied an antiseptic 424 as recommended. A comparable proportion applied herbs whereas others used antibiotics, black 425 stone, charcoal, acid, powder made by burning hair of biting dog, split beans, paraffin, salt, 426 monetary coins and others. The inquiry into application of non-recommended material revealed 427 that such practices were driven by a number of factors including individual beliefs on efficacy, 428 lack of funds to pay for medical services as well community influences and advice. Such practices 429 have been reported elsewhere with higher magnitudes being reported in both community [36, 37] 430 and hospital based surveys [40]. 25 431 Victims who were bitten by a dog with known ownership were 35% less likely to comply. 432 Sometimes it is intuitive that a person bitten by a dog of known ownership might be more confident 433 with regard to the health status of the dog compared to a bite by a dog they do not know. If the 434 owner is known, it is easier to inquire about the health aspects of the dog like the rabies vaccination 435 status. However, this practice of victims assessing the risk of rabies to be low based on knowing 436 the dog ownership is dangerous and should be discouraged. Nonetheless, our findings are in 437 concurrence with another study in Ethiopia which found that the likelihood of the dog bite victim 438 visiting a healthcare facility more than doubled when the victim was bitten by a dog of unknown 439 ownership [41]. 440 Victims who were employed at the time of the bite were approximately one and half times more 441 likely to comply with the guidelines than those who were not. This may probably be due to the 442 fact that the employed tend to have higher education levels. Besides, employment has been 443 associated with appropriate health seeking behavior in some studies [42]. Similar, those who 444 perceived the biting dog as being sick were more than twice likely to comply compared to those 445 who perceived them as being healthy. It may be that victims attached the sickness perception to an 446 elevated risk for rabies and therefore complied to the pre-clinical preventive measures. Besides, 447 some studies have described the health status of a biting animal as a drive to PET compliance [43] 448 Participants who had attained at least secondary education or higher were more likely to comply 449 with pre-clinical guidelines compared to those with no formal education. People who are more 450 educated tend to have a higher ability of interpreting health education messages. Moreover, in our 451 study, those with secondary education or more were more likely to access information on dogs and 452 dog bites than those with less education level. Our findings and probable explanation are coherent 26 453 with research that has suggested that people with higher education tend to have more knowledge 454 about rabies than illiterate ones [44, 45]. 455 Patients aged above fifteen years were less likely to comply compared to those below 15 years. 456 this finding is in consistent with that of a study in China that found age ˃15 as being at more risk 457 of failure to begin PEP [46]. Prioritization of younger ones to receive healthcare was evident in 458 this study. When asked why the daughter was sent to receive treatment and the mother stayed home 459 yet they had both been bitten by the same dog, the mother explained that the young one had more 460 need for treatment. So, with limited resources and lack of knowledge that treatment was free, 461 priority seemed to be given to younger ones. 462 Patients who did not know whether the biting dog had gone on to bite other people were 65% less 463 likely to comply. Cases of a single dog being responsible for multiple bite cases have been widely 464 described [47, 48] and this is typical of wandering dogs. The finding that some people did not 465 know may be a reflection of the care-free attitude of such individuals towards the risk of bite 466 consequences. Not caring to find out whether the dog bit other people makes them not likely to 467 comply as they may not know the value in ascertaining the status of the dog. Nonetheless, deeper 468 inquiries revealed that some respondents did not comply even after knowing that the dog had gone 469 on to bite other people. However, they attributed this to lack of funds to seek treatment. 470 The main limitation of this study may be the self-reports about the events which might have 471 introduced recall bias through inaccuracies in detailing the events. However, we verified the 472 information where possible by triangulation. In addition, we used a hospital-based convenience 473 sample and this limits the representativeness of our results for the entire population of dog bite 474 victims. An example of this may be that there may be specific factors that influenced our 27 475 respondents to report to hospital but not those who stayed home and used domestic remedies. 476 Therefore, the findings should be interpreted within this context. 477 Conclusions 478 The study presents evidence to show that dog bites in the study area are widespread across gender 479 and age. The bites are both provoked and unprovoked and are majorly by wandering dogs in public 480 places like roads. Compliance to recommended pre-clinical guidelines is low and mainly due to 481 inadequate awareness about dangers of alternative treatments and availability of therapy. There is 482 therefore need for holistic targeted health education programs and regulation of herbalist activities. 483 In addition, approaches that reduce human-dog interactions in public places for example reduction 484 of stray dog populations, need emphasis. 485 Acknowledgements 486 We are grateful to the respondents who participated in this study and the staff of Mulago National 487 Referral Hospital and Entebbe General Referral Hospital for providing a conducive environment 488 for successful data collection. 489 This research was supported by the Consortium for Advanced Research Training in Africa 490 (CARTA). CARTA is jointly led by the African Population and Health Research Center and the 491 University of the Witwatersrand and funded by the Carnegie Corporation of New York (Grant No- 492 -B 8606.R02), Sida (Grant No:54100113), the DELTAS Africa Initiative (Grant No: 493 107768/Z/15/Z) and Deutscher Akademischer Austauschdienst (DAAD). The DELTAS Africa 494 Initiative is an independent funding scheme of the African Academy of Sciences (AAS)’s Alliance 495 for Accelerating Excellence in Science in Africa (AESA) and supported by the New Partnership 28 496 for Africa’s Development Planning and Coordinating Agency (NEPAD Agency) with funding 497 from the Wellcome Trust (UK) and the UK government. The statements made and views 498 expressed are solely the responsibility of the Fellow. 499 References 500 1. 501 Burden of Endemic Canine Rabies. PLOS Neglected Tropical Diseases. 2015;9(4):e0003709. doi: 502 10.1371/journal.pntd.0003709. 503 2. 504 Technical Report Geneva, Swtizerland: World Health Organization, 2013. 505 3. 506 of rabies post-exposure prophylaxis of animal bite victims in China: Prevalence and determinants. PLOS 507 Neglected Tropical Diseases. 2017;11(7):e0005663. doi: 10.1371/journal.pntd.0005663. 508 4. 509 Switzerland World Health Organisation,; 2012. 510 5. Wilde H. Failures of post-exposure rabies prophylaxis. Vaccine. 2007;25(44):7605-9. 511 6. Song M, Tang Q, Rayner S, Tao XY, Shen XX, Liang GD. Factors influencing the number of rabies 512 cases in children in China. Biomed Environ Sci. 2014;27(8):627-32. 513 7. 514 301. 515 8. 516 exposure prophylaxis and deaths in a region of endemic canine rabies. PLoS Negl Trop Dis. 517 2008;2(11):25. Hampson K, Coudeville L, Lembo T, Sambo M, Kieffer A, Attlan M, et al. Estimating the Global WHO. World Health Organization Expert Consultation on Rabies. World Health Organization Liu Q, Wang X, Liu B, Gong Y, Mkandawire N, Li W, et al. Improper wound treatment and delay WHO. Current WHO Guide for Rabies Pre and Post-exposure Treatment in Humans. Geneva, Gilchrist J., Sacks JJ., White D., Kresnow M. Dog bites: still a problem? . Inj Prev. 2008;14:296- Hampson K, Dobson A, Kaare M, Dushoff J, Magoto M, Sindoya E, et al. Rabies exposures, post- 29 518 9. Aghahowa SE, Ogbevoen RN. Incidence of dog bite and anti-rabies vaccine utilization in the, 519 University of Benin Teaching Hospital, Benin City, Nigeria: A 12-year assessment. Vaccine. 520 2010;28(30):4847-50. 521 10. 522 victims & management practices in a tertiary care institute in Mumbai, Maharashtra, India. The Indian 523 Journal of Medical Research. 2014;139(3):459-62. PubMed PMID: PMC4069742. 524 11. 525 Knowledge-Attitude and Practices in the Awash Basin, Eastern Ethiopia. PLOS Neglected Tropical 526 Diseases. 2016;10(2):e0004471. doi: 10.1371/journal.pntd.0004471. 527 12. 528 Living in the Same Home: A Bi-National Comparison. Frontiers in veterinary science. 2018;5:66-. doi: 529 10.3389/fvets.2018.00066. PubMed PMID: 29780810. 530 13. 531 in rabies endemic South Bhutan: a community-based contact-tracing survey. BMC Public Health. 532 2019;19(1):237. doi: 10.1186/s12889-019-6559-x. 533 14. 534 about Rabies Prevention and Control: A Community Survey in Tanzania. PLoS Negl Trop Dis. 2014;8. doi: 535 10.1371/journal.pntd.0003310. 536 15. 537 Himachal Pradesh, India: a qualitative study. International Journal Of Community Medicine And Public 538 Health. 2016;3(8):2064-9. 539 16. 540 Health, 2016. Gogtay NJ, Nagpal A, Mallad A, Patel K, Stimpson SJ, Belur A, et al. Demographics of animal bite Tschopp R, Bekele S, Aseffa A. Dog Demography, Animal Bite Management and Rabies Messam LLM, Kass PH, Chomel BB, Hart LA. Factors Associated With Bites to a Child From a Dog Penjor K, Tenzin T, Jamtsho RK. Determinants of health seeking behavior of animal bite victims Sambo MLT, Cleaveland S, Ferguson HM, Sikana L. Knowledge, Attitudes and Practices (KAP) Dhiman AK, Thakur A, Mazta SR. Treatment seeking behavior of the dog bite patients in MoH. Injuries: District Health Information Software (DHIS-2). Kampala, Uganda: Ministry of 30 541 17. Masiira B, Makumbi I, Matovu JKB, Ario AR, Nabukenya I, Kihembo C, et al. Long term trends and 542 spatial distribution of animal bite injuries and deaths due to human rabies infection in Uganda, 2001- 543 2015. PLoS One. 2018;13(8):e0198568. doi: 10.1371/journal.pone.0198568. 544 18. 545 doi: 10.1371/journal.pntd.0003709. 546 19. 547 2010. 548 20. 549 of poverty on dog ownership and access to canine rabies vaccination: results from a knowledge, 550 attitudes and practices survey, Uganda 2013. Infectious Diseases of Poverty. 2017;6(1):97. doi: 551 10.1186/s40249-017-0306-2. 552 21. 553 Med Educ. 2012;4(1):1-3. doi: 10.4300/JGME-D-11-00307.1. PubMed PMID: 23451297. 554 22. 555 Literature and Case Series. Int J Clin Pediatr Dent. 2017;10(1):67-72. 556 23. 557 sectional survey of prevalence, incidence and factors associated with dog bites in a UK community. 558 Journal of Epidemiology and Community Health. 2018. doi: 10.1136/jech-2017-209330. 559 24. 560 room of a teaching hospital South East Nigeria. African Journal of Medical and Health Sciences. 561 2018;17(1):35-40. doi: 10.4103/ajmhs.ajmhs_59_17. 562 25. 563 of animal bite injuries in Uganda and projections of the burden of rabies. Trop Med Int Health;10(8):790- 564 8. 2005. Hampson K. Estimating the global burden of endemic canine rabies. PLoS Negl Trop Dis. 2015;9. MAAIF, UBOS. The National Livestock Census Report 2008. Kampala: Government of Uganda, Wallace RM, Mehal J, Nakazawa Y, Recuenco S, Bakamutumaho B, Osinubi M, et al. The impact Sargeant J. Qualitative Research Part II: Participants, Analysis, and Quality Assurance. J Grad Agrawal A, Kumar P, Singhal R, Singh V, Bhagol A. Animal Bite Injuries in Children: Review of Westgarth C, Brooke M, Christley RM. How many people have been bitten by dogs? A cross- Omoke N, Chukwueloka ON. Incidence and pattern of dog bite injuries treated in the emergency Fèvre EM., Kaboyo RW., Persson V., Edelsten M., Coleman PG., Cleaveland S. The epidemiology 31 565 26. Abubakar SA, Bakari AG. Incidence of dog bite injuries and clinical rabies in a tertiary health care 566 institution: a 10-year retrospective study. Ann Afr Med. 2012;11(2):108-11. 567 27. 568 C8. 569 28. 570 2015;14(2):142-53. Epub 10/10. doi: 10.1007/s12663-013-0593-5. PubMed PMID: 26028828. 571 29. 572 Delhi. Indian J Commun Med. 2003;28(4):157-9p. 573 30. 574 KwaZulu-Natal, South Africa. PLOS Neglected Tropical Diseases. 2013;7(4):e2059. doi: 575 10.1371/journal.pntd.0002059. 576 31. 577 in a rural area of Davangere, Karnataka, India. 2017. 2017;3(1):4. Epub 2017-01-31. doi: 10.18203/2394- 578 6040.ijcmph20151556. 579 32. 580 features and risk factors for human rabies and animal bites during an outbreak of rabies in Maputo and 581 Matola cities, Mozambique, 2014: Implications for public health interventions for rabies control. PLOS 582 Neglected Tropical Diseases. 2017;11(7):e0005787. doi: 10.1371/journal.pntd.0005787. PubMed PMID: 583 PMC5542695. 584 33. 585 children in Trinidad. BMC Public Health. 2008;8(1):85. doi: 10.1186/1471-2458-8-85. 586 34. 587 10.1136/bmj.39105.659919.BE. PubMed PMID: PMC1804160. Mannion CJ, Graham A. Dog bite injuries in hospital practice. Br J Hosp Med. 2016;77(10):C165- Chhabra S, Chhabra N, Gaba S. Maxillofacial injuries due to animal bites. J Maxillofac Oral Surg. Khokhar A, Meena G, Mehra M. Profile of dog bite cases attending MCD dispensary at Alipur, Hergert M, Nel LH. Dog Bite Histories and Response to Incidents in Canine Rabies-Enzootic Anandara R., Balu PS. Compliance to anti rabies vaccine and animal bite management practices Salomão C, Nacima A, Cuamba L, Gujral L, Amiel O, Baltazar C, et al. Epidemiology, clinical Georges K, Adesiyun A. An investigation into the prevalence of dog bites to primary school Morgan M, Palmer J. Dog bites. BMJ : British Medical Journal. 2007;334(7590):413-7. doi: 32 588 35. Oxley JA, Christley R, Westgarth C. Contexts and consequences of dog bite incidents. Journal of 589 Veterinary Behavior. 2018;23:33-9. doi: https://doi.org/10.1016/j.jveb.2017.10.005. 590 36. 591 and its Management in the Context of Prevention of Rabies Among the Victims of Dog Bite Attending the 592 OPD Services of CHC Muradnagar. Journal of Family Medicine and Primary Care. 2014;3(4):355-8. doi: 593 10.4103/2249-4863.148107. PubMed PMID: PMC4311341. 594 37. 595 Delhi: A Community-based Study. Annals of Medical and Health Sciences Research. 2016;6(2):115-9. doi: 596 10.4103/2141-9248.181836. PubMed PMID: PMC4866364. 597 38. 598 the communities’ knowledge, attitude and practices of rabies incidence/human exposures: Cases of 599 selected settings in Ethiopia and Kenya. Ethiopian Journal of Health Development. 2018;32(1). 600 39. 601 Bite Victims Attending an Anti-rabies Health Center in Jimma Town, Ethiopia. PLOS Neglected Tropical 602 Diseases. 2015;9(6):e0003867. doi: 10.1371/journal.pntd.0003867. PubMed PMID: PMC4482645. 603 40. 604 Joshua Inubile, Adekunle Bamidele Taiwo, Oyinkansola Tolulope Agaja, et al. Pattern and outcome of 605 dog bite injuries among children in Ado-Ekiti, Southwest Nigeria. The Pan African Medical Journal 606 2017;27(81). doi: doi:10.11604/pamj.2017.27.81.7360. 607 41. 608 following rabies exposure in Ethiopia. Zoonoses and Public Health. 2018;65(4):443-53. doi: 609 doi:10.1111/zph.12458. Jain P, Jain G. Study of General Awareness, Attitude, Behavior, and Practice Study on Dog Bites Sharma S, Agarwal A, Khan AM, Ingle GK. Prevalence of Dog Bites in Rural and Urban Slums of Menghistu HT, Thaiyah AG, Bajitie M, Bundi J, Gugssa G, Bsrat A, et al. Free roaming dogs and Kabeta T, Deresa B, Tigre W, Ward MP, Mor SM. Knowledge, Attitudes and Practices of Animal Ezra Olatunde Ogundare, Oladele Simeon Olatunya, Isaac Oluwadare Oluwayemi, Adekoya Beyene TJ, Mourits MCM, Revie CW, Hogeveen H. Determinants of health seeking behaviour 33 610 42. Inche Zainal Abidin S, Sutan R, Shamsuddin K. Prevalence and determinants of appropriate 611 health seeking behaviour among known diabetics: results from a community-based survey. Advances in 612 Epidemiology. 2014;2014. 613 43. 614 pediatric patients seen at the Research Institute for Tropical Medicine. PIDSP Journal. 2013;14(2). 615 44. 616 Pasteur Institute of Dakar, Senegal: trends and risk factors. BMC Infectious Diseases. 2019;19(1):321. 617 doi: 10.1186/s12879-019-3928-0. 618 45. 619 dog control in urban slums in India. Journal of infection and public health. 2012;5(6):374-80. 620 46. 621 and clinical characteristics of human rabies cases in China, 2006–2012. Scientific Reports. 622 2018;8(1):17188. doi: 10.1038/s41598-018-35158-0. 623 47. 624 near Bengaluru, Karnataka, India. PLoS neglected tropical diseases. 2018;12(8):e0006682-e. doi: 625 10.1371/journal.pntd.0006682. PubMed PMID: 30080901. 626 48. 627 income area of Bengaluru, South India. APCRI Journal. 2016;17(II):6-10. 628 Supporting information 629 S1 Fig. Map showing the location and geographical details of the two districts from which 630 the study participants were got. Kampala Capital City Authority serves as the capital city of 631 Uganda. It is divided into five administrative units called “divisions” and has an estimated human 632 population of 1.5 million. Approximately 8% of households in Kampala own dogs. The average Romero-Sengson RF. Factors affecting compliance to rabies post-exposure prophylaxis among Diallo MK, Diallo AO, Dicko A, Richard V, Espié E. Human rabies post exposure prophylaxis at the Herbert M, Basha R, Thangaraj S. Community perception regarding rabies prevention and stray Guo C, Li Y, Huai Y, Rao CY, Lai S, Mu D, et al. Exposure history, post-exposure prophylaxis use, Masthi NRR, S P. An exploratory study on rabies exposure through contact tracing in a rural area Masthi NR, Vairavasolai P. Estimation of animal bites using GPS and google earth in an urban low 34 633 number of dogs owned per household is 1.7. Wakiso district is divided into 18 administrative units 634 called sub-counties with a population of approximately 2 million people. Approximately 14% of 635 households in Wakiso own dogs. The ownership stands at 1.9 dogs per household. However, the 636 population of stray and free-roaming dogs in both districts is not known. Much as Mulago and 637 Entebbe Hospitals are located in Kampala and Wakiso respectively, dog bite patients from either 638 district can report to any of the healthcare facilities to receive PET. 639 S1 Table. Variables that were studied, indicating their categorization and measures. The 640 factors were organized into host / patient factors, including socio-demographics and those that 641 influence the vulnerability of people to dog bites. Some factors on biting dogs were also studied 642 to give a clear indication of how they influence the dogs to bite as well as the practices of the 643 victims after the bite. Factors on the circumstances of the particular dog bite event were categorized 644 into pre-bite, during the bite and post-bite factors. The intention of this was to study why the event 645 happened and the wound management practices thereafter. The measures indicate how the variable 646 was recorded and / or categorized. 647 S2 Table. Characteristics of biting dogs as reported by the 376 study participants from 648 Wakiso and Kampala districts, Uganda. The characteristics were reported by the dog bite 649 victims or their caretakers, if they knew the details of the biting dog. The frequencies of such 650 characteristics are presented by the specific district of residence and the differences in distribution 651 of characteristics is indicated by the corresponding p-value. Notably, all study participants were 652 residents of the district where the bite event happened. 653 S3 Table. Summary of circumstances of the dog bites, immediate actions taken by victims 654 and reasons for different applications and health seeking behavior. Ten themes were 655 synthesized out of the in-depth interviews to explain the circumstances in which bite events 35 656 happened. Immediate actions taken by bite victims were categorized into four. For those who 657 applied different substances to the bite wounds, the reasons for their choice and actions were 658 recorded into five categories. The same was done to explain why victims went to herbalists, 659 healthcare facilities or had a simultaneous resort. 660 661 662 36