key: cord-0292081-3jlhjdqk
authors: Muylaert, Renata L.; Davidson, Ben; Ngabirano, Alex; Kalema-Zikusoka, Gladys; MacGregor, Hayley; Lloyd-Smith, James O.; Fayaz, Ahmed; Knox, Matthew A.; Hayman, David T. S.
title: Community health and human-animal contacts on the edges of Bwindi Impenetrable National Park, Uganda
date: 2021-07-23
journal: bioRxiv
DOI: 10.1101/2021.07.23.453553
sha: e8c19ec14e8b64ab3f6ce19379c35d5e693bd2e9
doc_id: 292081
cord_uid: 3jlhjdqk

Cross-species transmission of pathogens is intimately linked to human and environmental health. With limited healthcare and challenging living conditions, people living in poverty may be particularly susceptible to endemic and emerging diseases. Similarly, wildlife is impacted by human influences, including pathogen sharing, especially for species in close contact with people and domesticated animals. Here we investigate human and animal contacts and human health in a community living around the Bwindi Impenetrable National Park (BINP), Uganda. We used contact and health survey data to identify opportunities for cross-species pathogen transmission, focusing mostly on people and the endangered mountain gorilla. We conducted a survey with background questions and self-reported diaries to investigate 100 participants’ health, such as symptoms and behaviours, and contact patterns, including direct contacts and sightings over a week. Contacts were revealed through networks, including humans, domestic, peri-domestic, and wild animals for 1) network of contacts seen in the week of background questionnaire completion, 2) network of contacts seen during the diary week. Participants frequently felt unwell during the study, reporting from one to 10 disease symptoms at different intensity levels (maximum of seven symptoms in one day), with severe symptoms comprising 6.4% of the diary records and tiredness and headaches the most common symptoms. Besides human-human contacts, direct contacts with livestock and peri-domestic animals were the most common. Wildlife contacts were the rarest, including one direct contact with gorilla with a concerning timeline of reported symptoms. The contact networks were moderately connected and revealing a preference in contacts within the same species or taxon and within their groups. Despite sightings of wildlife being much more common than touching, one participant declared direct contact with a mountain gorilla during the week. Gorillas were seen very close to six animal taxa (including themselves) considering all interaction types, mostly seen closer to other gorillas, but also people and domestic animals. Our findings reveal a local human population with recurrent symptoms of illness in a location with intense exposure to factors that can increase pathogen transmission, such as direct contact with domestic and wild animals and proximity among animal species. Despite significant biases and study limitations, the information generated here can guide future studies, such as models for disease spread and One Health interventions.

domesticated animals. Here we investigate human and animal contacts and human health in a 23 community living around the Bwindi Impenetrable National Park (BINP), Uganda. We used contact 24 and health survey data to identify opportunities for cross-species pathogen transmission, focusing 25 mostly on people and the endangered mountain gorilla. We conducted a survey with background 26 questions and self-reported diaries to investigate 100 participants' health, such as symptoms and 27 behaviours, and contact patterns, including direct contacts and sightings over a week. Contacts were 28 revealed through networks, including humans, domestic, peri-domestic, and wild animals for 1) network 29 of contacts seen in the week of background questionnaire completion, 2) network of contacts seen 30 during the diary week. Participants frequently felt unwell during the study, reporting from one to 10 31 disease symptoms at different intensity levels (maximum of seven symptoms in one day), with severe 32 symptoms comprising 6.4% of the diary records and tiredness and headaches the most common 33 symptoms. Besides human-human contacts, direct contacts with livestock and peri-domestic animals 34 were the most common. Wildlife contacts were the rarest, including one direct contact with gorilla with 35 a concerning timeline of reported symptoms. The contact networks were moderately connected and 36 revealing a preference in contacts within the same species or taxon and within their groups. Despite 37 sightings of wildlife being much more common than touching, one participant declared direct contact 38 with a mountain gorilla during the week. Gorillas were seen very close to six animal taxa (including 39 themselves) considering all interaction types, mostly seen closer to other gorillas, but also people and 

In fact, most human infections have their origins in wild or domestic animals. Therefore, despite 53 providing many benefits to health-such as overall well-being and services, including medicinal herbs 54 and provision of other natural resources [2-4]-green natural areas can also pose risks to human health, 55 especially when living in proximity to green areas is linked to social inequality and low-socio-economic 56 status.

In low to middle-income countries, and particularly in tropical regions, people also still suffer 58 from a high burden of numerous and treatable endemic diseases. These infections may be human-59 adapted (e.g., malaria [5]) or zoonotic (e.g. leptospirosis [6] ). Contacts between humans and other 60 animals can, therefore, pose hazards to human health, but the converse is also true [7] . Human behaviour 61 has contributed to species extinction rates much higher than pre-Anthropocene levels [8] . Habitat 62 destruction, climate change, pollution, species invasion, wildlife trade, and overharvesting [9] [10] [11] can 63 negatively impact ecosystems, leading to the declining health of many species [12] . In some cases, this 64 may be in the form of increased risk of pathogen transmission, with many examples where infectious Moreover, pathogens spread by the faecal-oral route can be readily shared among species that overlap 95 spatially and share resources such as waterways. On the other hand, regarding the gorilla-human 96 interface, there are no reports of HIV transmission between humans and gorillas. However, HIV causes 97 immunosuppression in humans, leaving people susceptible to tuberculosis and other respiratory 98 diseases, which pose a threat to gorilla populations [25, 33, 34] . Nevertheless, there are 99 immunodeficiency viruses that had their ancestry in apes [35] , such as SIVcpz from chimpanzees in 100 Cameroon [36] , which is thought to be the ancestral group for the pandemic HIV-1 group M. Thus, it 101 is essential to investigate contact patterns between humans and other primates at the boundaries of 102 natural areas.

Interdisciplinary frameworks consider key ultimate and proximate components that drive cross-104 species infection transmission to understand the process of disease emergence [37, 38] . Studying and 105 addressing these components requires multidisciplinary teams comprising local community members, 106 ecologists, anthropologists, physicians, and more [39, 40] . Ultimate drivers include broad 107 environmental, social and politico-economic drivers operating at different timescales and spatial scales 108 [38, 41] . Living conditions and higher-level structural drivers, beyond the actions of individuals, may 109 have a large impact on people's livelihoods and the outcomes of the interactions between humans and 110 nature. Human communities in poor regions in proximity to natural areas frequently suffer from various 111 diseases and take the toll of inadequate health facilities, the precarity of livelihoods, food insecurity, 112 illiteracy, and concurrent proximity to livestock and wildlife, making them susceptible to cross-species 113 transmission. The proximate drivers comprise the 1) prevalence in the reservoir, which falls in the scope 114 of disease ecology; 2) contact rates, a prevalent knowledge gap for many systems; and 3) probability of 115 infection given exposure, which falls in the domain of microbial risk assessment.

Regarding contact rates, proximate driver 2, quantifying behaviour and contact patterns to 117 inform disease spread models can be challenging. Large projects attempting to measure demographics 118 and patterns of human-human contact have been conducted worldwide at small [42] and large scales

[43], with essential applications in epidemiological models [44] . While identifying human contact 120 patterns is important to predict onward disease spread, contact with wildlife and livestock is a crucial 121 factor allowing pathogen cross-species transmission to happen. However, data sets describing human-animal contacts together with animal-animal contacts (such as Narat et al., in Central Africa [42] ) are 123 rarely reported [45, 46] . Contact data between numerous species can be cumbersome to gather by direct 124 observation but can be estimated from survey participants' perspectives, whereas self-declared 125 information can reveal relevant general contact patterns and health. Questionnaire research is hardly 126 ever objective [47] ; nonetheless, self-declared information can be used as feedback to guide actions 127 most needed in small communities. Interventions like keeping distance from wildlife and handwashing 128 before preparing meals can greatly reduce the cross-over of disease-causing pathogens [48, 49] .

Initiatives to promote these types of needed actions have been made in the BINP, which economically 130 relies heavily on subsistence agriculture [50] . In this ecosystem, close relationships people have with 131 other people, their livestock and other animals can be explored through social science approaches and 132 measured by further focused enquiry informed by conceptual models of cross-species transmission 133 [37, 39, 40] . Understanding the potential contacts between humans and other animals and their network 134 structure can help inform the control measures required to manage and prevent future disease outbreaks.

This study used questionnaires to investigate people's health and potential risk factors for 136 infectious diseases investigating the contacts among humans, domestic, peri-domestic and wildlife, with 137 a particular focus on humans and the endangered mountain gorilla. We describe contact patterns in 138 humans and other animals living close to BINP through self-reported surveys for one week. We 139 analysed contact patterns for 1) human-human contacts, including the participants' social contacts and 140 their reported observations; 2) animal-human contacts, including direct contacts and sightings and 3) 141 their observation of animal-animal contacts in proximity to the animal-human interface. We expected 142 human contact patterns to follow patterns reported previously for southwestern Uganda [51]. We 143 expected wildlife sightings would be much more common than touching, with the existence of close 144 contact preferences within-taxon and within groups (domestic, peri-domestic, people, and wild 145 animals). Our questionnaire provides insights on community health and human-animal interface in two 146 villages near BINP.

149 This study took place in Mukono parish at Bwindi Impenetrable National Park's Buhoma sector 150 in Uganda, between June and August 2018. Questionnaires were completed by local villagers living 151 next to the northwest border of BINP, Uganda. BINP covers an area of 331 km 2 in southwestern Uganda 152 ( Fig 1A, Fig 1B) , and the villages are adjacent to BINP at the north-western edge of the park. This 

However, there is still local concern regarding the distribution of the benefits to the community. The 

Questionnaires were designed by our multidisciplinary team, where the key information 223 required included human-human, human-animal and animal-animal contacts. The questions were 224 written in a manner appropriate to the local community and included pictures to provide clarity and to 225 be as easy as possible to complete. Interviewees were recruited by AN, from CTPH, who is a local 226 community member who has worked for CTPH since 2005. Small field trials were performed by DH and AN and the questions were refined and tested to ensure the data collected would be in a useable 228 form and culturally acceptable. All further community engagement was by CTPH alone, led by AN. By 229 asking the same question in different ways unreliable answers could be identified and these were 230 excluded from the final analysis.

Participants were recruited through community meetings and visiting homes. Participant 232 selection was through convenience sampling and based in part on participants' willingness or perceived 233 ability to complete the questionnaire. Only participants who gave their informed consent were included 234 in this study, all older than 18 years. The participants' answer sheets were identified by unique 235 identification numbers and their identity was not known to the authors, except AN.

Two types of questionnaire were used: first, interview-type questions (S1 File) and, second, a 237 seven-day questionnaire diary (S2 File); both were completed between June and July 2018. The 238 questionnaire and diaries were introduced to the study participants in community meetings led by AN.

During the meetings, AN would speak in English and Rukiga as necessary. The completed questionnaire 240 sheets were returned to the CTPH centre and stored in a locked room. In August 2018 data was entered 241 into a Microsoft® Access database by DH and BD, keeping the identity of interviewees anonymous. If 242 clarifications were required these were performed by AN.

The background questionnaire consisted of 29 questions, most of which could be answered by 245 ticking an appropriate box (S1 Data File). Information collected was about age, gender, ethnicity, 246 education, number of people living in the household, marital status and the number of children.

Questions including the types and number of latrines used, number of mosquito nets, and any ill health 248 and associated symptoms experienced were also asked. Symptom severity was indicated as mild ('does 249 not interfere with your day too much'), moderate ('interferes a lot with your life, but you can be 250 independent') or severe ('you were unable to behave normally and needed substantial help'). A request 251 for the vaccination history of each participant and their family members, as per their vaccination 252 records, was included. The final question asked the study participants to identify any animals from a 253 list that they had seen, touched or got close to. The background information was confirmed in a follow-254 up meeting, which was useful to get the participants ready for completing the diary.

The diary included 49 questions, 48 of which were in a checklist format, allowing participants 257 to answer with a tick (S2 Data File). People were asked to record questions relating to hygiene, such as 258 toileting frequency and location, as well as details around handwashing and food preparation.

Behaviours associated with an increased risk of zoonotic pathogen transmission, such as contact with 

Participants were also asked if they touched other people, saw other people touching each other, 268 or attended any group meetings that day. The final question asked participants to identify any animal-269 to-animal contact that they had observed with this question: "Today have you seen the following touch 270 or got very close (e.g. be together in the same field, trees or plantation)". This was done by drawing 271 lines (links) between two identical lists of animals to form an interaction network. Each line then 272 represented that the animals were observed to have close contact (were together in the same field, trees 273 or plantation). Each animal was a node in a pictorial network. The list included 22 taxa: humans, eight 274 domestic (cow, goat, sheep, pig, cat, dog, chicken, rabbit), two peri-domestic considered as one taxa 275 (mouse and rat were summed altogether for the sake of comparison as data from the diary did not 276 include mouse), and twelve wild animal taxa (gorilla, monkey, baboon, bushpig, civet, chimpanzee, 

Human social contact data 296 For direct social contacts, data was collected in age classes and the frequency (0-5 and above 297 contacts per day) for each pairwise interaction. Frequency histograms for contacts between all age class 298 combinations were created. We calculated all human-human direct contacts (counts) and their non-299 parametric 95% confidence intervals based on a bootstrap procedure with 1000 interactions per unique 300 pair of age classes, containing the interviewee and their direct number of contacts. Then, observed 301 contact rates were calculated as the number of days on which the interaction between age classes was 302 seen (presence or absence) over seven days. While children were not included as participants in our 303 study, we believe asking for observed contacts goes some distance to collect unbiased data about their 304 contact patterns. This means the calculation for observed contact rates (sum of days on which the interaction between age classes were seen for all individuals, separated by age class) is different from 306 the self-declared direct contacts (sum of contacts on for each interaction for all individuals, separated 307 by age class).

First, data were gathered for each participant, tabulating how many times they had close contact 310 with animals over the week. Then, timeline plots were produced to illustrate the number of self-reported 311 direct contacts (touching) or sightings of animals or their faeces per individual over the week.

Human-animal interface networks were produced from the observations of contacts from two 314 sources: 1) Contacts from the background questionnaire week (link weight = number of participants 315 reporting that contact), 2) Sum of all contacts along the week (link weight = sum of the contacts seen 316 along the week/7). To explore contact information consistency, we correlated the link weights between 317 the background questionnaire week and diary week with Spearman's correlation test. We believe this 318 approach helps to complement the actual contacts each participant had with contacts that happened in 319 the background questionnaire week and give a broad picture of the local context of contacts. We 320 calculated the animal-animal contact network structure using a network science approach, where a node 321 is a taxon (a species or group of species in several cases, like duiker).and the links between each pair of 322 taxa have weights given by the total number of events along the week when the participants have seen 323 a direct contact or a close contact (be together in the same field, trees or plantation). Where appropriate 324 for network analyses, we converted the edges to binary (present or absent).

To measure how connected the two networks of contacts were we calculated two descriptors: 326 connectance and assortativity. Connectance measures the proportion of binary links observed in a 327 network in relation to the number of potential links it could maximally have. We also quantified the 328 relative frequency of interactions within versus between particular species or groups, using the nominal 329 assortativity metric "r" [67] . We calculated the metric r for two levels: node-node assortativity and 330 assortativity based on broader groups (domestic, peri-domestic, human, wild). This metric is positive 331 when there is a pattern of more contacts with that same node (or group) than with other nodes (groups), and negative when the network presents a disassortative pattern, i.e. when a node (group) has more 333 contacts with other nodes (groups) than with the same. Networks were built in R 4.0.3 [68]; the 334 calculation of metrics and network drawing were performed with the packages igraph and ggplot2 335 [69, 70] .

Demography and health 338 Background information 339 All the participants recruited for this study completed the background questions (S1 Data File) 340 and the diaries (S2 Data File), except participant 83, who did not complete the background information.

In a small proportion of the questionnaires, individuals did not declare their age (N=2), village (N=1), 342 ethnicity (N=2) or gender (N=1). All participants filled the diaries, but there was variation in the number 343 of days completed: one participant completed the diaries for two days, two people completed them for 344 five days, ten people completed them for six days, 78 participants completed them for seven days, seven 345 participants completed them for eight days, one person completed them for nine days, and one person 346 completed them for ten days. These answers add up to 700 days of daily data for 100 participants.

The sample comprised 81% male to 19% female participants, from 19 to 70 years old (median The background questionnaire included a question that sought to identify diseases that any of the 355 participants had ever suffered from, but unfortunately this question was generally misunderstood, since 356 many identified diseases that they had been previously vaccinated against. Because of that, the answers were not included in the analysis. Most interviewees used pit latrines (98%; 73% had covered pit 358 latrines) and had at least one bed net in the house (Table 1) 

Age (mean ±SD) 37.9 ± 12.6

Married 77

Age when married (mean ±SD) 20.7 ± 3.7

Primary complete 52 Unwell in the last week

No 48

Yes 99

No 69

Yes 30

No 79

Yes 29

No 65

Yes 34

No 93

Yes 6 369 370

Regarding meat preparation, 58 of the participants reported in their diaries that they had handled 372 raw meat (N= 101 entries for handling raw meat), 14 had been bitten or scratched by an animal (N=22 373 entries) and 41 participants had come into contact with the blood of an animal (N=66 entries for 374 touching blood) at least once along the week. All participants reported washing their hands before eating 375 at least once during the week, but did not always use soap to wash their hands, which was done 72% of 376 the time (N=700 entries for all days, N=504 answered "yes" to wash hands with soap before meals).

Participants who reported having washed hands with soap before meals and also washed hands with 378 soap after going to the toilet at least once a week summed 94, and using soap for both situations (used 379 soap before meals and after toilet) happened 63% of the time. Regarding nutrition, it was more likely 380 that people would report eating two meals a day (42.3%) than one meal (25.8%) or three meals (25.5%).

The staple diet consisted mainly of posho, beans, matooke and porridge with 97.3% of all meals recorded consisting of one or more these carbohydrates. Meals containing animal protein were recorded 383 in 10.7% of all recorded entries.

In the diary, people reported feeling unwell on 281 of the 700 total days recorded (40%) with 385 11.6% of these records indicating moderate symptoms and 6.4% reflecting severe symptoms. Most 

Of the syndromes reported, headaches and tiredness were the most common (Fig 2) . Pain and 393 coughing were reported often (8.7% and 9.5% of the time, respectively), while diarrhoea and fever were 394 less common (2.4% and 4.5% of the time, respectively); itching and skin rashes were reported the least.

Two individuals reported seven symptoms on a single day. A person was more likely to feel unwell for 396 one, two or three days (20.2%, 12.9% and 5.1% respectively), compared to four or more days (less than 397 1%). The average number of symptoms reported was 1.7 per person (min=0, max=7 symptoms per day),

where it was most common to report no symptom (59.8%), or 1 symptom per day (20% of entries).

People were more likely to pass a bowel movement more than twice in a day compared to once 

We report self-reported close contacts (touching) for 100 individuals (Table 1, Fig 3) .

Participants reported a total of 5777 direct contacts, but 113 contacts from 2 participants were removed 414 because we could not retrieve the participants' ages. Thus, we ended up with 5664 human-human direct Table) . 

From the diary questions, where participants were asked if they had seen a certain type of animal 429 or their faeces that week, the domestic animals such as chickens, goats, cows and pigs were reported as 430 the most common animals seen compared to wild animals. Peri-domestic animals were also frequently 431 seen. Seeing was more common than touching, and wild animals were rarely reported to be touched 432 (Fig 4) . The animals that were reported to be touched most frequently were goats, chicken, pigs and 433 cows, cats and dogs. However, there were very rare records of touching wild animals. One person reported to touch a civet and a monkey at the same day. One participant reported to have touched a 435 duiker, and there were two events of touching a squirrel. Moreover, 4 participants reported having 436 touched a monkey, with one participant reporting having touched a monkey twice in a week (S3 Table) .

Contact with animal faeces/dung-only reported for domestic and peri-domestic animals, since their 

The contact networks illustrate that people involved in this study had regular contact or had 468 been in close proximity to both domestic animals and wildlife (Fig 5) . Both networks were moderately 469 connected; The connectance value for the human-animal contact network using the diary data (divided 470 by 7) was 0.31, and 0.26 for the background questionnaire week network. The observations of contacts 471 revealed many close contacts between groups of humans, domestic, peri-domestic and wildlife.

Regarding assortativity (preferential links) in the network, taxa tended to contact their similar taxa over 473 other taxa (r=0.24 for diary network, r=0.29 for the background questionnaire week network). In 474 addition, the same group (groups = domestic, human, peri-domestic or wildlife) also tended to contact 475 the same group over contacting other groups (r=0.40 for diary network, r=0.32 for background 476 questionnaire week network). The complete interaction list for the contact networks considering 477 background questionnaire week interactions and summed diary interactions is available in S3 Table. 478

From the close contact networks, the most common interaction observed between groups were cow-479 person (S3 Table) , followed by goat-person, chicken (domestic)-person, followed by pig-person.

Person-person contacts made up for the highest number of interactions, followed by goat-goat, cow-481 cow and chicken-chicken. Peri-domestic contacts were the 12th most common interaction, while the 482 wildlife species with most reported interaction events were primates. The most reported interaction 483 between wildlife taxa in close contact with humans were baboons, then gorillas, then monkeys. Most 484 common wildlife-wildlife in-taxon contacts were gorilla-gorilla (N=56) and monkey-monkey (N=56).

Wildlife-domestic close contacts were rarely seen, such as squirrel-rabbit, gorilla-cow, chimpanzee- Table. During the diary completion week, humans interacted with 18 taxa, followed by dogs and cows, 490 which were seen in close contact with 13 and 10 taxa. The wild taxon in close contact with the smallest 491 number of other taxa was the civet (1 taxa), followed by duiker and porcupines (all interacting 2 taxa 492 or less, including themselves). Gorillas had rarely been observed in close proximity or were seen 493 touching domestic animals (dogs and cows) and wildlife (baboon and monkey). and peri-domestic animals were very common, and human-human contacts were the most common 513 contacts reported, as expected. Seeing wild animals was much more common than touching them, with 514 only nine directly reported events of touching a wild animal, and one case of person touching a gorilla.

These numbers make up much less than 1% of total animal-person direct self-reported contacts (N=1424), with most direct human-animal contacts being with domestic or peri-domestic animals.

However, people's perceptions of close contacts along the week revealed hundreds of close contacts 518 between humans and 20 animal species considered in the survey (S3 Table, S4 Table) . Below we discuss 519 the self-reported health patterns, focusing on the interaction between people and mountain gorillas, the 520 local context, social contact patterns and the broad human-animal interface. Then, we discuss caveats 521 of this study and future directions.

Demography and health 

With a local human prevalence of intestinal parasitism ranging from 25% to 57% and the prevalence of 529 other enteric pathogens at 49%, including Giardia and Cryptosporidium [71], it is possible that 530 diarrhoea was under-reported in this study (17 cases in 700 entries, 2.4%), given that people reported 531 passing a bowel motion up to five times in a day without reporting diarrhoea. In addition, over a quarter 532 of the participants reported drinking un-boiled water, potentially exposing them to water-borne disease 533 risk.

Regarding disease symptoms, although guidance was given in the questionnaire regarding 535 severity levels, all information gathered was self-reported and therefore subjective. Moreover, if the 536 'normal' state for an individual is feeling constantly tired with headaches (Fig 2) , possibly due to poor 537 nutrition, heat and dehydration, the individual may not report these as symptoms. This may also be the 538 case for individuals suffering from chronic diseases who have lived with the symptoms for many years 539 and have learnt to tolerate them. This may mean that the symptoms' severity could be under-reported 540 so these results should be interpreted with caution, such as the difference between reported diarrhoea 541 and the frequency of defecation reported. It is common for people in this community to only eat two 542 meals a day, with a diet that is high in carbohydrates, low in protein, and generally of low nutritive value, suggesting that most of the local population is likely to be malnourished. Food insecurity in the 544 area was beyond the focus of this study and should be assessed in depth using survey methods designed 545 for that purpose (e.g., [72] ).

Considering risk factors for disease exposure, 68% of people reported handling raw meat and 547 animal blood, and 28 people indicated that they had cuts on their hands during the week. Eleven 

However, these data give more insights into contacts for children and teenagers because, while 585 participants were always older than 18 years old, they could observe the presence of a variety of age 586 classes interacting during their daily routine (S6 Figure) . 

The networks of close contacts plus direct reports of contacts revealed hundreds of interaction 599 events, with a positive correlation (Spearman's correlation test rho=0.67, p<0.0001) between the 600 interactions reported for the background questionnaire week and the interactions occurring during the 601 diary week (Fig 5) . From those, despite rare events, wild animals were mentioned to be having close 602 contacts with humans, for instance, gorillas, baboons, squirrel, monkey, and duiker (S3 Table, S4 603 week) and dogs and gorillas (also 0.14 contacts/day and zero for the background questionnaire week,

S3 

This study has significant limitations, such as recall bias, typically associated with diary-based 667 research [94]. However, filling diaries for a short period tends to minimize recall bias due to the small-668 time lag between an event occurring and being recorded compared to a questionnaire [95] . Another 669 bias is the non-random nature of the recruitment of participants since their selection was based on 670 literacy and age. Although the questionnaire and diary questions asked about the health of all household 671 members, the ages of any unwell members were not captured. Because of these limitations, it is possible 672 that the frequency of disease symptoms reported in the diaries in this study underrepresents the actual 673 disease in the community. Moreover, literacy can be associated with the tasks that a person might do in 674 their day-to-day lives, with low literacy levels associated with more prolonged episodes of sickness 675 [96] . More literate adults are more likely to work outside of the home in occupations such as teaching, 676 nursing, or running their own business, whereas nonliterate individuals are more likely to be doing manual work at home and working with animals and crops [3] . The latter activities present a higher risk 678 for human contact with livestock (which is the most common type of cross-species contact, see S3 679 Table) or wildlife and associated pathogen transmission. We will investigate community activities in 680 future works but note that most activities in the area point to subsistence farming [92, 97] , which is 681 noticeable in our results showing the frequent contacts with livestock.

Furthermore, male participants were overrepresented in this study (81% males, 19% females).

This disparity in the selection of study recruits may partly reflect convenience, but partly because of 684 gender biases and patriarchal systems, meaning more men volunteer as our community member who 685 administered the questionnaires (AN) was male. As different genders can perform different activities, 686 the diseases associated with these tasks can be gender-sensitive [98] . An example of this is the gender-687 specific division of labour, which is a feature of the Bakiga tribe, with men responsible for hunting and 

The Batwa were resettled from within BINP when the park was established, and they have been defined 

This study highlights a local human population that is generally unwell and may act as a 733 reservoir for pathogens. Humans and domestic animals form a tightly coupled system that may act as a 734 reservoir for infectious diseases that could spill over to endangered wildlife, including the mountain 735 gorilla. Despite having several caveats, our findings are relevant to inform local community workers 736 and stakeholders of the overall health condition and contact patterns of the local community around 737 BINP. Several behaviours that may increase zoonotic disease transmission risk were identified for 738 Mukono and Nkwenda villages in the Buhoma sector. With fewer than 1000 mountain gorilla mature 739 individuals remaining in their natural habitat [15] , it is a matter of urgency that we gain the knowledge 740 that will allow us to best mitigate the risks and prevent the transmission of infectious diseases to this 741 endangered species. More broadly, the improvement of healthcare and living conditions in the region 742 is of utmost importance to reduce the burden of pathogens in humans and wildlife. In this case study, 743 we provide evidence of an active interface between humans and other species, including cattle, goats, 744 baboons, and the endangered mountain gorilla around the BINP. We offer the first characterization of 745 the human-animal contact network around BINP based on survey data and report frequent interactions 746 indicating strong potential for pathogen cross-species transmission among various animal species.

We thank the team from CTPH in Uganda for their invaluable services to the community. We 

Pardini 757 R. Intention of preserving forest remnants among landowners in the Atlantic Forest: 758 The role of the ecological context via ecosystem services

Local disease -761 ecosystem -livelihood dynamics: Reflections from comparative case studies in Africa

Ecosystem services of urban green spaces in 765 African countries -perspectives and challenges

Leptospira 771 Seroprevalence and Risk Factors in Health Centre Patients in Hoima District, Western 772 Uganda

Estimating the normal 778 background rate of species extinction

Alien species as a driver of recent extinctions

Emergence of diseases from wildlife reservoirs. Veterinary 786 Pathology

Global factors driving emerging infectious diseases: Impact 788 on wildlife populations

A canine distemper virus epidemic in Serengeti lions (Panthera leo)

Rabies and 794 distemper outbreaks in smallest ethiopian wolf population. Emerging Infectious 795 Diseases

797 Gorilla beringei ssp. beringei. The IUCN Red List of Threatened Species 2018. The 798 IUCN Red List of Threatened Species

Phylogeny and geography predict pathogen community 801 similarity in wild primates and humans

The grand challenge of great ape health and 804 conservation in the anthropocene

Bas Huijbregts, Pauwel De Wachter, Louis Sosthe` ne NO and MEA. Ebola and the 810 decline of gorilla (Gorilla gorilla) and chimpanzee

The nexus between forest 814 fragmentation in Africa and Ebola virus disease outbreaks

United Nations, Department of Economic and Social Affairs PDivision. World 824 Population Prospects: The

Understanding Different Types of 827 Subsistence Economies: The Case of the Batwa of Buhoma

Pathogen Transmission 830 from Humans to Great Apes is a Growing Threat to Primate Conservation

833 Mountain gorilla genomes reveal the impact of long-term population decline and 834 inbreeding

Habituating the great apes: The disease 836 risks

Codetection of Respiratory Syncytial Virus in Habituated Wild Western Lowland 842 Gorillas and Humans During a Respiratory Disease Outbreak

Hyperkeratotic mange 848 caused by Sarcoptes scabiei (Acariformes: Sarcoptidae) in juvenile human-habituated 849 mountain gorillas (Gorilla gorilla beringei)

A 862 new human immunodeficiency virus derived from gorillas

870 Pathways to zoonotic spillover

Is there really such a thing as "one health"? Thinking about a more than 873 human world from the perspective of cultural anthropology

Views from many worlds: Unsettling categories in 876 interdisciplinary research on endemic zoonotic diseases

A Unified Framework for the Infection Dynamics of Zoonotic Spillover and Spread

Using physical 882 contact heterogeneity and frequency to characterize dynamics of human exposure to 883 non-human primate bodily fluids in central Africa

Projecting social contact matrices in 152 countries using 886 contact surveys and demographic data

The effect of 889 control strategies to reduce social mixing on outcomes of the COVID-19 epidemic

Craft ME. Infectious disease transmission and contact networks in wildlife and 893 livestock

Why infectious disease research needs 896 community ecology

A community-based One Health education program for disease risk mitigation at 901 the human-animal interface

Characteristics of human encounters and social mixing patterns relevant to 911 infectious diseases spread by close contact: A survey in Southwest Uganda. BMC 912 Infectious Diseases

A 918 retrospective cross sectional study of the effectiveness of a project in improving infant 919 health in Bwindi, South Western Uganda. Frontiers in Public Health

Determining the Human Diseases Transmissible To the 926 Great Apes of Western Uganda

OIE 928 Manual of Diagnostic Tests and Vaccines for Terrestrial Animals

Ape tourism and human diseases: how close should we get? Tourism

Conservation in a Region of 938 Political Instability: Bwindi Impenetrable Forest

East Africa Handbook: The Travel Guide. Footprint Travel Guides. 7th 941 Revise. Footprint Handbooks

Conservation through Community use of Plant Resources: 946 Establishing collaborative management at Bwindi Impenetrable and Mgahinga Gorilla 947 National Parks, Uganda. People and Plants Initiative. 1996; 1-45. 948 64. Rwego IB. Prevalence of clinical signs in mountain gorillas, Bwindi Impenetrable 949 National Park

Surveys: monitoring mountain gorillas, other select mammals, and human activities

An integrated health approach to the conservation of 958 Mountain gorillas Gorilla beringei beringei. International Zoo Yearbook

963 68. R Core Team. R: A language and environment for statistical computing. Vienna, 964 Austria: R Foundation for Statistical Computing

The igraph software package for complex network research. 967 InterJournal

ggplot2 Elegant Graphics for Data Analysis (Use R!)

Medical survey 971 of the local human population to determine possible health risks to the mountain 972 gorillas of Bwindi impenetrable forest National Park Uganda

Gastrointestinal bacterial 981 transmission among humans, mountain gorillas, and livestock in Bwindi Impenetrable 982 National Park

Campylobacteriosis, salmonellosis, and shigellosis in free-ranging human-habituated 986 mountain gorillas of Uganda

Cryptosporidium parvum Genotype 2 infections, in free-ranging mountain gorillas 990 (Gorilla gorilla beringei) of the Bwindi Impenetrable National Park

Gastrointestinal Parasites 993 of Mountain Gorillas (Gorilla Gorilla Beringei) in the Parc National Des Volcans

Minimizing pathogen transmission at primate 997 ecotourism destinations: The need for input from travel medicine

Risk of Disease Transmission between 1000 Conservation Personnel and the Mountain Gorillas: Results from an Employee Health 1001 Program in Rwanda

Self-reported medical history survey of 1003 humans as a measure of health risk to the chimpanzees (Pan troglodytes 1004 schweinfurthii

Human-1007 Wildlife Interactions Predict Febrile Illness in Park Landscapes of Western Uganda

Diseases of humans and their domestic 1013 mammals: Pathogen characteristics, host range and the risk of emergence

Prevention of 1017 Cryptosporidium and GIARDIA at the human/Gorilla/livestock interface. Frontiers in 1018 Public Health

Primate conservation: The prevention of disease transmission

Parasitic health of olive baboons in Bwindi 1026 Impenetrable National Park

A novel adenovirus species 1029 associated with an acute respiratory outbreak in a baboon colony and evidence of 1030 coincident human infection

Cross-1032 species transmission of a novel adenovirus associated with a fulminant pneumonia 1033 outbreak in a new world monkey colony

92. Namara A. From Paternalism to Real Partnership with Local Communities? 1041 Experiences from Bwindi Impenetrable National Park (Uganda)

Demand and proximity: Drivers of illegal forest resource 1044 extraction

Smith-Greenaway E. Are literacy skills associated with young adults' health in 1051 Africa? Evidence from Malawi

Penetrating the Impenetrable: Establishing profiles and motivations of 1054 resource users at Bwindi Impenetrable. 2013; 100. 1055 98

The implications of socioeconomic status bordering commnunities on 1059 sustainability of natural resources with and adjacent to protected area: the case of 1060

Vulnerability of indigenous health to climate change: A case study of Uganda's Batwa 1063 Pygmies

1066 Seasonal variation of food security among the Batwa of Kanungu, Uganda. Public 1067 Health Nutrition

Equity and poverty alleviation understanding profiles and motivations of resource 1070 users and local perceptions of governance at Bwindi Impenetrable National Park

Modeling infectious disease dynamics in the complex landscape of global health

Supporting Information

File Background questions.pdf" Virtual version of the questionnaire that was printed to the 1079 participants in Buhoma

S2 File Diary questions.pdf". Virtual version of the dairy that was printed to the participants 1081 in Buhoma

Data_File_S1_Survey.xlsx

Data_File_S2_Daily.xlsx". Relevant data from the dairy containing the answers from 1085 to the participants in Buhoma

Uganda, showing elements that are part of the local living 1087 conditions around Bwindi Impenetrable National Park. A. Set for washing hands and water usage 1088 outside a house. B. Water supply in plastic containers. C. Water supply from artesian well. D. Local 1089 butchery entrance. D. and E. goats very close to the Park border. F. Cooking area in one of the village's 1090 houses

Pictures from Buhoma, Uganda, showing parts of the village. A. Panoramic view of the 1092 street. B. View of the village close to the Bwindi Impenetrable National Park and livestock in the 1093 background. C. Closer view of one of the local houses

Gender and age distribution of the participants. Gender bias in our sampling is a 1095 significant study limitation

Average number of children per participant age group. The average number of children 1097 per age group of participants varied from no children older than six years old for younger participants 1098 (between 16-20 years old) to more than three children for participants above 30 years old

Health care behaviours in Buhoma during a self-reported survey conducted

Treatment choices of those who declared to be unwell or seek treatment

Self-reported observed human-human contacts around BNP. Observed contacts were 1104 collected along one week in a self-reported diary per age class

Self-reported direct contacts between humans and animal faeces around BNP. Contacts 1106 were collected along one week in a self-reported diary. A. Sightings of dung in gray. B. Direct contact 1107 with dung in gray

Timeline with symptoms reported for a participant who reported touching a 1110 mountain gorilla. Sources for the gorilla and hand sillhouettes: Creazilla

Age classes of self-reported direct contacts (n=98) with the people they touched along 1114 seven days in Buhoma, Uganda. Two individuals from the total sample (n=100) did not declare their 1115 age and were removed from the analysis

Self-reported events of actual touching certain animals or their dung the previous week 1117 in Buhoma

The number of contact events is in decrescent order based on the 1121 diary week sum of events divided by seven days

Taxon-level metrics reported in Buhoma according to the participant's observations of 1123 human and animal close contacts. Data is arranged in ascending order of the number of species a node 1124 interacted with