RESEARCH ARTICLE Open Access

The first mile: community experience ofoutbreak control during an Ebola outbreakin Luwero District, UgandaDaniel H. de Vries1*, Jude T. Rwemisisi1,2, Laban K. Musinguzi1,3, Turinawe E. Benoni1,3, Denis Muhangi3,Marije de Groot1, David Kaawa-Mafigiri4 and Robert Pool1

Abstract

Background: A major challenge to outbreak control lies in early detection of viral haemorrhagic fevers (VHFs)in local community contexts during the critical initial stages of an epidemic, when risk of spreading is itshighest (“the first mile”). In this paper we document how a major Ebola outbreak control effort in centralUganda in 2012 was experienced from the perspective of the community. We ask to what extent thecommunity became a resource for early detection, and identify problems encountered with communityhealth worker and social mobilization strategies.

Methods: Analysis is based on first-hand ethnographic data from the center of a small Ebola outbreak inLuwero Country, Uganda, in 2012. Three of this paper’s authors were engaged in an 18 month period offieldwork on community health resources when the outbreak occurred. In total, 13 respondents from theoutbreak site were interviewed, along with 21 key informants and 61 focus group respondents from nearbyKaguugo Parish. All informants were chosen through non-probability sampling sampling.

Results: Our data illustrate the lack of credibility, from an emic perspective, of biomedical explanations whichignore local understandings. These explanations were undermined by an insensitivity to local culture, amismatch between information circulated and the local interpretative framework, and the inability of theemergency response team to take the time needed to listen and empathize with community needs.Stigmatization of the local community – in particular its belief in amayembe spirits – fuelled historical distrustof the external health system and engendered community-level resistance to early detection.

Conclusions: Given the available anthropological knowledge of a previous outbreak in Northern Uganda, it issurprising that so little serious effort was made this time round to take local sensibilities and culture intoaccount. The “first mile” problem is not only a question of using local resources for early detection, but alsoof making use of the contextual cultural knowledge that has already been collected and is readily available.Despite remarkable technological innovations, outbreak control remains contingent upon human interactionand openness to cultural difference.

Keywords: Ebola, Viral haemorrhagic fever, Outbreak control, Uganda, Amayembe, Spirits, Witchcraft,Community health resources, Anthropology, Luwero

* Correspondence: d.h.devries@uva.nl1Department of Anthropology, University of Amsterdam, Amsterdam, TheNetherlandsFull list of author information is available at the end of the article

© 2016 de Vries et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

de Vries et al. BMC Public Health (2016) 16:161 DOI 10.1186/s12889-016-2852-0

BackgroundThe first mile in disease outbreak controlThere is general agreement that community membershave a key role to play in early detection of Viral Haemor-rhagic Fevers (VHFs) such as Ebola. Indeed, one of thegreatest risks in the spread of VHFs is people lingering inthe community without knowing they have contracted thedisease [1]. At the start of the 2014 West African out-break, Ebola spread unnoticed for three months [2].Recent data suggest that the period of infectiousness inthe community was five days on average, and at worst,more than 40 days [3]. Without effective isolation, it isestimated that each Ebola carrier transmitted the virus tobetween one and eight additional people, leading to adoubling time of around twenty days [4]. Early detectionin the community presents a major challenge however,due to the difficulty for community members of recogniz-ing nonspecific and common early symptoms such asfever, nausea, vomiting, diarrhea, and weakness. This isexacerbated by a general lack of local knowledge aboutbiomedical explanations and by competing indigenousexplanatory frameworks. The problem of how to engagelocal community members effectively in early detectionwhen they remain typically outside the reach of the healthsystem is reminiscent of an inverted scenario dubbed bydisaster managers as the people-centered “last mile” ofearly warning [5, 6]. In the last mile, scientific informationmay be present about an imminent hazard such as anearthquake, but the early warning does not reach thecommunities on the ground in time, and is thereforefutile. We suggest that the “first mile” of VHF outbreakcontrol presents us with a comparable challenge inreverse: to try to capture timely information about animminent outbreak from the community on the ground.To remedy the first mile gap, investment is focused

increasingly on technologies that accelerate bottom-upinformation provision from communities to public healthlevel. Location-tagged community level mobile reportingsystems are linked directly to national command unitsand international emergency aid agencies [1]. Uganda’s E-health infrastructure, for example, provides a foundationfor “real time information sharing”, where information isdistributed instantaneously [7]. This is often seen as thebest solution to the first-mile problem: “With smart-phones in the hands of all community health workers,drivers, facility managers, and district health officials, itshould be possible to create a map of the epidemic in realtime” [4]. Yet, as in the last-mile problem – where in-vestment in high technology fails to reach vulnerablepopulations at risk – technology alone may fail to solvethe first-mile problem if community members are notengaged. The real-time ideal collapses when those margin-alized from the development process are left out of thecommunity detection process [5, 6].

Can a more “people-centered” early detection systemalso be achieved? The standard response to this questionis to suggest an enhanced role for community healthworkers who should be “quickly trained and deployed toidentify people in the community who develop suspi-cious symptoms” [8]. Secondly, a “social mobilization”agenda is often called for, where entire communities par-ticipate in “quickly identifying and isolating infected in-dividuals before they can transmit the virus” [9] – or, asthe World Health Organization plainly puts it, “involveeveryone” [10]. Literature acknowledges that communityhealth worker and social mobilization strategies for thefacilitation of rapid outbreak detection should be com-bined with culturally appropriate community-based trustbuilding [11, 12].In this paper we ask to what the extent the community

became a resource for early detection and outbreak con-trol during the critical initial stages of the epidemic,when risk of spreading was at its highest (“the firstmile”). We identify problems encountered with commu-nity health worker and social mobilization strategies.

Indigenous beliefs and EbolaIndigenous representations of local responses to VHFs inunder-resourced contexts such as rural Africa often includeimages of ignorance, exoticism and superstition, with “riskytraditional burial practices” typically cited to illustrate thecommunity as barrier [12–14]. As anthropologists have ar-gued for decades, popular simplifications of community re-lationships create the impression that the local is stagnant,illogical and ignorant, while representations of internationalhealth communities tend to overemphasize collaboration,cooperation and organization [15]. Previous to the 2014West African outbreak, the anthropologists Hewlett &Hewlett argued that representations of Ebola in the mediaare seldom contextualized; “one is left with a feeling that anoutbreak is controlled only through Western biomedicalknowledge and technology, in spite of, not because of, theactions of local peoples” [16].Unfortunately, for the international health response to

Ebola, only a few case studies are available in publishedliterature detailing how such an outbreak is locally under-stood, responded to, and embedded within communityrelationships. This may be because anthropologists wereinvited late to participate in fieldwork relevant to VHFoutbreak control [16, 17]. The Hewletts published the firstanthropological account of indigenous reactions to Ebola25 years after the first reported Ebola outbreak; they wereinvited only after cases continued to increase duringUganda’s 2000 outbreak, due to perceived “problems withlocal people” [16]. The Hewletts reported that Americanand Italian physicians working in isolation units had noidea how local people explained or viewed Ebola, whilelocal northern Uganda health workers appeared unwilling

de Vries et al. BMC Public Health (2016) 16:161 Page 2 of 12

to bring local cultural beliefs to the fore in an inter-national healthcare response context. As the Hewlettswrote: “They feared that others hearing them talkabout local beliefs would view them as backward,primitive, and exotic” (p41).Investigating the local cultural beliefs of northern

Uganda in the 2000 outbreak, the Hewletts documentedgemo as a “bad” spirit that arrives suddenly and causes amysterious illness within a very short time. They ob-served that gemo was associated with prescriptive behav-iors and protocols for isolation of suspected cases, aswell as burial at the edge of villages; it was commonknowledge to children who had not learned about it inschool. Further, in both the Congo and Uganda casesstudied, they showed that as deaths continued despiteindigenous treatments, people shifted their explanatorymodels from community illnesses to biomedical models.From this perspective, the Hewletts and others haveargued for more respect and understanding and stressedthe central importance of trust and rapport with thelocal community [13].

Ebola in LuweroThis paper investigates this trust and rapport more than adecade later, in the same country that the Hewlettsworked in. After Sudan Ebola was confirmed in LuweroCounty, the Ministry of Health set up a District TaskForce (DTF) with case management, surveillance, eco-logical, psychosocial and social mobilization teams. InLuwero, the International Federation of Red Cross andRed Crescent Societies (IFRC) funded and organized socialmobilization and psychosocial support. The IFRC pledgedto produce and disseminate context-specific information,education and communication materials (38 K posters,50 K leaflets and 120 T-shirts), conduct media campaigns(5 talk shows and 400 radio spots), and conduct one fieldmedia visit [18]. The IFRC enlisted and trained localvolunteers and local community health workers (VHTs) toprovide psychosocial support and built community trustand confidence. According to the evaluation, this processwas successful and “put to a halt the spread of the diseaseand limited mortality as well as contributed to the well-being of the community” [19] (P16). The evaluationclaimed that sensitization measures dissolved any mythsor misconceptions, influenced behaviors and practices andprovided psychosocial support, while information flowenhanced visibility and public confidence in responseactivities [19].How did community members understand and ex-

perience the disease, and the external interventionsthat arrived in its wake? Did the community becomea resource and partner – instead of a “barrier” – tooutbreak control efforts, effectively bridging the firstmile gap?

MethodsThe Ebola outbreak occurred in Kakute village inSsambwe Parish (population 9400), just 30 minutes by carfrom our main field site. We were conducting 18 monthsof ethnographic research in a number of villages inKaguugo Parish (population 4000), part of Luwero district,75 km north of Kampala. As well as proximity, our mainfield site was similar to Kakute village in terms of rurality,Buganda culture and socioeconomic status.Two months after Ssambwe Parish was declared free of

Ebola, on January 16, 2013, one of our Ugandan doctoralresearchers visited Kakute village to conduct interviewswith an Ebola survivor, the DTF Chair, village chair, twonurses at isolation units, and one neighbour of the in-volved family. These key informant interviews were com-plemented by a focus group discussion comprising sevenKakute village participants. A total of 13 Kakute res-pondents were interviewed. The informants in SsambweParish were chosen through non-probability “snowball”sampling. A snowball sample is a non-probability sam-pling technique in which the researcher collects data onthe few members of the target population that he or shecan locate, then asks those individuals to provide informa-tion to locate other members of that population whomthey know. Snowball sampling is appropriate to use whenmembers of a population are difficult to locate, but is notlikely to provide a representative sample. In this case,snowball sampling was used because outbreak messagesto the community had warned village residents to distancethemselves from possibly infected residents and their“contacts”. As a result, key respondents had to some ex-tent “gone underground” with respect to their beliefs. Ourresearcher traced the infected family and communitymembers who had played a significant role influencing thecommunity’s Ebola history. In addition to this targetedfieldwork in Ssambwe Parish, data from ethnographicresearch in nearby Kaguugo Parish by all three Ugandandoctoral students was used for this paper. Our teamobserved and documented the outbreak response controlmeasures, and spoke about Ebola with an additional 21key informant interviews and 61 focus group respondents.Table 1 summarizes the total number and location ofall respondents.All interviews were held in Luganda, the native

language of the Baganda, spoken by the doctoral re-searchers and their Baganda assistants. Interviewswere translated into English. Interview questionswere open-ended. No specific standard interviewinstrument was designed for the purpose of do-cumenting Ebola. Questions were developed in dia-logue, guided by the ethnographic experience of theinterviewers. All interviews and focus groups wereaudiorecorded, transcribed and analysed using theNVivo qualitative software analysis program. Textual

de Vries et al. BMC Public Health (2016) 16:161 Page 3 of 12

information was analysed using grounded theory.Through review of collected data, repeated ideas,concepts or elements became apparent, which weretagged with codes and extracted from the data.These codes were grouped into thematic clusters.This resulted in the following themes: 1) critical dis-course, 2) emergency response, 3) history of spreadin Luwero, 4) perceptions of Ebola, 5) post-Ebolanormalization, and 6) sources of information. Thesethemes are further explored and organized below,divided into major subthemes summarizing the com-munal and indigenous experience.Limitations to this method include subjective bias

that is always part of qualitative study. In this casestudy, a major limitation is the short length of timewe were able to spend on the ground in Kakute com-munity. We also only had opportunity to interview arelatively small number Kakute residents, even thoughthe interviews lasted long enough to obtain neededinsight. However, we feel strongly that the ethnographicunderstanding provided by three separate ethnographersin nearby Kakuugo Parish brings in-depth substanti-ation of prevailing community attitudes. We observelittle cultural difference with nearby community ofKakute. Further, while qualitative and/or ethnographicstudies are also limited in generalizability, results pro-vide conceptual frames of reference that help under-stand dynamic processes otherwise not easily capturedusing statistical methods.The manuscript was reviewed by all authors, and

further insights were added, arising from on-goingdialogue in weekly group meetings of the researchteam. Ethical clearance was obtained from the School ofPublic Health at Makerere University College of HealthSciences (Reg. Number 158) and Uganda National Councilfor Science and Technology (Reg. Number SS 2754).Written informed consent for participation in thestudy was obtained. Observation was conducted withthe approval of village chiefs. No children participatedin the study.

Results and discussionThe outbreakThe Ebola outbreak in Luwero can be traced to a32-year-old Islamic Kakute village motorbike taxi ridercalled Kabugo. His illness began with a fever in October2012, which his family initially took for malaria. Hesuffered for a week before being taken to a health center.After a week of treatment, he spent a night at home wherehis family saw that he was still not well, so they took himto the hospital at the local army barracks. There, hestarted vomiting blood. The hospital staff shifted him toan isolated ward and took blood samples to establish thecause of his illness. Kabugo died shortly thereafter, beforethe results from the tests could be relayed to the family.During his Muslim funeral, his body was laid for finalviewing during the night, and relatives and communitymembers touched and washed his body. He was buriedthe following day according to tradition. During hisfuneral, Kabugo’s widow Halima developed similar symp-toms and was admitted to the health center. She wasattended by her mother-in-law, two sisters in-law, and anaunt who spent the nights with them and left early in themorning to attend to her business. The health staff at theclinic suspected that the widow had malaria, but as shetested negative the doctor concluded it was depressionbecause she had lost a husband, and allowed her to bedischarged. Her father took her to a traditional healer toinvestigate the mysterious cause of her husband’s death.She died shortly after. Two days later one of thesister-in-law developed symptoms and died. Two daysafter that, the second sister-in law died.Staff at the health center started to suspect Marburg

disease, and samples obtained from the patients wereshipped to the Uganda Virus Research Institute. Whenthese tested positive for Sudan Ebola, the Ministry ofHealth declared an Ebola outbreak on November 16,2012, and established an isolation center. By November28, there were six confirmed cases and one probablecase, including four deaths (two from the same family), inLuwero and Kampala [19]. Twelve suspected cases testednegative. Medical personnel noted that their experience ofearlier outbreaks of Ebola and Marburg in nearby regionshelped them to deal with this outbreak; measures includedpooling staff at national level from other parts of thecountry, extra training, mixing experienced staff with newstaff, and provision of psychosocial support for medicalstaff. One medical respondent noted:

In Gulu over 200 people died; in Bundibugyo, over30 people died; in Kibale about 17 died. So whatI could ultimately say is that controlling theepidemic in a short time with minimal loss of lifeis the best thing we gave to the community. Itmight not be easily visible to some people, but as

Table 1 Number of key informants and focus group respondentsby location and Parish in Luwero County

Location Parish # of keyinformants

# of focus grouprespondents

Totals

Bukuma Kaguugo 3 8 11

Bunyenye Kaguugo 0 16 16

Dekabusa Kaguugo 5 15 20

Kakute Ssambwe 6 7 13

Undefined Kaguugo 3 8 11

Kyetume Kaguugo 2 0 2

Ssakabusolo Kaguugo 2 7 9

Totals 21 61 82

de Vries et al. BMC Public Health (2016) 16:161 Page 4 of 12

part of the system I think that was a very bigachievement or reward. It is the lowest number.

One DTF member said that as death in Africa cansometimes be attributed to witchcraft, the main focus ofthe social mobilization team was to strengthen thecommunity component which delivered the “actual def-inition by the health system of the problem, and how itcan be prevented”. The member noted that to convincethe community, the outreach response team used aclinical presentation that focused on clinical symptomsin order to dismiss ideas of witchcraft. The team’s mainaim was to change people’s behaviour quickly:

And what you are aiming at in an epidemic is how tochange people’s behaviour very fast. So we aimprimarily at changing behaviour to control the spreadof the infection. All the other things come later.

This behavioural change was generally perceived to bea “fight against misconception”. According to our DTFrespondent, this fight was won by the social mobilizationteam, which succeeded in getting the facts across. “Thatis why we were able to contain it. We lost only fivepeople from the same village, the same family.”

SpiritsHowever, in Kakute, there was little evidence to supportthe idea that social mobilization had been a success. Themajority of villagers continued to attribute Kabugo’sillness to witchcraft by his mother. One communitymember noted:

There is nobody in this village who believes that itwas Ebola. The people did not even think of any othername for the disease because they were sure that itwas amayembe.

Respondents noted that even the local motorcycle(bodaboda) boys who had worked with Kabugo didnot believe that he had died of Ebola. The local explana-tory framework supporting this interpretation was in placebefore Ebola was diagnosed; when Kabugo fell ill, a trad-itional healer argued that someone had bewitched him. Itwas said that when Kabugo and his kin sought charms,called mayebe, to remedy the witchcraft, one of his broth-ers refused to be involved, fearing that the spirits wouldmix with evil spirits – kifaluu amayembe – and cause thewhole family to die. His fears seemed to be substantiated.His father, accused of witchcraft, was murdered; his ownbrothers among the killers. Upon his death, Kabugocollected his father’s spirits, and unable to meet theirneeds, became mortally ill. A local traditional healer thenwarned that his death would be followed by many more in

his family. Stories were also circulating that Kabugo’smother-in-law had used evil spirits to kill his wife, withwhom she had had an acrimonious relationship. Commu-nity members asserted that in one of her many quarrelswith Kabugo’s wife, the mother-in-law brazenly undressedand showed her bare buttocks, a sign that she cursed herson’s family. They also believed that when Kabugo’s bodywas brought home, a bat entered the house and disap-peared into the body of one of his sisters. The communityaccused his mother of using amayembe to kill her childrenand obtain their property. After Kabugo’s death, commu-nity members went to her home to burn her house downand demand that she leave the village. Although herhusband was able to prevent the burning, she did leave.Members of the Ebola team got wind of this event, foundher on the road and took her to Mulago Ebola Response/Treatment center in Kampala for monitoring.Community members believe in various types of

amayembe: some benign, such as sources of fertility, andothers malicious (kifaluu amayembe), used if a personhas ill feelings towards someone, or is simply greedy andenlists supernatural forces for their own gain. Movingaround at night, these spirits “eat” people and animalsby sucking blood, knocking on doors, and rapingwomen. When out of control, however, amayembe mayturn against their buyers, their families, and sometimesbeyond, particularly if their owners fail to provide suffi-cient sacrifice, in the form of goats or people, in returnfor the wealth that has been promised. In stories, com-munity members describe seeing amayembe or theirtraces. For example, a 25-year-old Kaguugo man vividlydescribed how his whole town went “on fire” when thelocal healer who came to deal with the amayembe fellout of a tree while chasing after it: “We thought he hadbroken his legs, but he would run while hitting his handsas flames would be seen, and there he was chasing theamayembe.” The idea of greed often plays a role in theamayembe narrative, suggesting that the concept repre-sents an embodied rationalization of paranoid suspicion[20] and a social leveling force, both undermining andfacilitating inequalities in wealth and power [21]. Onerespondent related the emergence of amayembe to a timewhen there was a lack of unity in the local communitybecause “some people didn’t behave well”. Amayembe alsotrigger social isolation. Those who get in trouble are notonly punished (through illnesses) by the amayembe, butthey are also ostracized by the community. This isessentially an outbreak prevention strategy. A Kaguugovillager noted:

You see there are those who buy amayembe in thisarea. Now, if those amayembe kill people in thevillage, we give up on such a person in the village.We stop dealing with such a person. So it means that

de Vries et al. BMC Public Health (2016) 16:161 Page 5 of 12

even when such a person loses someone, he has tocare for his own needs (Tumusamu).

Amayembe may also be thought of as a system of beliefthat falls outside governmental control because – as manycommunity members noted – the government does notacknowledge its existence. Amayembe are the domain oftraditional healers at odds with a biomedical public healthsystem. As a Kaguugo community member said:

So the greatest challenge is that when you report thatpeople are dying because of amayembe they will ask‘what is amayembe?’ and you really cannot prove theirexistence. So that is why many people are dying andthe government cannot do anything.

The difficulty of dispelling amayembeThe amayembe interpretative framework described aboveleads to a credibility problem for biomedical explanations,which are seen as unclear and illogical interpretations ofthe genesis of particular outbreaks in the context oftraditional aetiologies. This is not to say that thecommunity was entirely united in this perception.Some community members likened the disease to small-pox (kawumpuli). But the belief that the disease was notEbola was corroborated by logical observations madeby the majority of the Kakute community and affirmed byKaguugo locals. Crucially, the deaths occurred only in onefamily, and specifically affected people who the motherdid not like. This despite the large number of people whohad been in physical contact with the first patient, includ-ing the traditional healer. It was simply seen as unbeliev-able that a very infectious disease would not affect any ofthese other people. Furthermore, Kakute villagers arguedthat while health workers stated that Ebola causes victimsto bleed from every pore and die within 24 hours, pa-tients in Kakute lived for at least a week, did notbleed from every pore, and did not always suffer fromhigh temperatures. In addition, it was locally knownthat previous Ebola outbreaks, in particular that inKibaale the same year, had also primarily affectedmembers of the same family (13 out of 17) [22].Luwero residents further argued that the public healthauthorities had not connected the case of those whodied in Kibaale with “eating infected animals”, awarning which figured prominently in the informationcirculated to the community, and there was no evi-dence that Kabugo or any of the other victims inLuwero had eaten wild animals such as monkeysbefore falling ill.To confirm whether this disease was indeed Ebola,

community members monitored one female neighborwho had interacted heavily with the diseased victims,as her house was surrounded by those of the family

involved. Community members said that “even if shehad just got a headache they would all scamper forsafety, knowing that it was indeed Ebola.” This casewas a yardstick for the community, and because shedid not suffer any health problems, community atti-tudes did not change. The woman herself believedthat God had saved her. Locals began to call for trad-itional healers to come and investigate exactly whathad attacked the community. As a Kakute communitymember noted:

Even when the health workers came, the peopleasked them if they had considered investigating ifthese could be spiritually instigated attacks. Theysuggested that since we had used side A [medicalapproach] and people continued to die, we shouldswitch to side B [traditional healers] to see if theproblem could be properly diagnosed. But this wasnot done because the authorities insisted thatpeople should accept that this was Ebola.

Community members asked the district chairman tosend traditional healers to make their assessment ofthe situation. But this was stopped by police intervention,to prevent revenge killings of those villagers whomight be identified as spreading amayembe. A Kakuteresident noted:

The head of Police in the district said that he wouldarrest anyone who promoted the idea that the deathswere due to amayembe, and threatened to take sternaction if anything happened to those suspected ofbringing amayembe. He read out some names ofthose who had been listed as promoting this thinkingamong the community, and people got scared ofdiscussing it much.

By the time the health authorities ruled that peopleshould avoid attending large gatherings, the communitywas still preoccupied with ways to establish whetheramayembe was to blame. Gatherings continued withpublic testimonies, one of which involved a daughterfrom the affected family “testifying” that her motherhad amayembe. Community members said that accord-ing to the daughter, the spirits had “asked for the livesof seven family members and thirty from outside thefamily after the seven had died.” This led to increasedfears within the community. One resident explainedthat he fled the village when the seven Ebola cases wereabout to be reached, for fear that he would be amongthe thirty who would subsequently be sacrificed fromoutside the family. Ironically, because people did not con-tinue to die, this discourse strengthened local convictionthat it was indeed, after all, amayembe, and not Ebola.

de Vries et al. BMC Public Health (2016) 16:161 Page 6 of 12

The community as barrier to outbreak controlWhen Ebola outreach workers tried to provide reliefitems to the children of the husband of one of thefemale Ebola victims, he refused them entry unless theyadmitted that his wife died not of Ebola, but becauseof amayembe. Such resistance to medical interpret-ation was widespread. The IFRC evaluation reportnoted a “strong negative community reaction”, whicheased over time, but did not disappear. One commu-nity member belonging to the social mobilizationteam noted:

It was like talking to mad people. They were angry,with some of them totally rejecting the message, whilesome agreed, but we did not really change thethinking of many that it was not amayembe. Generallythe outbreak did not change the cultural beliefs of thepeople, and the people were not scared to talk tomedical personnel who had touched patients.

One of the surviving Ebola victims described how hewas returned to Kakute village in an ambulance becausethe doctors wanted the villagers to see that the survivorshad indeed been treated at the hospital, and not bytraditional healers. But community members shoutedthat the survivors had bribed the medics to say it wasnot amayembe and they argued that he had survivedEbola by escaping from Mulago to visit a traditionalhealer back home.Eventually, the community’s resistance to biomedical

interpretation was met with government oppression. AKaguugo district community leader told us that whilecommunities were disputing health workers’ claimsabout Ebola, district level government officials arrivedwith armed law enforcement to convey the seriousnessof the issue. One Kakute respondent noted:

They warned that anyone who continued insistingthat this was not Ebola would be dealt withaccording to the law. So that is why people startedto listen quietly and behave [laughs]. So we neededthat kind of authority to sit and listen to the healthworkers. We attended those sensitization sessionsthrough the barrel of a gun. Before the governmentused force, people would come to the sensitizationsessions, argue with the health workers and disrupttheir sessions. But when the guns came in they hadto sit and listen.

Even the youths, “the most critical audience within thevillage” , were forced to attend and listen, but it isunlikely that this strategy changed their minds. Rather, itreinforced historical mistrust of a chronically under-funded medical system.

Community members themselves provided hints inour interviews of how trust could have been built if theapproach had been different. They argued that muchcould have been achieved had the response been speed-ier, with a concerted effort showing the community thatthe medical system actually cared for their thoughts andwishes. After confirmation of Ebola, our respondentsnoted that social mobilization came too slowly to coun-ter the growing amayembe interpretation effectively.This delay in outreach had serious consequences. As aKakute villager noted:

But for us when Ebola was suspected we keptexpecting the health teams, but they did not turn upuntil a week later. So we thought that they did notcare about us because for a whole week after theoutbreak there was no health personnel to inform thepeople on how to go about it…

She argued that this perceived lack of concern andinformation from health workers escalated the idea thatit was amayembe: “So I think if they had come earlier thehealth workers would have had less difficulty convincingthe community about Ebola.”The health system may have been slow to respond for

several reasons. There is the practical challenge of get-ting samples, confirming lab results, and assembling andcoordinating the response team and necessary logistics.One of our health worker respondents noted that thereappeared to be conflict among the leadership over whowas in charge. She also said it was difficult to persuadeher colleagues to take the disease seriously:

I really felt that they were not serious. At that time wethought that the disease was Marburg and were veryworried and anxious to know the results. Ourcolleagues from other sections had begun to distancethemselves from us. We needed the results to knowwhat we were dealing with, but it took them awhole week to give us the results after they hadtaken the samples. They took them on Thursdayand told us on Wednesday. It did not please usthat they were that lazy.

Respondents indicated that these delays were probablythe result of insufficient funds to transport materials.However, the IFRC evaluation report noted that nobodywanted to drive the volunteers or lend them their trans-port, because outreach workers were seen as potentialcarriers, and people feared being associated with witch-craft [18]. When social mobilization did eventually start,the 20 volunteers sent door-to-door by the IFRC lackedsufficient material support, and experienced considerableanxiety regarding their own family contexts. In addition,

de Vries et al. BMC Public Health (2016) 16:161 Page 7 of 12

they were not well acquainted with the wishes of thecommunity. For example, they advised communities notto shake hands but instead to knock fist (kubongo), apractice locally associated with drug users and marginal-ized people. Moreover, the surveillance team appeared ill-trained in methods to assess the logic and needs of thelocal community. For example, Kabugo’s family neighbour,the same woman who had been the local yardstick forwhether the disease was Ebola or amayembe (see above),was denied testing. As she herself said:

A team of health workers came to our village and tookblood samples from selected people who were on theirlist. I told them to take my blood too, but they insistedthat they were taking samples only from those whowere on their list. My neighbour here was also tested.

While the interview does not tell us who made up thislist, or what the criteria for selection were, ignoring thecommunal yardstick for Ebola illustrates a lack of sensi-tivity to the value of local knowledge and relationships.The IFRC evaluation report is strikingly focused on thepatients and the medical system. When suspected caseswere found, their isolated and even starving conditionswere exclusively attributed to witchcraft. The reportimplied that it was witchcraft which caused “normal”social support elements (e.g. bringing food and water) atbay. When volunteers were asked about their motivationto volunteer “in such a setting”, the report notes that theyfelt it was “good to have the feeling that you were saving alife and keeping people safe by sensitizing the communitywho otherwise might have acted as far as killing thepatients” [19] (P19). Focused mostly on the health ofEbola patients, volunteers, and health workers, the issueof community stigmatization or marginalization was ad-dressed only at the level of economic marginalization.This discourse of framing witchcraft exclusively as amedical risk is also present in the accompanying IFRCstrategy report, which explicitly noted that the diseaserequired – through “culturally acceptable” interventions– “total behavior change from the community members(they still believe they are being attacked by witchcraftnot Ebola) if the outbreak is to be controlled” [18] (p4).While biomedically accurate, this approach clearlyfailed to connect to local community health ideologies.The discourse in the reports is against witchcraft – andby extension, the community. This view that the com-munity either needed change or was a barrier, insteadof a necessary ally, appeared to be promoted in themedia as well. For example, in an article on the UgandaRadio Network, the Luwero District Health Educatorwas quoted as having said that witchcraft allegationswere “dangerous because they water down efforts tocontain the deadly Ebola virus” [23].

Mistrust of the health systemKaguugo residents told us that the sensitization meet-ings failed to change their minds, and their interviewnarratives suggest that any contradiction in the meetingswas seen as further evidence of a corrupt health system.For example, it had been announced on the radio thatthe government had released Ush600 million (approx.USD 175,900) to assist areas affected by Ebola, but localresidents noted that they did not see any of this money.In Kakute, a respondent noted:

For us, we told them that we do not have Ebola, butthey kept coming and we concluded that the healthworkers just wanted to eat the money that they hadbeen given [laughs]. Yes. The truth needs to be toldsometimes. The health workers came just forceremonial purposes.

In a focus group held as part of sensitization meetingsat Luwero town council during the outbreak, one of therespondents noted cynically that officials taught themhow to prevent the spread of Ebola, but did not provideany water to wash their hands. Kaguugo communitymembers asked why there been so little help from relieforganizations despite calls for help by school heads. Itwas also noted more than once that schools were givenjust half a bottle of disinfectant to clean the entireschool, and that no health camp had been set up nearthe community. And, as one focus group respondentpointed out, “These health workers would tell us nottouch our patients who are ill, and at the same timeexpect us to deliver them to the health centers!?”Many statements exemplify a sentiment of marginalization.

People simply believed that while equipment and drugswere available, they were not helped because they werepoor and unimportant. These words appear to have sometruth in them. The surviving brother noted that some arti-cles allocated to him and his brother-in-law at Mulagohospital – including new clothes and food items – neverreached them. He discovered about these allocationsthrough the health workers, who confirmed, on beingqueried, that they were supposed to have receivedthese items.Many locals remembered using an effective health care

system in the past, but explained that they had becomehostile to the system as the result of disillusionmentwith competition and bribery in government hospitals.To community members, this bribery was alsoentrenched in the much heralded community healthworker system (VHTs) that consisted mostly of betteroff community members, typically including villagechairs. Each village could pick one member to be sentfor DTF training in Kampala, but community membersargued that PLAN International would only select from

de Vries et al. BMC Public Health (2016) 16:161 Page 8 of 12

the village leadership because they would hand out feesand thus control the agenda. Many conversations in of-fices and verandahs led to the conclusion that Ebola wasan overblown conspiracy, developed to justify the allow-ances of medical personnel. Medical cases that seemedto counter the amayembe interpretation, like the sur-vival of Kabugo’s brother and brother-in-law, were ex-plained as the result of the corrupt health system itself.For rexample, according to the brother, some peoplethought that he and his mother-in-law had become richas a result of the ordeal, as the radio had announced thatthe government of Uganda had given each Ebola patientthree million Uganda shillings (approx. $860), whichthey were accused of pocketing.While the community mistrusted the health sys-

tem’s agenda, it also simultaneously experiencedstrong stigmatization from the outside world duringthis period, suffering serious economic consequences.As everyone was seen as contagious, the area hadbecome a “danger zone”. A Kakute respondent said:

The discrimination was such that wherever we went,people would say that the people of Kakute villagehave Ebola. Even the people of Ndejje [a nearbycommunity] did not want to pass through our roadsaying that there was Ebola at Kakute. You see? Andeven at the hospitals, the Kakute people were lookedat with suspicion. They associated us with Ebola. Thepregnant women who went for antenatal clinicssuffered most, as the nurses told them to wait or sitaside as they treated others first.

As the idea of the contagion spread, so did thegeographic isolation. In a sense, community memberswere right about amayembe contagion; now the entireLuwero area had become a danger zone of culturalstigma. Even health workers working within this com-munity, helping the victims, were shunned by the public.

The crisis of careLocal aetiologies, a discourse of community as barrier,and mistrust of the health system were major factors de-termining resistance, but it is equally, if not more, im-portant to recognize that underlying this resistance wasa local need for care. This need reached crisis propor-tions. A Kaguugo respondent explained:

My heart is not settled because I am used to otherdiseases that can allow your people to take care ofyou… But for this disease, you cannot take care ofyour people even when they are your father, yourmother, your child or even your husband. You justrun away from them! So the main question is ‘whatshould we do?’

Another Kakute resident noted that the inability tohelp a parent or someone close when in need “mayhaunt you forever”:

First of all she is my mother, and you are saying thattouching her will also make me sick. So I think theyhave scared us too much, to the extent that one cannoteven show love to their loved ones. In fact your mothercan die cursing you if you don’t care for her when she issick. She can say ‘my child refused to care for me whenI was sick’ yet you haven’t really refused.

If one’s child catches Ebola, should a parent not touchto see what is happening? Would a parent run awayfrom a child to save their own life? Should a neighbornot help the same people that helped her the last timeshe fell ill? The dilemma facing community memberswas that complying with quarantine requirements alsointroduced the risk of no longer being able to take careof loved ones:

I think they should calm down and tone down thescare. We know that the disease kills. I think theyshould also distribute the items which they tell peopleto buy like gloves, gumboots and soap. With thoseitems people can take care of their relatives whodevelop the symptoms and even visit them at theEbola clinic.

The Ebola control team collected all those who hadbeen close to Ebola victims to monitor them in an Ebolaisolation center. The IFRC evaluation describes how sus-pected cases –children or adults – were whisked away tohealth facilities, where they arrived potentially frightenedand disoriented, “taken out of the car and sprayed withdisinfectant and their possessions burnt in front of theireyes” (p15) [19]. Luwero respondents noted that withinthe emergency facilities, water and electricity were notavailable at first, while another prominent issue was thelack of food to sustain patients. According to a regis-tered nurse working at the health center, local commu-nity members were generally not allowed to attend totheir kin inside this center other than a brief glance.This inability to provide care was present not only in

terms of attendance, but also at communal level duringburial rites. During a focus group with community mem-bers, the experience of burials appears to have been themost traumatic and visible violation of their communityidentity, leading to several incidents, involving the localNubian community in particular: “They just dump themlike dogs and return running fast from the burial groundlike it is done for those who have committed suicide; that’sthe way they do it.” There were also problems with caseswhich may not have been Ebola-related, but were also

de Vries et al. BMC Public Health (2016) 16:161 Page 9 of 12

hurriedly dispatched. As the surviving brother noted,these disrespectful practices were relatively unnecessary:

For us Muslims, we have a special cloth to wrap thebodies in and the Christians use bark cloth. I think thatif they didn’t want us to touch the bodies they wouldhave asked us to buy these items so that our loved onesare wrapped and buried with the respect and in theclothes that we traditionally use for burial. They alsoshould not just dump them. I saw them personallyswing the body of a person and dump it on the vehicle.

According to our main medical informant, socialmobilization “made them appreciate the reasons as towhy we wanted the burial conducted that way.”Claiming that the subsequent burials were then han-dled in the right fashion, the respondent also notedin the same interview that the leaders within theburial group were not Muslims, instead referring tothe affected, discordant group as “fanatics.” Overallthe community felt that they had been made irrele-vant in the struggle against Ebola. It appeared thatone of the few places that had remained unaffectedby the Ebola crisis was the local church, where resi-dents were still allowed to convene. Ironically, thesame spirituality that was heavily denied by author-ities when it came in the form of amayembe, washere allowed to provide spiritual guidance and care.As a Kakute resident summarized it:

They told the people not to shake hands, not toconverge, and schools were even closed because theydid not want us to be in groups. But we used to go tothe churches and mosques. These were allowed. Wewent there because God is the one who can deal withany situation including Ebola. So there we had nooption but to seek God’s intervention.

ConclusionsTo solve the first mile problem, public health managersare quick to involve community health workers andsocial mobilization teams, ideally equipping them withsmartphones to map the outbreak in real time. Thisstudy shows that in a major Ebola outbreak control ef-fort in central Uganda in 2012, these interventionsproved far from sufficient to assure collaboration fromthe community. Instead, the “people-centered approach”reaffirmed a historical mistrust between “scientific” bio-medical and local explanations of illness and misfortune.This theme goes back as far as Evans-Pritchard’s Witch-craft, Oracles and Magic among the Azande (1937) [24]and is the subject of a truly immense body of anthropo-logical literature. It illustrates the lack of credibility(from an emic perspective) of biomedical explanations

that ignore local understandings, combined with the rela-tively logical and empirical basis of local explanations.While biomedicine offered no clear evidence of Ebola

as an infectious pathology, and no explanation as towhy this outbreak occurred suddenly in this particularlocation and affected these particular individuals, theemic explanatory model involving amayembe was easilyable to account for this by showing that the deaths hadoccurred in only one family, and that they involvedpeople whom Ruth – who was accused of using theamayembe to bewitch them – did not like. Moreover,many people had had contact with the first patient,Kabugo, during his initial illness and during his funeral,but had not become infected. From an emic perspec-tive, the evidence was clear.While biomedical information appeared to be emitted

into the community in relatively arbitrary and discretebursts, these local interpretations were continuous andbased on a sort of collective and cumulative empiricism,with local experiences, gossip and discussion feedinginto and supporting an increasingly clear picture of whatwas “really” going on. People were proactive in collectingevidence to support their suspicions, for example bymonitoring the woman who lived close to the Ebolavictims and interacted with them, to see whether shealso became infected.The credibility of biomedical explanations was further

damaged by the lack of fit between the general informa-tion that was circulated and the local situation on theground. For example, the advice to avoid wild monkeymeat might have been based on campaigns in areaswhere previous outbreaks occurred, and where this prac-tice was common, but in an area where people did noteat monkey (and where there aren’t even monkeys tospeak of ) it did not inspire confidence in the veracity ofthe information. Similarly, the advice to use gloves andgoggles when dealing with Ebola victims was bothimpractical (people have no access to these resources)and socially and culturally unacceptable as a way ofcaring for both sick and deceased relatives.Although lack of funding, training and capacity ap-

peared to have played a role, what characterizes thiscase study is the shared experience of communitymembers that the DTF and its partners lacked open-ness and willingness to take the time needed to listento and empathize with community needs. After all,what these community members were seeking firstand foremost were strategies to take care of theirfellow human beings and to diagnose and explainwithin their own cultural framework. The resistanceto openness may be related to a pre-existing assump-tion that openness to other interpretations is thesame as “accepting” them, and would automaticallylead to a lack of control.

de Vries et al. BMC Public Health (2016) 16:161 Page 10 of 12

We acknowledge that the amayembe paradigm is diffi-cult to reconcile with the biomedical view. Yet, ideasabout amayembe do not necessarily exclude prompttreatment-seeking, a finding corroborated elsewhere[25]. Local practices and aetiology did motivate isolationand social control of those who were most likely to becarriers of the disease, because they were also, to anextent, the ones suspected of using the amayembe.Moreover, the interpretation of amayembe is not asuniform and static as might appear at first sight. Rather,we document a community discussion in which there isroom for interpretation, doubt, and different viewswhich may be held simultaneously. The existence of alocal medical system characterized by resilience and syn-cretism is an observation supported by well-establishedanthropological facts [26]. It seems, therefore, counter-productive to deny this process.Ironically, as the village and Luwero district became

increasingly stigmatized and isolated by the outsideworld, the spread of amayembe from one family out intothe community and district became its own case ofsocial contagion that mimicked Ebola, but rested ondifferent actors and different explanatory frameworks.While the outbreak in Luwero was ultimately controlledby an authoritative stance, one might wonder if this washelpful only in the short-term, as reemergence of thedisease may only reinforce a divide with long historicalroots. To be truly effective in bridging the first mileoutbreak control, we believe respect and engagement areneeded first and foremost, and despite remarkabletechnological innovations including real-time commu-nity information systems, these interventions remaincontingent upon human interaction and openness tocultural difference. To develop a more “people-centered”early detection system, collaboration with the commu-nity must be at its core, and ideally the communityshould be seen as a resource instead of a barrier. Toachieve this, the big missing element is knowing how torespectfully “agree to disagree” while simultaneouslyfinding commonalities in humanity to facilitate collabor-ation during the first mile of outbreak control.Given the mass of anthropological literature on trad-

itional African aetiologies, and anthropological knowledgeon a previous Ebola outbreak in Northern Uganda [16], itis surprising that so little serious effort was made to takelocal sensibilities and culture into account. It beggarsbelief, given this knowledge base and the extent of com-monality in local aetiologies across sub-Saharan Africa,that this was largely ignored in the early response to amajor epidemic in West Africa. Consequently, the “firstmile” problem is not only a question of using local re-sources for early detection, but also of making use, in pub-lic health outreach, of the contextual cultural knowledgethat has already been collected and is readily available.

These observations underscore the crucial importanceof initiatives such as the Ebola Response AnthropologyPlatform1 aimed at providing access to regional and inter-national rapid-response anthropological experts whocould advise health authorities and governments as soonas an outbreak is suspected. This could also apply to otherinfectious diseases, because the underlying principles arethe same. Another large body of literature which is rele-vant for explaining the distrust of the formal healthsystem, and which also tends to be ignored when preven-tion campaigns do not work, is that which exploresrumours and local distrust of “outsiders” perceived to bepowerful and a potential threat to local interests. Thesemay be “foreign” medical researchers thought to be steal-ing local people’s blood under the cover of clinical trials,or national health authorities secretly sterilizing women inthe guise of immunisation campaigns, or disinformationcampaigns deliberately intended to mislead people aboutthe “real” cause of a disease [27, 28].

Endnotes1http://www.ebola-anthropology.net/

AbbreviationsDTF: District task force; IFRC: red cross and red crescent; VHFs: viralhemorrhagic fevers; VHT: village health team.

Competing interestsThe authors declare that they have no competing interests.

Authors’ contributionsDDV analyzed the data and drafted the manuscript. JR, LM and TB carriedout fieldwork and provided input in the analysis. DM, MG and DKMsupported fieldwork and analysis. RP conceived the larger research project,participated in its design and coordination, and helped to draft themanuscript. All authors read and approved the final manuscript.

AcknowledgementsWe acknowledge the people of Luwero who graciously provided their timeand input during the period of study. This study has been made possiblethrough a Global Health Policy and Health Systems Research grant by theDutch Scientific Organization (NWO) called “Developing SustainableCommunity Health Resources in Poor Settings in Uganda” (CoHeRe).

Author details1Department of Anthropology, University of Amsterdam, Amsterdam, TheNetherlands. 2Makerere University College of Humanities and Social Sciences,CeSSRA office, Kampala, Uganda. 3Department of Social Work and SocialAdministration, Makerere University, Kampala, Uganda. 4Center for SocialScience Research on AIDS (CeSSRA), School of Social Sciences, MakerereUniversity, Kampala, Uganda.

Received: 10 June 2015 Accepted: 10 February 2016

References1. Frieden TR, Damon I, Bell BP, Kenyon T, Nichol S. Ebola 2014 — New

Challenges, New Global Response and Responsibility. N Engl J Med.2014;371:1177–80.

2. Nguyen V. Ebola: How We Became Unprepared, and What Might ComeNext. Cultural Anthropology Website. 2014;October 07. Available: http://www.culanth.org/fieldsights/605-ebola-how-we-became-unprepared-and-what-might-come-next.

de Vries et al. BMC Public Health (2016) 16:161 Page 11 of 12

3. Ebola Response Team WHO. Ebola Virus Disease in West Africa - TheFirst 9 Months of the Epidemic and Forward Projections. N EnglJ Med. 2014;371:1481–95.

4. Dhillon RS, Srikrishna D, Sachs J. Controlling Ebola: next steps. Lancet.2014;384:1409–1411.

5. Taubenback H, Goseberg N, Lammel G, Setiadi N, Schlurmann T, Nagel K,et al. Risk reduction at the Last-Mile: An attempt to turn science into actionby the example of Padang, Indonesia. Nat Hazards. 2013;65:915–45.

6. Collins A. Early warning: a people-centered approach to early warningsystems and the 'last mile'. In: Anonymous World Disasters Report.Switzerland: International Federation of Red Cross and Red CrescentSocieties; 2009. pp. 39–67

7. Mandela W. Uganda roots for electronic early-warning. East AfricanBusiness Week. Available: http://www.busiweek.com/index1.php?Ctp=2&pI=2175&pLv=3&srI=68&spI=107&cI=10. AccessedNovember 30, 2015.

8. Crigler L, Glenton C, Hodgings S, LeBan K, Lewin S, Perry H. Developing andstrengthening community health worker programs at scale. A ReferenceGuide for Program Managers and Policy Makers. Jhpiego Corporation,Report. 2013. Available: http://www.mchip.net/sites/default/files/mchipfiles/CHW_ReferenceGuide_sm.pdf.

9. Attiah K. The world is still far from the finish line on Ebola. WashingtonPost. Available: http://www.washingtonpost.com/blogs/post-partisan/wp/2014/11/06/the-world-is-still-far-from-the-finish-line-on-ebola/. AccessedNovember 30, 2015.

10. World Health Organization. Involving everyone: social mobilization is key inan Ebola outbreak response. Features. 2014. Available: http://www.who.int/features/2014/social-mobilisation/en/#. Accessed November 30, 2015.

11. Fowler RA, Fletcher T, Fischer WA, Lamontagne F, Jacob S, Brett-Major D,et al. Caring for Critically Ill Patients with Ebola Virus Disease. Perspectivesfrom West Africa. Am J Respir Crit Care Med. 2014;190:733–7.

12. Bah SM, Aljoudi AS. Taking a religious perspective to contain Ebola.2014;384: 951

13. Sayegh J. Ebola and the Health Care Crisis in Liberia. Cultural AnthropologyWebsite. 2014;October 07. Available: http://www.culanth.org/fieldsights/595-ebola-and-the-health-care-crisis-in-liberia.

14. Piot P, Muyembe J, Edmunds WJ. Ebola in west Africa: from diseaseoutbreak to humanitarian crisis. 2014;14: 1034–1035

15. Wayland C, Crowder J. Disparate Views of Community in Primary HealthCare: Understanding How Perceptions Influence Success. Med AnthropolQ. 2002;16:230–47.

16. Hewlett BS, Hewlett BL. Ebola, Culture and Politics: The Anthropology of anEmerging Disease. Belmont: Wadsworth, Cengage Learning; 2008.

17. Brown H, Kelly AH. Material Proximities and Hotspots: Toward an Anthropologyof Viral Hemorrhagic Fevers. Med Anthropol Q. 2014;28:280–303.

18. International Federation of Red Cross and Red Crescent. Uganda: EbolaOutbreak. Disaster Relief Emergency Fund (DREF). Disaster EmergencyRepair Fund. 2012;EP-2012-000195-UGA.

19. Thormar SB. Evaluation of Ebola response – Uganda. Report. Ugandan RedCross Society, Kampala. 2013. Available: http://adore.ifrc.org/Download.aspx?FileId=42478&.pdf.

20. Robertson AF. Greed: Gut Feelings, Growth, and History. Oxford: Blackwell; 2001.21. Geschiere P. The Modernity of Witchcraft, Politics and the Occult in

Postcolonial Africa. Charlottesville, VA: University of Virginia Press; 1997.22. World Health Organization. End of Ebola outbreak in Uganda. Global Alert

and Response (GAR) Disease Outbreak News (DONs). 2012.23. Luwaga B. Witchcraft Claims Hamper Ebola Fight in Luweero. Uganda Radio

Network. Available: http://ugandaradionetwork.com/a/story.php?s=47494.Accessed November 30 2015.

24. Evans-Pritchard EE. Witchcraft, Oracles and Magic among the Azande.Oxford: Clarendon Press; 1937.

25. Ackumey MM, Gyapong M, Pappoe M, Maclean CK, Weiss MG. Socio-culturaldeterminants of timely and delayed treatment of Buruli ulcer: Implicationsfor disease control. 2012;1: 6–6.

26. Stoner BP. Understanding Medical Systems: Traditional, Modern, andSyncretic Health Care Alternatives in Medically Plurastic Societies. MedAnthropol Q. 1986;17:44–8.

27. White L. Speaking with Vampires: Rumor and History in Colonial Africa.Oakland, CA: University of California Press; 2000.

28. Geissler PW, Pool R. Editorial: Popular concerns about medical researchprojects in sub-Saharan Africa? A critical voice in debates about medicalresearch ethics. 2006;11: 975–982

• We accept pre-submission inquiries

• Our selector tool helps you to find the most relevant journal

• We provide round the clock customer support

• Convenient online submission

• Thorough peer review

• Inclusion in PubMed and all major indexing services

• Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step:

de Vries et al. BMC Public Health (2016) 16:161 Page 12 of 12