Institutional trust and misinformation in the response to the 2018–19 Ebola outbreak in North Kivu, DR Congo: a population-based survey
Affiliations & Notes
aHarvard Medical School, Harvard University, Cambridge, MA, USA
bCenter for Research on Democracy and Development in Arica, Free University of the Great Lakes Countries in the Democratic Republic of the Congo, Goma, DR Congo
cAnthrologica, Oxford, UK
Article Info
Publication History:
Published March 27, 2019
DOI: 10.1016/S1473-3099(19)30063-5 External LinkAlso available on ScienceDirect External Link
Copyright: © 2019 Elsevier Ltd. All rights reserved.
Linked Articles (2)
- Ebola: public trust, intermediaries, and rumour in the DR CongoThe Lancet Infectious DiseasesMarch 27, 2019
- Acknowledging the limits of public health solutionsThe LancetMay 25, 2019
Download started
OkIntroduction
On Aug 1, 2018, the DR Congo declared its tenth outbreak of Ebola virus disease (EVD). Responding to EVD outbreaks entails a multifaceted control strategy that includes the early detection, active finding and isolation of suspected cases, identification and tracing of contacts, as well as risk communication, specialised medical care, support for safe burial practices, and vaccination for individuals at high risk.1,2 These measures are particularly challenging in an active conflict zone.3 They assume some level of public cooperation to report suspect cases, confidence in response workers, and freedom of movement for public health teams.
1.
Faye, O ∙ Boëlle, PY ∙ Heleze, E ∙ et al.
Chains of transmission and control of Ebola virus disease in Conakry, Guinea, in 2014: an observational study
Lancet Infect Dis. 2015; 15:320-326
2.
Walldorf, JA ∙ Cloessner, EA ∙ Hyde, TB ∙ et al.
Considerations for use of Ebola vaccine during an emergency response
Vaccine. 2017;
published online Sept 7. DOI:10.1016/j.vaccine.2017.08.058
3.
Nakkazi, E
DR Congo Ebola virus outbreak: responding in a conflict zone
Lancet. 2018; 392:623
Eastern DR Congo has had violent conflicts for over two decades, leading to the protracted presence of humanitarian agencies. Over the past 4 years, the epicentre of the current EVD outbreak around the city of Beni has endured recurring violence resulting in thousands of deaths and injured people.4 Such violence limits the free movement of the population and health workers and leads to the collapse of trust and social capital.5–7 Longitudinal data from the region show that the perception of security and trust in government, security, and humanitarian workers has been declining, complicating what is already a complex and challenging EVD response operation.8 Community trust and support for the EVD response is undermined by misinformation exploited by local politicians9 and the contrast between the rapid mobilisation to contain the EVD outbreak and chronic failure to protect civilians.10 Violent incidents, including targeted attacks on response teams, led to the temporary suspension of EVD response activities.11 These activities include important local engagement efforts by the International Federation of Red Cross and Red Crescent Societies, UNICEF, WHO, and others to dispel misinformation and provide crucial information and services for prevention and care.
4.
Congo Research Group
Mass killings in Beni territory: political violence, cover ups, and cooptation
Center on International Cooperation, New York, 2017
5.
Colletta, NJ ∙ Cullen, ML
Violent conflict and the transformation of social capital: lessons from Cambodia
World Bank Publications, Rwanda, Guatemala, and Somalia, 2000
http://documents.worldbank.org/curated/en/799651468760532921/pdf/multi-page.pdf
Date accessed: February 28, 2019
6.
De Juan, A ∙ Pierskalla, JH
Civil war violence and political trust: microlevel evidence from Nepal
Confl Manag Peace Sci. 2016; 33:67-88
7.
Rohner, D ∙ Thoenig, M ∙ Zilibotti, F
Seeds of distrust: conflict in Uganda
J Econ Growth. 2013; 18:217-252
8.
Vinck, P ∙ Pham, PN ∙ Makoond, A ∙ et al.
Voices from Congo Report #13. Harvard Humanitarian Initiative
http://www.peacebuildingdata.org/sites/m/pdf/DRC_Poll13_english_final.pdf
Date: 2018
Date accessed: February 28, 2019
9.
Weber, L
WHO fears Ebola in DRC will spin out of control due to political violence. Huffington Post
https://www.huffingtonpost.com/entry/ebola-drc-congo-perfect-storm-who_us_5baa6616e4b0f143d10ddff9
Date: Sept 25, 2018
Date accessed: February 28, 2019
10.
Murphy, R
UN peacekeeping in the Democratic Republic of the Congo and the protection of civilians
J Conflict Secur Law. 2016; 21:209-246
11.
DR Congo EVD outbreak: Beni attack halts outreach work. BBC News
The role of public trust is recognised in medicine and public health. Trust is central to the legitimacy of health systems.12 Trust and legitimacy have been associated with acceptance of and compliance with preventive or curative interventions, including vaccine acceptance and changes in individual behaviours to reduce risk.13–19 Public trust in the health-care system, the understanding of EVD transmission risks, perceptions of and experiences with EVD survivors, and long-term effects of violence also appear to be associated with compliance.20,21 Fear and the perception of risk might also influence the adoption of EVD preventive behaviours, as shown during the 2014 west Africa outbreak.22 Despite this evidence, the association between institutional trust and individual-level responses to outbreaks during conflict is underexplored, especially concerning EVD.
12.
Gilson, L
Trust and the development of health care as a social institution
Soc Sci Med. 2003; 56:1453-1468
13.
O'Malley, AS ∙ Sheppard, VB ∙ Schwartz, M ∙ et al.
The role of trust in use of preventive services among low-income African-American women
Prev Med. 2004; 38:777-785
14.
Mohseni, M ∙ Lindstrom, M
Social capital, trust in the health-care system and self-rated health: the role of access to health care in a population-based study
Soc Sci Med. 2007; 64:1373-1383
15.
Meredith, LS ∙ Eisenman, DP ∙ Rhodes, H ∙ et al.
Trust influences response to public health messages during a bioterrorist event
J Health Commun. 2007; 12:217-232
16.
Goold, SD
Trust, distrust and trustworthiness
J Gen Intern Med. 2002; 17:79-81
17.
Leavitt, JW
Public resistance or cooperation? A tale of smallpox in two cities
Biosecur Bioterror. 2003; 1:185-192
18.
Mechanic, D ∙ Schlesinger, M
The impact of managed care on patients' trust in medical care and their physicians
JAMA. 1996; 275:1693-1697
19.
Yaqub, O ∙ Castle-Clarke, S ∙ Sevdalis, N ∙ et al.
Attitudes to vaccination: a critical review
Soc Sci Med. 2014; 112:1
20.
Nuriddin, A ∙ Jalloh, MF ∙ Meyer, E ∙ et al.
Trust, fear, stigma and disruptions: community perceptions and experiences during periods of low but ongoing transmission of Ebola virus disease in Sierra Leone, 2015
BMJ Glob Health. 2018; 3:e000410
21.
Betancourt, TS ∙ Brennan, RT ∙ Vinck, P ∙ et al.
Associations between mental health and Ebola-related health behaviors: a regionally representative cross-sectional survey in post-conflict Sierra Leone
PLoS Med. 2016; 13:e1002073
22.
Tenkorang, EY
Effect of knowledge and perceptions of risks on Ebola-preventive behaviours in Ghana
Int Health. 2018; 10:202-210
Research in context
Evidence before this study
We searched PubMed and Google Scholar for publications in English and French published from Jan 1, 1950, to Aug 20, 2018, using various combinations of the terms “trust”, “Ebola outbreaks”, “behaviour change”, “protective behaviours”, and “preventive behaviours”. We obtained few studies and broadened the search to include other epidemic-prone diseases, using the terms “cholera”, “malaria”, “Zika”, “HIV-AIDS”, and “infectious diseases”. The relevant published studies suggested that mistrust and misinformation are obstacles to public health interventions. However, most studies were qualitative and few had attempted to rigorously characterise and quantify these issues, and only one had done so during an Ebola outbreak. No previous studies have attempted to define these issues in the context of active conflict.
Added value of this study
We present the most comprehensive study of trust and misinformation and their association with preventive behaviours during an outbreak of Ebola virus disease in an active conflict environment. Our data indicate that low institutional trust and belief in misinformation about Ebola are inversely associated with preventive behaviours on an individual level. This study more precisely defines the socioanthropological factors that are important for outbreak control, which provides evidence to guide prioritisation of response activities.
Implications of all the available evidence
Trust and the circulation of accurate information by reliable sources are crucial to control Ebola outbreaks and pose a major challenge in conflict environments. Despite substantial advances in the public health response to Ebola outbreaks in general, the precise mechanisms by which misinformation and mistrust undermine outbreak response are poorly understood, especially in conflict settings.
We aimed to explore two hypotheses: (1) whether institutional trust is associated with the adoption of preventive measures, including exposure avoidance and vaccination; and (2) whether belief in EVD misinformation is associated with lower adoption of preventive measures. This study builds on underdeveloped but crucial research examining misinformation about the motives or results of interventions as important obstacles to public health programmes.23
23.
Kaler, A
Health interventions and the persistence of rumour: the circulation of sterility stories in African public health campaigns
Soc Sci Med. 2009; 68:1711-1719
Methods
Study design and participants
We did a population-based survey 1 month after the EVD outbreak was declared (Aug 1, 2018), in the cities of Butembo (population 670 000) and Beni (230 000), in the province of North Kivu, eastern DR Congo (figure). As of Feb 25, 2019, WHO reported 875 probable and confirmed cases of EVD and 486 confirmed deaths due to the current outbreak, including 235 (26·9%) of those confirmed and probable cases in Beni and 78 (8·9%) in Butembo.
30063-5/asset/22a606a7-3b62-4e3e-b302-f1f7996d7db3/main.assets/gr1_lrg.jpg)
Figure Study sites
Respondents were selected using a multistage cluster sampling procedure (appendix). Out of 412 avenues and cells (the lowest administrative structures; 99 in Beni and 313 in Butembo), we randomly selected 30 in each city, using a list of all avenues and cells and with a systematic random approach, generating a random number then using a constant interval to select the next avenue or cell. In each avenue or cell, 16 households were selected using the WHO Expanded Programme on Immunization's random walk approach. In each household, we randomly selected one adult for interview. When first contacted, any individual that was present in the randomly selected household was asked how many adults (aged ≥18 years) lived in the household. Three attempts were made to contact the selected individuals over the course of 1 day. If the third attempt failed, interviewers proceeded to the next randomly selected respondent, first within the same household, or in the next household if no other eligible adult was available. Female interviewers selected eligible adult women, and male interviewers selected eligible adult men. Because interview teams comprised equal numbers of men and women, this method ensured sex-matching and equal sex representation in the sample.
Participation was anonymous, voluntary, and no compensation was provided. Interviewers were trained on preventive measures to reduce risk of exposure, and all households were provided with EVD information sheets in local languages at the end of the interview. The survey was approved by the Human Research Committee at Brigham and Women's Hospital (Boston, MA, USA) and a similar body that was convened by the Research Center on Democracy and Development in Africa, Free University of the Great Lakes Countries in the DR Congo (Goma, DR Congo).
Procedures
Interviews were done using a standardised, structured questionnaire developed by a team with experience in public health, medicine, anthropology, and EVD in DR Congo and west Africa. The questionnaire was developed in English and translated into French and Swahili. Independent experts reviewed and validated the translation. Local experts established face validity. Pilot interviews were done to test and validate the questionnaire. Questions covered demographics, measures of institutional trust, trust in the EVD response, exposure to EVD-related information, knowledge about and changes in behaviour due to the EVD outbreak, and health-seeking behaviour. Interviewers were experienced in community survey methods and participated in a 1-week training course on the questionnaire content and sampling protocol. The survey was done using tablet computers.
We computed several scores to combine the outputs of multiple questions covering similar topics. The government trust score, developed for longitudinal research in eastern DR Congo,8 was computed using questions related to community perceptions of how authorities represent the interest of the population at different administrative levels (neighbourhood, city, provincial government, and national government). Each of the four questions was scored from one to five with one reflecting the lowest level of trust. A cumulative score was computed and scaled from one to five.
8.
Vinck, P ∙ Pham, PN ∙ Makoond, A ∙ et al.
Voices from Congo Report #13. Harvard Humanitarian Initiative
http://www.peacebuildingdata.org/sites/m/pdf/DRC_Poll13_english_final.pdf
Date: 2018
Date accessed: February 28, 2019
A summative EVD information score quantified exposure to six specific categories of EVD information (prevention, symptoms, where to seek health care, what to do if a relative or neighbour has EVD, updates on EVD in the province, overall EVD response). An open-ended question assessed where, if at all, respondents would seek care if they suspected they had EVD. Formal health services included hospitals or health centres and informal settings included friends and traditional healers. Acceptance of EVD vaccine was assessed using two items: belief that the vaccine was safe and acceptance of the vaccine, if offered. The EVD risk-perception score was a measure of perception of personal risk of contracting Ebola in the month following the survey.
To understand the relationship between the intermediate variables of interest (trust, exposure to EVD information, and exposure and belief in misinformation) and the outcomes of interest (EVD avoidance or preventive behaviours), we explored changes in behaviour at the individual level that participants reported undertaking because of the EVD outbreak. Respondents were asked whether they engaged in a list of specific behaviours, as well as an open-ended question on changes in behaviour. A total of 12 behaviours were grouped into five categories: direct avoidance (three), social interaction avoidance (three), physical contact avoidance (two), public space avoidance (three), and hygiene (one). The items were identified in consultation with local and international anthropologists, risk communication specialists, and epidemiologists with expertise in EVD. Items were scored one and zero for change versus no change in reported behaviour.
Statistical analysis
The sample size was calculated to estimate proportions in the given population for a 95% CI and 10% precision. We used a 0·5 proportion estimate, for a target sample size of 96. The sample size was multiplied by four (384) to allow for comparison by sex within cities and account for an estimated design effect of two. The sample size was increased by 20% to account for non-response and further adjusted to reflect logistical constraints, resulting in a target sample size of 480 interviews per city (960 total).
We calculated frequencies, odds ratios (ORs), 95% CIs, and all other analyses using the complex sample module in SPSS (version 25). Data were weighted to reflect the unequal probability of sampling between the two cities. We built four separate stepwise binary logistic models with direct avoidance behaviour (all and any), formal health seeking, and EVD vaccine acceptance as the outcomes of interest. The outcomes represent key aspects of EVD response—behaviour change, care seeking, and vaccination. For the four models, the independent variables were location, age, sex, education level, wealth, ethnicity, the government trust score, government EVD trust score, health professionals EVD trust score, EVD information score, EVD risk perception score, and belief in rumours. The independent variables of greatest interest were the measures of trust and belief in rumours. For belief in rumours, statements were tested independently and in combination. One item (belief in the rumour that Ebola does not exist) entered the model. The selection of independent variables was guided by the literature review, measures of associations, and expert opinion.
Role of the funding source
The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Results
Between Sept 1, and Sept 16, 2018, 961 adults from 977 households that we approached were interviewed in Beni (480) and Butembo (481; appendix). Ten households declined, and no one was home at the time of visitation at six households. The mean age was 34·3 years (SD 14·0) and the median age was 31·0 years (IQR 22–42). As per study design, there was equal sex representation.
Overall trust in how administrative authorities represent the interests of the population was low and decreased from local, to city, to provincial, to national levels (table 1). When considering the EVD response specifically, 419 respondents trusted the government and 620 trusted health professionals to act in the best interests of the public (table 1). These items were used to compute the government trust score, government EVD trust score, and health professional EVD trust score (table 2).
| Unweighted (n) | Weighted (% [95% CI]) | ||
|---|---|---|---|
| Total | 961 | . | |
| Generalised government trust | |||
| Trust local authorities | 349 | 31·9% (27·4–36·9) | |
| Trust city authorities | 198 | 15·1% (11·9–19·0) | |
| Trust provincial authorities | 75 | 4·9% (3·6–6·7) | |
| Trust national authorities | 29 | 2·1% (1·2–3·4) | |
| Ebola-related trust | |||
| Trust government for Ebola response | 419 | 40·5% (36·8–44·3) | |
| Trust health professionals for Ebola response | 620 | 61·5% (56·9–65·9%) | |
Trust in state and institutions
| Government trust score (mean [SD]) | Government EVD trust score (mean [SD]) | Health professional EVD trust score (mean [SD]) | ||
|---|---|---|---|---|
| Total | 2·69 (0·67) | 3·20 (0·84) | 3·59 (0·84) | |
| City | ||||
| Beni (n=481) | 2·86 (0·60) | 3·36 (0·72) | 3·73 (0·67) | |
| Butembo (n=480) | 2·64 (0·68) | 3·15 (0·87) | 3·54 (0·88) | |
| Sex | ||||
| Female (n=482) | 2·65 (0·80) | 3·08 (0·97) | 3·49 (0·99) | |
| Male (n=479) | 2·74 (0·50) | 3·33 (0·66) | 3·68 (0·63) | |
| Age group (years) | ||||
| 18–30 (n=450) | 2·67 (0·67) | 3·18 (0·84) | 3·63 (0·85) | |
| 31–45 (n=295) | 2·78 (0·64) | 3·22 (0·80) | 3·51 (0·79) | |
| ≥46 (n=216) | 2·63 (0·70) | 3·23 (0·90) | 3·60 (0·88) | |
| Education | ||||
| None or incomplete primary (n=255) | 2·76 (0·62) | 3·13 (0·79) | 3·54 (0·84) | |
| Primary completed (n=363) | 2·63 (0·69) | 3·15 (0·90) | 3·53 (0·91) | |
| Secondary completed or higher (n=343) | 2·72 (0·67) | 3·31 (0·80) | 3·69 (0·74) | |
| Wealth | ||||
| Poorest quartile (n=270) | 2·71 (0·71) | 3·08 (0·89) | 3·48 (0·96) | |
| Second quartile (n=206) | 2·67 (0·73) | 3·22 (0·86) | 3·60 (0·81) | |
| Third quartile (n=263) | 2·68 (0·60) | 3·26 (0·76) | 3·67 (0·76) | |
| Richest quartile (n=222) | 2·71 (0·63) | 3·27 (0·83) | 3·62 (0·77) | |
| Ethnicity | ||||
| Non-Nande (n=105) | 2·78 (0·75) | 3·15 (0·95) | 3·72 (0·70) | |
| Nande (n=856) | 2·69 (0·66) | 3·21 (0·83) | 3·58 (0·85) | |
Table 2
Trust score by demographic characteristics
EVD=Ebola virus disease.
All respondents had heard about the EVD outbreak, including 932 (97·6%, 95% CI 96·1–98·6) in the past week. Most respondents had received information about how to protect themselves, where to seek care, and symptoms of EVD (table 3). Fewer received information about what to do if a relative was affected, cases in the province, and ongoing efforts to control the outbreak (table 3). This information was used to compute an EVD information score summing the types of information received by respondents. The mean EVD information score was 4·7 (SD 1·48) out of a maximum of 6.
| Unweighted (n) | Weighted (% [95% CI]) | ||
|---|---|---|---|
| Total | 961 | . | |
| Type of information received | |||
| Cases of Ebola in the province | 605 | 63·7% (54·3–72·2) | |
| Intervention to combat Ebola in the province | 641 | 63·7% (54·5–72·1) | |
| Symptoms of Ebola | 831 | 85·0% (81·2–88·2) | |
| How to protect oneself | 896 | 91·6% (89·1–93·5) | |
| Where to seek care | 824 | 80·3% (77·0–83·2) | |
| What to do if a relative has Ebola | 747 | 72·3% (68·8–75·7) | |
| Heard misinformation | |||
| Ebola does not exist | 850 | 86·5% (82·9–89·4) | |
| Ebola is fabricated for financial gains | 826 | 84·7% (80·2–88·3) | |
| Ebola is fabricated to destabilise the region | 837 | 86·1% (81·8–89·4) | |
| Heard any of the three statements | 899 | 92·2% (88·8–89·4) | |
| Heard all three statements | 768 | 78·0% (73·0–82·4) | |
| Belief in misinformation | |||
| Ebola does not exist | 230 | 25·5% (21·7–29·6) | |
| Ebola is fabricated for financial gains | 312 | 32·6% (28·2–37·3) | |
| Ebola is fabricated to destabilise the region | 371 | 36·4% (32·1–41·0) | |
| Believe any of the three statements | 446 | 45·9% (41·7–50·2) | |
| Believe all three statements | 171 | 18·2% (14·3–22·7) | |
Respondents who had received or heard information or believed misinformation about Ebola
Respondents received Ebola information from friends and family (863 [88·8%, 95% CI 86·0–91·1]), community radio stations (803 [82·4%, 77·4–86·5]), national radio stations (657 [67·9%, 59·9–75·0]), religious leaders (691 [73·1%, 66·4–78·9]), and health professionals (562 [52·8%, 46·8–58·8]). Fewer had heard about EVD from local authorities (248 [21·3%, 16·3–27·2]) or national government (305 [28·7%, 23·5–34·5]).
Most respondents had heard statements that the EVD outbreak does not exist, was fabricated by the authorities for financial gains, or was fabricated to destabilise the region (table 3). One in four respondents believed in the statement that Ebola does not exist (table 3). A higher proportion of respondents believed that the Ebola outbreak was fabricated for financial gains, or was fabricated to destabilise the region; 446 (45·9%) respondents believed at least one misinformation statement to be true and 171 (18·2%) believed all were true (table 3). The EVD information and government trust score, government EVD trust score, and health professionals EVD trust scores were significantly lower among those who believed in all or any misinformation statements than among those who did not believe in misinformation statements (appendix).
Among all respondents, 220 (33·9%, 95% CI 25·7–43·3) thought contracting EVD was likely or very likely, 239 (21·9%, 17·4–27·2) were uncertain, and 502 (44·2%, 38·0–50·5) thought that contracting EVD was unlikely or very unlikely.
Respondents said that they engaged in protective behaviours following the outbreak declaration, including avoiding physical contact with people exposed or potentially exposed to EVD (table 4). Few people reported general avoidance of social interaction, such as avoiding visiting family members or neighbours (table 4). More respondents reported avoiding any public space, including church or public transport. Reduction in any physical interaction, including shaking hands or hugging with relatives and others, was commonly reported (table 4).
| Unweighted (n) | Weighted (% [95% CI]) | ||
|---|---|---|---|
| Total | 961 | . | |
| Direct avoidance | |||
| Avoid contact with people suspected to have Ebola | 757 | 75·5% (68·1–81·7) | |
| Avoid contact with body of suspected Ebola death | 801 | 78·8% (72·2–84·2) | |
| Avoid contact with people suspected of recent contact with someone infected by Ebola | 743 | 74·9% (68·9–80·0) | |
| Any direct avoidance | 846 | 82·2% (75·7–87·2) | |
| All direct avoidance | 613 | 67·5% (60·7–73·6) | |
| Social interaction avoidance | |||
| Avoid visiting extended family members | 11 | 1·1% (0·4–3·1) | |
| Avoid visiting neighbours | 10 | 0·9% (0·3–2·8) | |
| Stay home more than usual | 21 | 2·3% (1·4–3·9) | |
| Any social interaction avoidance | 26 | 2·8% (1·7–4·7) | |
| Physical contact avoidance | |||
| Reduce physical interactions with relatives | 369 | 30·6% (26·7–34·7) | |
| Reduce physical interactions with others | 591 | 53·9% (50·1–57·6) | |
| Any physical contact avoidance | 601 | 54·9% (51·0–58·8) | |
| Public space avoidance | |||
| Avoid public spaces like markets or stadiums | 121 | 11·5% (9·1–14·5) | |
| Avoid going to church | 40 | 3·7% (2·3–5·8) | |
| Avoid taking public transport | 83 | 7·9% (5·8–10·6) | |
| Any public space avoidance | 196 | 19·6% (16·3–23·4) | |
| Hygiene (washing hands more frequently) | 885 | 89·9% (87·2–92·0) | |
Respondents who adhered to preventive behaviours
Most respondents (876 [89·7%, 95% CI 86·1–92·4]) reported that they would first seek care from formal rather than informal sources if they believed they had EVD. However, among those who believed in all misinformation statements, 123 (70·8%, 59·6–80·0) of 171 would seek care from formal health service providers, compared with 753 (93·9%, 90·6–96·1, p<0·0001) of 790 among those who did not believe the statements. The government trust score was significantly higher among respondents who would seek care from formal sources than it was among those who would seek care from informal sources (2·7 (SD 0·65) vs 2·4 (0·72) out of 5, p<0·0001). Similarly, the government EVD trust score (3·3 SD 0·78) and health professional EVD trust score (2·5 SD 1·04) were significantly higher among respondents who would seek care from formal sources (p<0·0001) than they were among those seeking care from informal sources (government EVD trust score 3·7 SD 0·73, and health professional EVD trust score 3·0 SD 1·12).
Confidence in vaccines in general was high and most respondents believed that vaccines work (899 [90·7%, 95% CI 87·0–93·4) and are safe (852 [88·5%, 85·4–91·0]). Fewer believed that EVD vaccines work (641 [65·7%, 59·9–71·0]). 589 reported they would accept the EVD vaccine (63·3%, 58·0–68·3). Reasons for not accepting the vaccine included that it was unsafe (225/313 [71·5%, 64·1–77·9]), did not work (75/313 [23·4%, 16·4–32·1]), or was not needed (45/313 [12·0%, 8·2–17·4]). Those who believed that the EVD vaccine is effective were more likely to accept vaccination if offered than were those who did not believe it is effective (crude OR 27·3, 95% CI 16·9–44·1).
Among those who believed all misinformation statements, 31 (24·2%, 95% CI 16·7–33·8) of 171 would accept the vaccine, compared with 558 (72·0%, 67·4–76·2) of 790 among those who did not believe all three statements. We did not find a significant difference in government trust score between those who would accept the vaccine (2·8 SD 0·68) and those who would not (2·6 SD 0·64). However, the government EVD trust score was significantly higher among respondents who would accept the vaccine compared to those who would not (3·4 SD 0·74 vs 2·9 SD 0·90, p<0·0001), as was the health professional EVD trust score (3·8 SD 0·74 vs 3·3 SD 0·91, p<0·0001).
Men had higher odds of avoidance behaviour and vaccine acceptance than did women; however, they had lower odds of seeking care from formal providers if they suspected having EVD (table 5, appendix). Each one-point increase in government trust, government EVD trust, health professionals EVD trust, and EVD information scores increased the odds of adoption of avoidance behaviours, Ebola vaccine acceptance, and seeking care from formal providers when suspecting Ebola (table 5). Belief that Ebola is not real was associated with lower odds on all the outcomes of interest. Each one-point increase in EVD risk perception score was associated with higher odds of avoidance behaviour. Among those who believed that Ebola is real, increased risk perception was associated with lower odds of care seeking and vaccine acceptance (table 5, appendix).
| Any avoidance behaviour (OR [95% CI]) | All avoidance behaviour (OR [95% CI]) | Formal health-care seeking (OR [95% CI]) | Ebola vaccine acceptance (OR [95% CI]) | ||
|---|---|---|---|---|---|
| Sex (men vs women) | 11·04 (10·82–11·27) | 3·04 (3·00–3·07) | 0·62 (0·61–0·63) | 1·10 (1·09–1·11) | |
| Age (1-year increase) | 1·00 (1·00–1·00) | 0·99 (0·99–0·99) | 1·03 (1·03–1·03) | 0·99 (0·99–0·99) | |
| Education (vs incomplete primary or less) | |||||
| Primary completed | 0·87 (0·85–0·88) | 0·86 (0·84–0·87) | 1·99 (1·95–2·03) | 0·98 (0·96–0·99) | |
| Secondary completed or higher | 0·81 (0·80–0·83) | 0·81 (0·79–0·82) | 1·80 (1·76–1·84) | 1·14 (1·13–1·16) | |
| City (Butembo vs Beni) | 0·17 (0·17–0·18) | 0·57 (0·57–0·58) | 0·61 (0·60–0·63) | 2·28 (2·25–2·31) | |
| Nande (Nande vs non-Nande) | 0·85 (0·83–0·88) | 0·93 (0·91–0·95) | 0·62 (0·60–0·64) | 0·72 (0·71–0·74) | |
| Wealth (vs poorest quartile) | |||||
| Second quartile | 0·86 (0·84–0·88) | 0·71 (0·70–0·72) | 0·97 (0·95–0·99) | 0·89 (0·88–0·90) | |
| Third quartile | 0·96 (0·94–0·98) | 0·62 (0·61–0·63) | 0·65 (0·63–0·66) | 0·92 (0·91–0·93) | |
| Richest quartile | 0·68 (0·66–0·69) | 0·64 (0·63–0·65) | 0·60 (0·58–0·61) | 0·80 (0·78–0·81) | |
| Belief Ebola does not exist (yes vs no) | 0·57 (0·55–0·59) | 0·45 (0·44–0·47) | 0·06 (0·05–0·06) | 0·22 (0·21–0·22) | |
| EVD risk perception score* | 1·60 (1·59–1·61) | 1·54 (1·54–1·55) | 0·95 (0·94–0·96) | 0·99 (0·98–0·99) | |
| Belief–risk interaction* | 1·30 (1·28–1·32) | 1·49 (1·48–1·51) | 1·73 (1·70–1·75) | 1·09 (1·08–1·10) | |
| Government trust score* | 1·61 (1·59–1·62) | 1·44 (1·43–1·45) | 1·40 (1·39–1·42) | 1·16 (1·15–1·17) | |
| Government EVD trust score* | 1·65 (1·63–1·66) | 1·32 (1·31–1·33) | 1·38 (1·37–1·40) | 1·47 (1·46–1·48) | |
| Health professionals EVD trust score* | 1·08 (1·07–1·09) | 1·20 (1·19–1·21) | 1·25 (1·24–1·26) | 1·28 (1·27–1·29) | |
| EVD information score* | 1·49 (1·48–1·49) | 1·32 (1·32–1·33) | 1·08 (1·08–1·09) | 1·22 (1·21–1·22) | |
Table 5
Adjusted odds ratios for EVD avoidance, health seeking, and vaccine acceptance behaviours
All p values <0·001. OR=odds ratio. EVD=Ebola virus disease.
*
One-point increase.
Discussion
We collected data during an active EVD outbreak in eastern DR Congo, with the aim to better characterise the role of institutional trust and misinformation on individual behaviours of EVD prevention. The EVD outbreak is occurring in an active conflict zone, where low institutional trust is linked to a long-term decline in security and political confidence.24,25 We identified low levels of trust in government institutions and widespread belief in misinformation about EVD. Exposure to violence reduces political trust in general. Local authorities were more frequently trusted than were provincial and national levels of government, which might reflect enhanced access, visibility, and direct delivery of services. Health professionals were more frequently trusted than were authorities, although we did not distinguish between government, private, and humanitarian health providers.
24.
Van der Bracht, K ∙ Flaam, H ∙ Vlassenroot, K ∙ et al.
Conflict and insecurity: a sociological perspective on perceptions of insecurity in conflict-affected Democratic Republic of Congo
Curr Sociol. 2017; 65:336-355
25.
Lake, M
Building the rule of war: postconflict institutions and the micro-dynamics of conflict in Eastern DR Congo
Int Organ. 2017; 71:281-315
Confidence in vaccines in general was high, reflecting similar findings in the region,26 but reduced when considering Ebola vaccines. This difference might reflect fear of contamination or considerations of cost if the vaccine is not perceived as being offered for free, or also because it is new. Our findings suggest that low institutional trust and belief in misinformation are linked to reduced adherence to EVD preventive behaviours. Nevertheless, reported overall adherence to selected preventive behaviours was high, including avoidance of direct and physical contact and hygiene measures. This finding probably reflects the important local engagement efforts that have taken place since the beginning of the outbreak.
26.
Merten, S ∙ Schaetti, C ∙ Manianga, C ∙ et al.
Local perceptions of cholera and anticipated vaccine acceptance in Katanga province, Democratic Republic of Congo
BMC Public Health. 2013; 13:60
Using crude and adjusted analyses for all composite measures of trust, the adoption of preventive behaviour was positively correlated with higher trust. Increased EVD risk perception was associated with reduced odds of care seeking and vaccine acceptance among those who believed that Ebola is real. Although this finding seems counterintuitive, it is consistent with previous findings.22 Possibly, as perceived risk increases, the likelihood decreases of respondents adopting behaviour that might ultimately expose them to EVD or be perceived to increase the risk of exposure.
22.
Tenkorang, EY
Effect of knowledge and perceptions of risks on Ebola-preventive behaviours in Ghana
Int Health. 2018; 10:202-210
Our findings on trust and EVD outbreak response align with the only previous quantitative study27 on this issue done in Monrovia, Liberia, during the EVD outbreak in 2013–15. That study reported low levels of trust in the national government, although substantially higher than those we found in North Kivu, and the authors identified a correlation between trust in government and compliance with government EVD control measures. A qualitative study reported that a lack of trust, information, and ownership undermined contact tracing during the 2013–15 EVD outbreak in west Africa.28 Other studies in various settings have identified an association between distrust and reduced adherence to recommended public health interventions, suggesting that our findings fit well within the existing literature.13–16,29 The general agreement around this issue across settings and methodologies reinforces our findings. Studies largely reaffirm the general principle of trust as an essential element for effective public health interventions, including outbreak control.30 However, there are many questions that we are not able to answer with these data. For example, how does trust in governments and local and national institutions interact with and relate to trust in local and international non-governmental organisations, especially in conflict settings? What are the most effective tools for building trust? And should limited outbreak response resources be directed to rebuilding institutional trust?
27.
Blair, RA ∙ Morse, BS ∙ Tsai, LL
Public health and public trust: survey evidence from the Ebola virus disease epidemic in Liberia
Soc Sci Med. 2017; 172:89-97
28.
Kutalek, R ∙ Wang, S ∙ Fallah, M ∙ et al.
Ebola interventions: listen to communities
Lancet Glob Health. 2015; 3:e131
13.
O'Malley, AS ∙ Sheppard, VB ∙ Schwartz, M ∙ et al.
The role of trust in use of preventive services among low-income African-American women
Prev Med. 2004; 38:777-785
14.
Mohseni, M ∙ Lindstrom, M
Social capital, trust in the health-care system and self-rated health: the role of access to health care in a population-based study
Soc Sci Med. 2007; 64:1373-1383
15.
Meredith, LS ∙ Eisenman, DP ∙ Rhodes, H ∙ et al.
Trust influences response to public health messages during a bioterrorist event
J Health Commun. 2007; 12:217-232
16.
Goold, SD
Trust, distrust and trustworthiness
J Gen Intern Med. 2002; 17:79-81
29.
Salmon, DA ∙ Dudley, MZ ∙ Glanz, JM ∙ et al.
Vaccine hesitancy: causes, consequences, and a call to action
Vaccine. 2015; 33:D66-D71
30.
Meredith, LS ∙ Eisenman, DP ∙ Rhodes, H ∙ et al.
Trust influences response to public health messages during a bioterrorist event
J Health Commun. 2007; 12:217-232
Although the level of exposure to EVD information and knowledge of symptoms and transmission was generally high, belief in misinformation was widespread. As anticipated, belief in certain misinformation was associated with lower exposure to EVD-related information, which suggests intensive communication initiatives might effectively counter the circulation of harmful misinformation. The belief that EVD does not exist was linked to low adoption of preventive behaviours. Low institutional trust was associated with reduced levels of EVD knowledge (including transmission routes and clinical symptoms) and a greater likelihood of believing certain misinformation. The Liberia study27 did not identify an association between trust and information about EVD and did not explore issues of misinformation. The reason for the different findings about information between the studies is unclear but might relate to study design, specific questions asked, or actual population differences.
27.
Blair, RA ∙ Morse, BS ∙ Tsai, LL
Public health and public trust: survey evidence from the Ebola virus disease epidemic in Liberia
Soc Sci Med. 2017; 172:89-97
There are limitations to this study. The survey was done in urban settings and findings might not reflect other affected areas, particularly rural areas. The survey explored trust in health workers regardless of their affiliation with government, private, or humanitarian and non-governmental service providers. Perceptions might differ on the basis of these affiliations. All data were self-reported, including behaviour changes. Social desirability bias (telling researchers what participants expect they want to hear, rather than what they actually do) can lead to over-reporting of adherence to prevention activities and there might be discrepancies between what people report doing and what they do.
There have been great advances in the response to many outbreak-prone pathogens over the past decade, but our understanding of the social dynamics and community perceptions related to behaviour during outbreaks require more research. The EVD outbreak in this study occurred in a highly insecure, densely populated environment, with attacks against civilians and a tense political situation, including delayed presidential elections. Attacks against health professionals, condemned by the UN, jeopardise the response to the EVD outbreak.31 A lack of institutional trust and widespread misinformation are, our findings suggest, additional factors that undermine control efforts. Engaging locally trusted leaders and service providers could help to build trust with Ebola responders who are not from these communities. If those involved in the EVD response are transparent and consistent in responding to the local needs to stop this outbreak, the trust established during this response could translate into long-term general trust in institutions. Until trust building is effectively translated into response strategies and communication protocols, the basic principles of intensive risk communication by trusted sources in a transparent, sincere, and consistent manner should be the cornerstone of the social mobilisation and community engagement efforts. Mediation (eg, by local and international peacebuilding organisations) and interactive dialogue between communities, community leaders, and local and international Ebola responders might address misinformation about the reality and politicisation of outbreaks, reducing the tensions between EVD responders and the community at risk.
31.
UN Security Council
General Assembly resolution 2439. The rule of law at the national and international levels: S/RES/2439
Contributors
PV and PNP obtained the data with input from KKB, JB, and EJN for the study and questionnaire design. PV, PNP, and EJN analysed the data. PV, PNP, and EJN wrote the first draft. All authors contributed to data interpretation and reviewed and edited the manuscript.
Declaration of interests
We declare no competing interests.
Acknowledgments
This study was funded by the Harvard Humanitarian Initiative Innovation Fund.
Supplementary Material (1)
Supplementary appendix
References
Faye, O ∙ Boëlle, PY ∙ Heleze, E ∙ et al.
Chains of transmission and control of Ebola virus disease in Conakry, Guinea, in 2014: an observational study
Lancet Infect Dis. 2015; 15:320-326
Walldorf, JA ∙ Cloessner, EA ∙ Hyde, TB ∙ et al.
Considerations for use of Ebola vaccine during an emergency response
Vaccine. 2017;
published online Sept 7. DOI:10.1016/j.vaccine.2017.08.058
Congo Research Group
Mass killings in Beni territory: political violence, cover ups, and cooptation
Center on International Cooperation, New York, 2017
Colletta, NJ ∙ Cullen, ML
Violent conflict and the transformation of social capital: lessons from Cambodia
World Bank Publications, Rwanda, Guatemala, and Somalia, 2000
http://documents.worldbank.org/curated/en/799651468760532921/pdf/multi-page.pdf
Date accessed: February 28, 2019
De Juan, A ∙ Pierskalla, JH
Civil war violence and political trust: microlevel evidence from Nepal
Confl Manag Peace Sci. 2016; 33:67-88
Rohner, D ∙ Thoenig, M ∙ Zilibotti, F
Seeds of distrust: conflict in Uganda
J Econ Growth. 2013; 18:217-252
Vinck, P ∙ Pham, PN ∙ Makoond, A ∙ et al.
Voices from Congo Report #13. Harvard Humanitarian Initiative
http://www.peacebuildingdata.org/sites/m/pdf/DRC_Poll13_english_final.pdf
Date: 2018
Date accessed: February 28, 2019
Weber, L
WHO fears Ebola in DRC will spin out of control due to political violence. Huffington Post
https://www.huffingtonpost.com/entry/ebola-drc-congo-perfect-storm-who_us_5baa6616e4b0f143d10ddff9
Date: Sept 25, 2018
Date accessed: February 28, 2019
Murphy, R
UN peacekeeping in the Democratic Republic of the Congo and the protection of civilians
J Conflict Secur Law. 2016; 21:209-246
DR Congo EVD outbreak: Beni attack halts outreach work. BBC News
Gilson, L
Trust and the development of health care as a social institution
Soc Sci Med. 2003; 56:1453-1468
O'Malley, AS ∙ Sheppard, VB ∙ Schwartz, M ∙ et al.
The role of trust in use of preventive services among low-income African-American women
Prev Med. 2004; 38:777-785
Mohseni, M ∙ Lindstrom, M
Social capital, trust in the health-care system and self-rated health: the role of access to health care in a population-based study
Soc Sci Med. 2007; 64:1373-1383
Meredith, LS ∙ Eisenman, DP ∙ Rhodes, H ∙ et al.
Trust influences response to public health messages during a bioterrorist event
J Health Commun. 2007; 12:217-232
Leavitt, JW
Public resistance or cooperation? A tale of smallpox in two cities
Biosecur Bioterror. 2003; 1:185-192
Mechanic, D ∙ Schlesinger, M
The impact of managed care on patients' trust in medical care and their physicians
JAMA. 1996; 275:1693-1697
Yaqub, O ∙ Castle-Clarke, S ∙ Sevdalis, N ∙ et al.
Attitudes to vaccination: a critical review
Soc Sci Med. 2014; 112:1
Nuriddin, A ∙ Jalloh, MF ∙ Meyer, E ∙ et al.
Trust, fear, stigma and disruptions: community perceptions and experiences during periods of low but ongoing transmission of Ebola virus disease in Sierra Leone, 2015
BMJ Glob Health. 2018; 3:e000410
Betancourt, TS ∙ Brennan, RT ∙ Vinck, P ∙ et al.
Associations between mental health and Ebola-related health behaviors: a regionally representative cross-sectional survey in post-conflict Sierra Leone
PLoS Med. 2016; 13:e1002073
Tenkorang, EY
Effect of knowledge and perceptions of risks on Ebola-preventive behaviours in Ghana
Int Health. 2018; 10:202-210
Kaler, A
Health interventions and the persistence of rumour: the circulation of sterility stories in African public health campaigns
Soc Sci Med. 2009; 68:1711-1719
Van der Bracht, K ∙ Flaam, H ∙ Vlassenroot, K ∙ et al.
Conflict and insecurity: a sociological perspective on perceptions of insecurity in conflict-affected Democratic Republic of Congo
Curr Sociol. 2017; 65:336-355
Lake, M
Building the rule of war: postconflict institutions and the micro-dynamics of conflict in Eastern DR Congo
Int Organ. 2017; 71:281-315
Merten, S ∙ Schaetti, C ∙ Manianga, C ∙ et al.
Local perceptions of cholera and anticipated vaccine acceptance in Katanga province, Democratic Republic of Congo
BMC Public Health. 2013; 13:60
Blair, RA ∙ Morse, BS ∙ Tsai, LL
Public health and public trust: survey evidence from the Ebola virus disease epidemic in Liberia
Soc Sci Med. 2017; 172:89-97
Kutalek, R ∙ Wang, S ∙ Fallah, M ∙ et al.
Ebola interventions: listen to communities
Lancet Glob Health. 2015; 3:e131
Salmon, DA ∙ Dudley, MZ ∙ Glanz, JM ∙ et al.
Vaccine hesitancy: causes, consequences, and a call to action
Vaccine. 2015; 33:D66-D71
Meredith, LS ∙ Eisenman, DP ∙ Rhodes, H ∙ et al.
Trust influences response to public health messages during a bioterrorist event
J Health Commun. 2007; 12:217-232
UN Security Council
General Assembly resolution 2439. The rule of law at the national and international levels: S/RES/2439
Related Hub
Ebola Resource Centre
Updated archives from the Lancet family of journals and free access to new content as it is published.
Figures (1)
Article metrics
-
508Citations
-
563Captures
-
145Mentions
Supplementary material (1)
PDF (148.14 KB)
Supplementary appendix
30063-5/asset/22a606a7-3b62-4e3e-b302-f1f7996d7db3/main.assets/gr1_lrg.jpg)
