(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Community trust of government and non-governmental organizations during the 2014-16 Ebola epidemic in Liberia

['Ronan F. Arthur', 'School Of Medicine', 'Stanford University', 'Stanford', 'California', 'United States Of America', 'Lily M. Horng', 'Fatorma K. Bolay', 'National Public Health Institute Of Liberia', 'Monrovia']

Date: 2022-02 

Abstract The West African Ebola Virus Disease epidemic of 2014-16 cost more than 11,000 lives. Interventions targeting key behaviors to curb transmission, such as safe funeral practices and reporting and isolating the ill, were initially unsuccessful in a climate of fear, mistrust, and denial. Building trust was eventually recognized as essential to epidemic response and prioritized, and trust was seen to improve toward the end of the epidemic as incidence fell. However, little is understood about how and why trust changed during Ebola, what factors were most influential to community trust, and how different institutions might have been perceived under different levels of exposure to the outbreak. In this large-N household survey conducted in Liberia in 2018, we measured self-reported trust over time retrospectively in three different communities with different exposures to Ebola. We found trust was consistently higher for non-governmental organizations than for the government of Liberia across all time periods. Trust reportedly decreased significantly from the start to the peak of the epidemic in the study site of highest Ebola incidence. This finding, in combination with a negative association found between knowing someone infected and trust of both iNGOs and the government, indicates the experience of Ebola may have itself caused a decline of trust in the community. These results suggest that national governments should aim to establish trust when engaging communities to change behavior during epidemics. Further research on the relationship between trust and epidemics may serve to improve epidemic response efficacy and behavior uptake.

Author summary Behavior change was critical to the West African Ebola epidemic response in Liberia. Previous studies show trust in the Liberian government was associated with behavior compliance, while hardships related to Ebola were associated with lower levels of trust. Trust of international non-governmental organizations (iNGOs) was consistently higher than trust in the Liberian government. However, studies have not compared exposed communities to non-exposed communities over time. This study measured trust in the government and trust in iNGOs in three communities with different exposure to the Ebola epidemic. Results corroborate that trust of iNGOs was higher than trust in the government across all five time periods. Trust decreased significantly during the peak of the Ebola crisis, especially in the community with the highest incidence. Individuals who believed that Ebola was real and had high levels of knowledge about Ebola had higher trust than those who did not. Being the frequent witness of Ebola-related events was associated with higher trust in iNGOs, but not in the government. These findings indicate that efforts to improve Ebola awareness and knowledge may positively influence trust, while exposure to the epidemic may reinforce prior mistrust of government institutions.

Citation: Arthur RF, Horng LM, Bolay FK, Tandanpolie A, Gilstad JR, Tantum LK, et al. (2022) Community trust of government and non-governmental organizations during the 2014-16 Ebola epidemic in Liberia. PLoS Negl Trop Dis 16(1): e0010083. https://doi.org/10.1371/journal.pntd.0010083 Editor: Marilia Sá Carvalho, Fundacao Oswaldo Cruz, BRAZIL Received: February 8, 2021; Accepted: December 11, 2021; Published: January 27, 2022 This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication. Data Availability: All data are made available in a public repository at Dryad doi:10.5061/dryad.h44j0zpmv. Funding: RA received support from a NSF Graduate Research Fellowship (2015); LH, SL received support from a DOD Global Emerging Infections Surveillance grant (#P0244_15_N3; url: https://www.health.mil/Military-Health-Topics/Combat-Support/Armed-Forces-Health-Surveillance-Branch/Global-Emerging-Infections-Surveillance-and-Response). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. Author Fatorma K. Bolay was unable to confirm their authorship contributions. On their behalf, the corresponding author has reported their contributions to the best of their knowledge.

Introduction The 2014–16 Ebola Virus Disease (Ebola) epidemic in West Africa, principally in Guinea, Liberia and Sierra Leone, was the deadliest Ebola epidemic in history, with more than 28,000 reported cases and 11,000 associated deaths [1]. Originating in Guinea in late 2013, the Ebola outbreak spread to Liberia’s northern Lofa County by March of 2014 [2] and from there to the urban Montserrado County by June 2014. In August 2014, locals in the Monrovia slum of West Point looted a health center and forcefully withdrew Ebola patients there. The government responded with an enforced quarantine that turned violent when government forces fired shots into the crowd during a protest, resulting in the death of a teenage boy. The event incited fear, mistrust, and public memories of recent civil war [3]. International response to the epidemic, other than through the few institutions already on the ground in Liberia, was slow to mobilize [4, 5]. The majority of Ebola treatment units (ETUs), for example, were constructed after incidence in Liberia had already begun to decline in September 2014 [1, 6]. Numerous health promotion, risk communication, and psycho-social support interventions by international organizations were likewise launched in August–November 2014 [7]. Thus, actions taken by international organizations late to arrive were likely not the sole driver of Ebola incidence decline. Indeed, modeling indicates a combination of institutional intervention and individual behavior change helped to finally contain and end the epidemic [7, 8]. In the absence of vaccines and approved therapies, non-pharmaceutical interventions were recommended, including handwashing, notifying Ebola response teams of infected persons, and safe burial practices [9]. Some of these behaviors and practices were widely adopted [10, 11], but a few behaviors recognized for their contribution to ongoing transmission, most notably healthcare provided at home by family members and traditional burial practices, proved more difficult to modify [12, 13]. Early government interventions to interrupt transmission often targeted these behaviors without meaningful dialogue with communities to problem-solve and build trust [14]. Many community members rejected quarantine and safe burial interventions and shunned treatment centers out of fear for the safety of the sick [12, 15]. Medical anthropologists found that healthcare avoidance and traditional burial practices were difficult behaviors to alter in a climate of fear, mistrust, and denial [12, 16]. Some individuals would hide themselves or their loved ones if infected, and, in some communities, there were incidents of violent resistance to community health worker groups [14]. Over time, community trust of government institutions and response efforts were eventually recognized as critical to controlling the epidemic as recommended practices would not be taken up in their absence [17]. Qualitative work has contextualized our understanding of individual trust and fear during the 2014–16 Ebola epidemic, including recent history, tradeoffs, and rational perspectives [12, 18], while quantitative studies have measured associations between trust, behavior, and knowledge. Cross-sectional surveys found higher trust in health institutions was associated with higher compliance with Ebola control measures [10]. Another study found differences in Ebola knowledge, attitudes, and practices between high-incidence communities and low-incidence communities: low-incidence communities, which had less exposure to outbreak response interventions, expressed less knowledge of Ebola and more belief in rumors about Ebola transmission [11]. A household survey conducted in Monrovia, Liberia during the Ebola epidemic found higher levels of trust in international non-governmental organizations (iNGOs), such as the Red Cross, Médecins Sans Frontières (MSF), Partners in Health, and Last Mile Health, than in the Liberian government [10]. Local trust of health authorities may thus be differentiated based on the object of trust. In Liberia, in a post-war, post-colonial context with ongoing national government corruption allegations, health-related trust is likely to be influenced by the historical and modern associations with these actors [3, 19]. Trust may also be associated with social capital, the set of resources available to an individual embedded in their social relations [20]—in other contexts, having good neighborly relations and access to resources through social connections has been associated with greater trust in one’s community [21, 22]. Despite the recognized importance of local trust and behavior change during epidemics, social factors that contribute to disease transmission are still comparatively neglected in the infectious disease dynamics literature [23–26], while the quantitative link between trust and public health outcomes is even less explored [10]. The majority of the quantitative literature on trust and behavior relates to US and UK domestic public health issues, such as for vaccination [27–29] or HIV/AIDS [30]. Few studies have examined trust in low-income country infectious disease contexts [31–33]. In the case of Ebola, mistrust in government was a major obstacle to response [12] as early government actions did not improve trust [14], and, in Sierra Leone, trust in the healthcare system was qualitatively found to have improved after the peak of the epidemic due to the perception of improved health management [18]. Trust in government and iNGOs has been quantitatively associated with behavior change in the 2014–16 West African Ebola epidemic [10] and in the 2018–19 DR Congo Ebola epidemic [33]. However, we lack quantitative studies of how trust changed over time during the Ebola epidemic, or indeed in any epidemic context, and possible drivers of such changes. This understanding would improve our knowledge of how trust is affected by epidemics and by health authorities, which could lead to improved epidemic intervention strategies. This study investigates trust in national government and trust in iNGOs over time in locations with differing levels of Ebola incidence and exposure to epidemic-related activity in Liberia. We conducted household surveys in three Liberian communities that had different experiences during the 2014–16 Ebola epidemic. We retrospectively measured self-reported levels of trust in the government and iNGOs during five time periods that cover the core time frame of the Ebola epidemic. Here we report differentiated trust over the course of the epidemic, between communities, and between the object of trust, and we begin a hypothesis-generating analysis of potential explanatory variables.

Results A total of 1,433 participants were surveyed in Duazon (n = 457), the peri-urban site with high exposure to Ebola response, Careysburg (n = 476), the rural low exposure site, and Tubmanburg (n = 500), the urban high incidence site (Table 1). The mean age of the sample population was 34 years (SD = 12, range = 18–92); 52% of all participants were female. Twenty-five percent of Tubmanburg respondents were Muslim, roughly double the national average of 12.2% [34] and significantly higher than the other two sites (p < 0.01). Tubmanburg residents were also more likely to support the Unity Party (the party of former president Ellen Johnson Sirleaf) in recent national elections (25%) as compared to respondents in Careysburg and Duazon (p < 0.01). Performance on the Ebola knowledge and beliefs quiz (Table 2) was high across the sample population (mean = 4.48 correct answers out of a possible 6, range = 0–6, SD = 1.22). It was well known that Ebola could be found in body fluids (89% agreed) and that a dead body could still be infectious (87% agreed). Common misconceptions included that Ebola can be carried by mosquitoes (42% agreed), that Ebola can be found in drinking water (33% agreed), and that Ebola is caused by witchcraft (22% agreed). Residents of Tubmanburg, the urban high-incidence study site, had a higher aggregate score on these questions than residents of Duazon, the peri-urban site with high exposure to Ebola response interventions (Tubmanburg mean of 4.63 correct answers vs. Duazon mean of 4.32, p < 0.001). Likert-scale trust responses of government and of iNGOs demonstrate differentiation by region and by object of trust over the five time periods (Fig 2). Respondents consistently reported trusting iNGOs to protect their health more than they trusted the Liberian government to protect their health in all five time periods (Table 3). The odds of reporting trust of iNGOs as higher than trust of the government ranged between 2.35 in Time Period 1 to 3.26 in Time Period 5. Overall, trust of government and of iNGOs was higher in Tubmanburg (urban, high incidence) than in the other two study sites, and this differentiation was greatest in Time Period 1. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Change over time in trust in government and iNGOs in each of three study locations: Duazon (peri-urban, high Ebola exposure), Careysburg (rural, low exposure), and Tubmanburg (urban, high incidence). Time period 1 = before Ebola came to Liberia, Jan 2014; Time period 2 = between 1st case in Lofa and 1st case in Monrovia; Time period 3 = between 1st case in Monrovia and quarantine in West Point; Time period 4 = between quarantine in West Point and the end of 2014; Time period 5 = beginning of 2015 as last cases of Ebola occurred and schools opened. https://doi.org/10.1371/journal.pntd.0010083.g002 Trust of government in Tubmanburg, the urban high incidence site, was significantly greater in Time Period 1 compared to Time Periods 2–5, with the greatest difference between Time Periods 1 and 3, for which participants were 40% less likely to rate their trust in government as higher in Time Period 3 than in Time Period 1 (Table 4). Trust of iNGOs was not found to change compared to Time Period 1 for all three regions, except for a small increase in Time Period 5 in Duazon, the peri-urban high exposure site (Table 5). The belief that Ebola was real, being a resident of the urban high-incidence community of Tubmanburg, being a resident of the rural low-exposure community of Careysburg, and higher knowledge of Ebola were all associated with trust in the Liberian government and trust in iNGOs (Fig 3), according to ordinal logistic regression analysis of trust during Time Period 3, the peak of the epidemic. Frequently witnessing Ebola-related events (at least once a day) was positively associated with trust in iNGOs. Having a relationship with someone infected with Ebola and social capital were negatively associated with trust in the government and in iNGOs, while being highly mobile (leaving the community more than once a week) was negatively associated with trust in the government. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Generalized linear model regression coefficient estimates for trust in government and trust in iNGOs as dependent variables in time period 3 using an ordinal regression model. Some independent variables were dropped via AIC stepwise reduction from one or both models. A negative coefficient indicates a negative association with trust, while a positive coefficient indicates a positive association with trust. Community descriptions: Careysburg = rural, low exposure; Tubmanburg = urban, high incidence. https://doi.org/10.1371/journal.pntd.0010083.g003

Discussion Data on community trust from this large household survey study carried out in Liberia in the aftermath of the 2014–16 Ebola epidemic indicate trust levels, especially of the Liberian government, were distinct in areas with different experiences during the crisis. Respondents were more likely to recall their trust of iNGOs as higher than their trust of the government across all five time periods in agreement with a cross-sectional household survey conducted in Monrovia, Liberia [10]. This result was qualitatively corroborated in focus groups conducted in these communities in a parallel study on post-Ebola era trust [39]. Many respondents in these open discussions drew a distinction between iNGOs and government, perceiving all foreign organizations, including foreign government organizations and iNGO-established Ebola treatment units, to be iNGOs. They perceived government as encompassing any institutions, activities, or representatives sponsored by the Liberian government, including government-run hospitals. Some rationalized the iNGO-government trust discrepancy by explaining that iNGOs had no ulterior motives than simply to help during the crisis, while the government was financially benefiting from the epidemic and was therefore incentivized to prolong the disaster and even to have deliberately caused it in the first place. This may also explain the association between trust in iNGOs and witnessing Ebola-related events frequently, an association which was not found for trust in the government. The a priori trust of each institution and perceived incentives for their epidemic activity may have led to the reinforcement of those perspectives when witnessing Ebola-related activity. In Tubmanburg, the area of study with the highest Ebola incidence, there was the highest degree of difference between trust in the government between Time Periods 1 and 3, from the beginning stage to the middle stage of the epidemic. This decrease was not replicated in Duazon, the peri-urban high exposure site, but it was replicated to a lesser extent in Careysburg, the rural, low exposure site. Other studies have found that the initial top-down Ebola interventions, such as enforced cremation, quarantines, and bushmeat prohibitions, decreased public trust [14]; hardships experienced in connection with the epidemic decreased trust [10]; and the public feared Ebola treatment units would lead to the death of sick loved ones [12, 39]. During the Ebola epidemic in Guinea and the subsequent epidemic in the Democratic Republic of the Congo, there were accounts of violent actions taken against healthcare teams, and trust overall was reported to be low [40, 41]. Thus, the incidence of Ebola in Tubmanburg may have led to negative experiences with Ebola response and caused the sharp decline in trust found in this study. In Careysburg, there were heavy quarantine restrictions forbidding the entry of foreign or returning persons to the community. Trust in the government to protect community health cannot be divorced from other trust-related perceptions of the government. In the case of Tubmanburg, public trust may be influenced by recent history of brutal civil war, as the city was occupied by a rebel group and experienced a major battle during the war [42]. More broadly, Liberia has experienced ongoing corruption allegations and a stalled economy with few employment opportunities, and is ranked among the poorest economies in the world by GDP per capita [43]. Given the importance of trust to epidemic intervention and the higher trust afforded to iNGOs than to the government, we may conclude that trust of iNGOs may be leveraged to support public health communication and behavior campaigns. However, this conclusion risks exacerbating the alienation of the government from the populace if iNGOs are perceived as more benevolent entities. The relationship between iNGOs and government in weak states, states with fragile institutions and stagnant economies, is complex and can take many forms, at times adversarial or competitive when the viewpoints of the two parties diverge [44]. In the case of epidemics, there is an opportunity for more cooperative relations because the goals of the national government and the iNGOs are largely aligned. Government and iNGOs may achieve goals through different means, undertaking complementary approaches toward a single objective. Closer collaboration may benefit perception of the government, thereby improving behavioral compliance, provided authorities take care to respectfully acknowledge and account for local beliefs, customs, and leadership. At the same time, however, government collaboration may pose a reputational risk for iNGOs, particularly in communities where trust in government is low. The decline of trust in government public health interventions during the peak of the Ebola epidemic in the community with highest Ebola incidence is concerning because it may indicate that in the absence of effective intervention, a high number of cases may beget more cases not simply due to the nonlinear growth of contagion, but also because of decreased trust. Health economic models which incorporate adaptive behavior change during epidemics typically assume a negative feedback between disease incidence and behavior—as incidence increases, behavioral compliance increases as well, reducing the reproduction number of the epidemic [45–47]. Empirical evidence of this negative feedback relationship has been found in high-income country contexts [48]. However, few studies of this sort have been conducted in low-income country or weak state contexts where low trust may compromise buy-in to behavior change interventions. Fear can drive behavior change, but without perceived response efficacy and confidence of the prescribed behaviors—which in turn requires trust—the set of adopted behaviors may not conform with those that best evidence suggests would actually reduce risk [49, 50]. If increased incidence degrades trust in a weak-state government, as was the case during the Ebola epidemic in Liberia [10], then the relationship between trust and incidence may actually produce a positive feedback of increased transmission and decreased trust, undermining the expected dampening effects of behavior change. These study results should be interpreted in light of the limits of our methods. We measured self-reported perceptions of trust during different parts of the 2014–15 Ebola epidemic in 2018, subjecting the data to recall bias and bias introduced by socially-reinforced post hoc narratives about the community’s shared experience. While blame-driven narratives may have led respondents to recall a heightened degree of mistrust in government, we consider it more likely that the emotional impact of the epidemic has diminished over time, leading to an underestimation of the effect of epidemic events on trust. Given the time elapsed and the diminishing accuracy of memory, we make no claim that we have accurately measured actual levels of trust over time during the Ebola epidemic, a metric we would have needed to acquire during the epidemic itself. However, these data are still useful as they describe and pertain to the community’s narrative of what happened during the epidemic and still allow for quantitative study of this narrative—differences between groups, time periods, and associations with other variables. Today’s perceptions of trust may inform community actions during the COVID-19 pandemic and future public health emergencies. Furthermore, the consistency of our findings with other studies of trust [10, 12, 18] that deployed somewhat different methods during the Ebola epidemic supports the validity of our primary conclusions. Ultimately, the Ebola epidemic in West Africa was largely under control by early 2015, but not before claiming the lives of over 11,000 people. The contributing problems associated with mistrust, misinformation, and behavioral non-compliance experienced in Liberia were repeated in the Ebola epidemic in the Democratic Republic of the Congo, beginning in 2018 [33], and in the ongoing coronavirus (COVID-19) pandemic [51]. We may expect such social reactions in future infectious disease outbreaks to recur, and therefore, the international community and national governments should plan accordingly. Specific attention should be paid to weak states where weak health institutions and their associated low levels of trust are often combined with locations of the greatest geographic risk of zoonotic spill over. We recommend further research on trust, including concurrent study of trust during future public health crises [52]. This and other exploratory studies could inform development of community-based interventions to build trust and improve buy-in during response efforts, similar to the interventions put into place during later stages of the Ebola epidemic in Liberia [53]. Trust-building interventions could serve not only to support responses to ongoing disease outbreaks, but also to prevent future epidemics by strengthening community involvement in public health.

Acknowledgments We gratefully acknowledge scientific collaboration with the US Naval Medical Research Unit Three in Cairo, Egypt. Logistical support was provided by the Liberian Institute of Biomedical Research, the Armed Forces Liberia, and Q&A Services, Inc. We thank the study participants in each of our three study sites for their time and attention. The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the views of the Department of the Navy, the Defense Health Agency, the Department of Defense, or the U.S. Government. This work was prepared as part of their official duties; and, as such, there is no copyright to be transferred.

[END]

[1] Url: https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0010083

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/