SOCIAL SCIENCE, COMMUNITY ENGAGEMENT, AND THE EVD RESPONSE

A socio-behavioral perspective is fundamental to understanding the historical and ongoing factors that have shaped EVD outbreaks, including historical trauma from Western colonization and cultural marginalization of communities within outbreak areas.^{Reference Mohatt, Thompson, Thai and Tebes8–Reference Bedson, Jalloh and Pedi11} Increased fear and lack of control coupled with activities of foreign organizations can lead to actions perceived as infringements on human rights by evoking past colonial activities that quickly degrade trust and respect.^{Reference Durojaye and Mirugi-Mukundi12} This is evident from the community perspective and mistrust toward the implemented policies around contact measures and burial practices, particularly in Guinea during the West Africa EVD outbreak.^{Reference Chan13} This understanding then provides context to the documented high levels of (1) mistrust toward responders, typically epidemiologists and clinicians; (2) attacks targeting health care workers and facilities; and (3) decreased health-seeking behavior amongst potential cases in the EVD outbreaks in Western and Central Africa.^{Reference Roca, Afolabi, Saidu and Kampmann4,Reference Vinck, Pham and Bindu14}

In the West Africa EVD outbreak, social science research and community engagement were eventually used to address these challenges and improve response efforts by drawing on previous ethnographic studies, as well as field research as part of the response.^{Reference Abramowitz, Hipgrave, Witchard and Heymann5,Reference Wilkinson, Parker, Martineau and Leach6,Reference Laverack and Manoncourt9,Reference Jacobsen, Aguirre and Bailey15,Reference Dada, McKay, Mateus and Lees16} These community engagement activities include recruiting local staff for social mobilization activities, working with religious leaders, engaging local journalists and radio, and hosting local community meetings.^{Reference Wilkinson, Parker, Martineau and Leach6} Effective community engagement, informed by social science research, can ensure that interventions are adapted to and by local communities. This leads to improving outbreak response by decreasing levels of anxiety and fear, improving dialogue between responders and communities to ensure effective communication around key public health measures, increasing support and reporting by the community, and encouraging acceptance and implementation of community interventions, which, ultimately, lower transmission rates.^{Reference Moon, Sridhar and Pate7,17,18} Documented successes of effective community engagement integrate elements of bottom-up approaches that empower local communities^{Reference Thiam, Delamou and Camara19} and respect both the local power dynamics^{Reference Laverack and Manoncourt9,Reference Enria, Lees and Smout10} and cultural norms.^{Reference Wilkinson, Parker, Martineau and Leach6} The involvement of social science research is also important in vaccine and treatment trials for emergent diseases. Such involvement can help address community and participant anxieties as well as hopes for the candidate vaccines or treatments, motivations for participation, and the social, political, economic, and historical context of trials.^{Reference Enria, Lees and Smout10,Reference Tengbeh, Enria and Smout20} Further to this, research from the Sierra Leone EVD vaccine trials described how, by interacting with various community leaders, researchers understood the fragmented nature of community hierarchies and were able to diversify engagement strategies to reach different populations within the community.^{Reference Enria, Lees and Smout10} Therefore, consideration of power dynamics, fairness, and trust in implementing an outbreak response program are critical.^{Reference Enria, Lees and Smout10} In order to address these elements, there has been the design and deployment of “rapid social science” research/tools in the 2018–2020 DRC outbreak. However, there is still a need for a validated process integrating social science and epidemiological data to support community engagement during outbreaks.^{Reference Bedson, Jalloh and Pedi11}

INCLUSION OF SOCIO-BEHAVIORAL VARIABLES INTO TRANSMISSION MODELS

Infectious disease transmission models can estimate a range of outbreak parameters, including number of cases, duration, dynamics, and/or impact of control interventions. Model outputs are frequently used to guide response strategies, policies, and resource allocation, and can be powerful tools given appropriate assumptions and variables. Human infectious disease transmission is invariably shaped by individual and population behaviors, yet these variables rarely, if ever, are included in mathematical transmission models.^{Reference Verelst, Willem and Beutels21} For example, World Health Organization (WHO) data suggest that burial practices were linked to 60%–80% of cases in West Africa, resulting in EVDs classification as a behavioral disease.^{Reference Sharareh, Sabounchi, Sayama and MacDonald22} Despite the importance of understanding behavior in transmission dynamics, a systematic review found that most models of behavioral change are purely theoretical and lack scientific validation.^{Reference Verelst, Willem and Beutels21}

Some studies have attempted to find variables for socio-behavioral modeling by using cross-sectional knowledge-attitude-practice (KAP) surveys to quantify socio-behavioral factors.^{Reference Wilken, Pordell and Goode23,Reference Tenkorang24} KAP studies are easily deployable with quick analysis, making them highly desirable for acute outbreaks. While these studies are useful for some basic modeling, they only indicate general behavioral factors and are unable to explain underlying social factors that shape disease transmission. For example, a KAP study conducted in Liberia found that EVD transmission awareness was poor and education was needed to improve outcomes, but was unable to describe how and to what degree education could modify behavior.^{Reference Wilken, Pordell and Goode23} Similarly, a cross-sectional KAP study conducted during the West Africa EVD outbreak found that myths and stigma affected risk perception and, critically, behavior.^{Reference Tenkorang24} While data from these studies could be used to improve prediction models of vulnerable populations, they do not elucidate the social forces that need to be understood to affect behavioral change.

Building behavioral change models requires an understanding of underlying sociocultural structures. For example, while a KAP study may be able to identify burial practices as a risky behavior, it will not explain how to mitigate or shape this behavior positively. Understanding that emotions are a principal driver behind the act of participating in funerary rituals reveals that behavior change will be successful only when risky practices are replaced by ritually and emotionally meaningful equivalent safer practices.^{Reference Keeling25} The importance of behavioral modeling and adaptation demonstrated by recent EVD outbreaks has created a call for systematic integration of these elements into an outbreak response.^{Reference Wilkinson, Parker, Martineau and Leach6,Reference Bedson, Jalloh and Pedi11,Reference Blanchet, Ramesh and Frison26}

CONSIDERATIONS FOR INTEGRATING SOCIAL SCIENCE AND EPIDEMIOLOGICAL APPROACHES

Challenges stem from core epistemological differences between traditional outbreak responders, typically epidemiologists and clinicians approaching problems from a scientific paradigm, and social scientists who approach similar problems from a humanistic paradigm; thus, effective collaboration is limited. Developments that integrate appropriate social science and epidemiological methodologies have not been fully tested and validated through scientific experimentation.^{Reference Spengler, Ervin and Towner27} Therefore, a more structured and rigorous approach for community engagement and socio-behavioral models is crucial for a successful outbreak response.^{Reference Abramowitz, Hipgrave, Witchard and Heymann5,Reference Blanchet, Ramesh and Frison26}

There are a number of theoretical frameworks for context-specific research that incorporates socio-behavioral elements, including alternative randomized controlled trials, process evaluation, multiple case study design, and realist evaluation (context-mechanism-outcome [CMO] studies).^{Reference Svoronos and Mate28} Of these developing methodologies, realist valuations, or CMO studies, present the greatest utility, adaptability, and applied outcomes. CMOs create (1) a generalized theory of how a system works and (2) specific actions/mechanisms that can be taken in described contexts to lead to desired outcomes.^{Reference Linsley, Howard and Owen29} An advantage of this study design is that theory and specific variables are tested and compared in multiple contexts. Theory and variables from CMOs can be tested by future CMO studies to continue evaluation of underlying theories, interventions, and outcomes. By design, this creates a systematic evaluation enabling necessary validation of research methods through an iterative process. Previously, CMO studies were used to better understand and modify complex socio-behavioral interactions, such as improving performance of community health workers in humanitarian crises^{Reference Gilmore, Adams and Bartoloni30} or studying influences of social networks on behavioral change in adolescent sexual and reproductive health in West Africa.^{Reference Van Belle, Marchal, Dubourg and Kegels31} Table 1 displays different CMO studies, the contexts that they have been used in, and example variables.

TABLE 1 Different Context-Mechanism-Outcome Studies

Another approach to standardizing variables and outcomes is a system dynamics approach, which uses multiple social and epidemiological inputs to demonstrate the impact that behavioral feedback has on outbreak outcomes. A post hoc system dynamics model using an integrated set of causal relationships from social science research on EVD was able to recreate epidemiological outcomes reported by the WHO^{Reference Sharareh, Sabounchi, Sayama and MacDonald22} and found that behavioral and sociocultural factors are needed to fully interpret transmission dynamics observed in the West African EVD outbreak.

Once standardized, these methods and models could be used for different aspects of response, including program design, formative research, community engagement programs, case investigation, socio-behavioral models, and post hoc assessment.^{Reference Stellmach, Beshar and Bedford32} All such methodologies should be scientifically studied for effectiveness, which presents a challenge for acute outbreak responses since there is no ability to test methods before implementation. One approach to mitigate this would be to test these methods in non-acute settings, where infectious diseases are endemic. This could also address ethical challenges implementing new methods in fragile outbreak settings with already vulnerable populations. While validation of methods/models may involve culturally specific elements, adaptable frameworks, like CMO studies, should be used to attempt to standardize across environments.