The Science of Human Connection: A Study of the Effect of Social Networks on Acute Gastrointestinal Illness in Rural Ecuadorian Communities

Abstract

Background Diarrheal disease is an important cause of childhood mortality and is spread by two main mechanisms: human contact and contamination of the environment. Though individual- and household-level Water, Sanitation, and Hygiene (WASH) interventions are primarily used to intersect these transmission pathways, seldom are community-level factors considered to ensure both intervention adoption and sustainability. Social constructs like social cohesion are believed to influence the quality and effectiveness of interventions, especially those based on action at the community-level. Few studies, however, identify a causal framework for how social constructs impact WASH interventions and diarrheal disease occurrence, and fewer use social network data. Previous studies in coastal Ecuador showed diarrheal disease spreads more slowly to and in rural villages that have a greater density of social ties, suggesting a greater spread of individual and collective water practices that help reduce transmission of diarrheal disease. Objective This dissertation research aims to extend previous work by methodically defining social cohesion as an important social construct using different types of social network data, examining temporal variability of the effect of social cohesion on diarrheal disease, whether this relationship is mediated by WASH, and the role that gender plays in social cohesion and WASH in rural, coastal Ecuador. Methods Using longitudinal sociometric data from villages in rural, coastal Ecuador, we identify important network determinants of social cohesion and in turn the temporal effect of social cohesion on WASH interventions and diarrheal disease incidence. We use statistics for the analysis of network graph data and a novel two-stage Bayesian hierarchical model. We importantly theorize a causal framework for the observed phenomena through use of qualitative methods. Results Different types of social networks illustrate the multidimensionality of social processes at the household- and community-levels that influence diarrheal disease incidence. While a network comprised of individuals who pass time together becomes a stronger measure of risk over time, due to density of people and increased travel, having a network of core discussants with whom to discuss important matters is a consistent measure of protection. Having a strong community network of core discussants results in 0.87 (0.71, 1.06) fewer odds of diarrheal disease in 2007 and 0.34 (0.26, 0.45) fewer odds of diarrheal disease by 2013. This protective effect is partially mediated by WASH related factors like community sanitation and improved water use over time, suggesting the importance of social constructs at the community-level for intervention implementation and in turn the reduction of diarrheal disease. Qualitative data collected in the same communities, however, revealed the contributions of infrastructural development and an increasing wage economy to the increasing importance of community. Qualitative data also revealed the importance of gender equity for both community social cohesion and adoption of WASH practices. Analysis of social network data shows communities that are more assortative by gender (i.e. that have less gender equity) are less likely to engage in WASH practices at the household-level over time. Significance By understanding how community correlates of social networks affect intervention practices and diarrheal disease transmission, we can leverage social networks to influence positive behavior change and WASH infrastructure. This research objective is in line with target 5 and 6b of the United Nations Sustainable Development Goals, which aim to achieve gender equality and support and strengthen participation of local communities in improving WASH.