Objectives: Consuming contaminated food is a well-documented individual-level risk factor for diarrheal disease. The sharing of food also influences the distribution of diarrheal disease risk through a community and region. Understanding this social process at a population level is therefore an important dimension of risk not captured by standard individual-level analyses. We examined social networks related to food-sharing in rural villages at 2 scales: within a village, examining whether connections within these networks clustered or were uniformly spread; and among villages, looking at whether food-sharing networks differed according to the village's remoteness from a population center.
Methods: We surveyed 2129 individuals aged 13 years and older in 2003-2004, within a representative (block-randomized) sample of 21 rural villages in Esmeraldas province, northern coastal Ecuador. We calculated degree (number of social contacts) for a social network defined by sharing food.
Results: Networks of households sharing food differ according to remoteness from a metropolitan center. On average, residents living in "far villages" had 2 more social contacts than those in "close villages," and 12 more years of residence in their village. Estimates of transmissibility (a measure of outbreak potential) based on network structure varied as much as 2-fold across these villages.
Conclusions: Food-sharing practices link particular households in rural villages and have implications for the spread of food-borne pathogens. The food-sharing networks in remote rural villages are heterogeneous and clustered, consistent with contemporary theories about disease transmitters. Network-based measures may offer tools for predicting patterns of disease outbreaks, as well as guidance for interventions.