High time-resolution simulation of E. coli on hands reveals large variation in microbial exposures amongst Vietnamese farmers using human excreta for agriculture

Infectious disease transmission is frequently mediated by the environment, where people's movements through and interactions with the environment dictate risks of infection and/or illness. Capturing these interactions, and quantifying their importance, offers important insights into effective interventions. In this study, we capture high time-resolution activity data for twenty-five Vietnamese farmers during collection and land application of human excreta for agriculture. Although human excreta use improves productivity, the use increases risks of enteric infections for both farmers and end users. In our study, the activity data are integrated with environmental microbial sampling data into a stochastic-mechanistic simulation of E. coli contamination on hands and E. coli ingested. Results from the study include frequent and variable contact rates for farmers' hands (from 34 to 1344 objects contacted per hour per hand), including highly variable hand-to-mouth contact rates (from 0 to 9 contacts per hour per hand). The frequency of hand-to-mouth contacts was substantially lower than the widely-used frequency previously reported for U.S. Office Workers. Environmental microbial contamination data highlighted ubiquitous E. coli contamination in the environment, including excreta, hands, toilet pit, handheld tools, soils, surfaces, and water. Results from the simulation suggest dynamic changes in E. coli contamination on hands, and wide variation in hand contamination and E. coli ingested amongst the farmers studied. Sensitivity analysis suggests that E. coli contamination on hands and ingested doses are most influenced by contamination of handheld tools, excreta, and the toilet pit as well as by frequency of hand-to-mouth contacts. The study findings are especially relevant given the context: no farmers reported adequate storage time of human excreta, and personal protective mask availability did not prevent hand-to-mouth contacts. Integrating high time-resolution activity data into exposure assessments highlights variation in exposures amongst farmers, and offers greater insight into effective interventions and their potential impacts.