A data-driven approach to optimize a sanitary survey for communities in Maputo, Mozambique

ABSTRACT Current approaches to align sanitary surveys to environmental fecal waste indicators have demonstrated limited success, leaving stakeholders focused on fecal waste management with limited and costly tools. We analyzed data from the Maputo Sanitation Project using exploratory factor analysis (EFA) and structural equation modeling (SEM) to enhance the survey's correlation with Escherichia coli concentration data. EFA grouped related survey questions and sampling locations into distinct latent factors. SEM was then used to assess the relationship between the survey latent factors and the mean E. coli concentrations from grouped locations. The results suggested four survey question subgroups: latrine structure, latrine cleanliness, compound/household waste management, and community waste management. In addition, three sampling location subgroups were identified: high-traffic areas (HTAs), food activity locations, and bodily cleaning areas. The largest significant effect size identified suggested that for every 1-unit increase in community waste management, there was a decrease of 1.94 in log10 E. coli per gram of soil in HTAs (p = 0.03), a substantial improvement from the initial 0.50 decrease reported for an expert-weighted metric of all survey questions. These results underscore the importance of community-level waste management and demonstrate the use of a data-driven approach in enhancing environmental health assessments and planning interventions.