Sociodemographic Disparities in Exposures to Inorganic Contaminants in United States Public Water Systems

Maximum contaminant level violations occur more frequently among some public water systems (PWS) serving marginalized communities across the United States. Spatially resolved PWS service area data are now available, allowing for better understanding of relationships between sociodemographic factors and water quality.

We examined associations between detection and concentrations of arsenic (As), chromium (Cr), manganese (Mn), and selenium (Se) and the sociodemographic characteristics of residents served by PWS at two spatial scales: PWS service areas and counties.

We leveraged data on inorganic contaminants in $>\mathrm{273,000}$ drinking water samples from $>\mathrm{21,000}$ PWS collected by state agencies and the US Environmental Protection Agency over the decade between 2005-2015 and curated in a national database developed by the Environmental Working Group. Sociodemographic data were from the 2006-2015 American Community Survey. We developed two-step hurdle models that included: a) a categorical model predicting contaminant detection and b) a continuous model predicting contaminant concentration. We developed these hurdle models using PWS service areas or counties as spatial units of analysis and compared results.

PWS serving 5% higher proportions of limited English-speaking households had significantly greater detections of As, Cr, and Se, ranging from 1.25 [95% confidence interval (CI): 1.18, 1.32] times higher probability of Cr detection to 1.69 (95% CI: 1.58, 1.80) times higher probability of Se detection. Small PWS ($\le \mathrm{10,000}$ customers) had higher concentrations of these same three contaminants. PWS serving greater proportions of Black residents had a lower probability of detecting As, Cr, and Se, but significantly higher concentrations of Cr, Mn, and Se, particularly outside large urban areas. The direction of significant associations was consistent between spatial units of analysis, but county results tended to exhibit greater effect sizes and fewer significant associations.

PWS service area data help to identify drinking water concerns specific to served communities and reduce issues related to statistical bias at larger spatial scales. Targeted subsidies for water infrastructure in communities served by small PWS would help reduce disparities in exposures to drinking water contaminants. https://doi.org/10.1289/EHP14793.