In the U.S., approximately 9.3 million lead service lines (LSLs) account for most lead contamination of drinking water. As the commitment to replace LSLs with safer materials grows, empirical evidence is needed to understand which households are benefitting most from current replacement practices. This exploratory study analyzes factors predictive of whether an LSL was replaced fully (from water main to premise) or partially (only the portion on public property). Conventional ordinary least squares, negative binomial, and geographically weighted regression models are used to test the hypothesis that full lead service line replacements (LSLRs) were less common in lower-income, higher-minority neighborhoods under a cost-sharing program design in Washington, D.C. between 2009 and 2018. The study finds supportive evidence that household income is a major predictor of full replacement prevalence, with race also showing significance in some analyses. These findings highlight the need for further research into patterns of full versus partial LSLR across the U.S. and may inform future decisions about LSLR policy and program design.

No abstract