Buy this Article for $10.95

Have a coupon or promotional code? Enter it here:

When you buy this you'll get access to the ePub version, a downloadable PDF, and the ability to print the full article.


  1. Jones, Rachael M. PhD, MPH
  2. Graber, Judith M. PhD, MPH
  3. Anderson, Robert PhD
  4. Rockne, Karl PhD
  5. Turyk, Mary PhD
  6. Stayner, Leslie T. PhD


Context: Environmental Public Health Tracking (EPHT) tracks the occurrence and magnitude of environmental hazards and associated adverse health effects over time. The EPHT program has formally expanded its scope to include finished drinking water quality.


Objectives: Our objective was to describe the features, strengths, and limitations of using finished drinking water quality data from community water systems (CWSs) for EPHT applications, focusing on atrazine and nitrogen compounds in 8 Midwestern states.


Methods: Water quality data were acquired after meeting with state partners and reviewed and merged for analysis.


Results: Data and the coding of variables, particularly with respect to censored results (nondetects), were not standardized between states. Monitoring frequency varied between CWSs and between atrazine and nitrates, but this was in line with regulatory requirements. Cumulative distributions of all contaminants were not the same in all states (Peto-Prentice test P < .001). Atrazine results were highly censored in all states (76.0%-99.3%); higher concentrations were associated with increased measurement frequency and surface water as the CWS source water type. Nitrate results showed substantial state-to-state variability in censoring (20.5%-100%) and in associations between concentrations and the CWS source water type.


Conclusions: Statistical analyses of these data are challenging due to high rates of censoring and uncertainty about the appropriateness of parametric assumptions for time-series data. Although monitoring frequency was consistent with regulations, the magnitude of time gaps coupled with uncertainty about CWS service areas may limit linkage with health outcome data.