Overview
This study compared two concurrent lead monitoring programs operating simultaneously in Pittsburgh, Pennsylvania: Lead and Copper Rule compliance sampling at Tier I high-risk sites and a voluntary customer-requested testing program serving the broader city. The Pittsburgh Water and Sewer Authority exceeded the US EPA Lead and Copper Rule action level of 15 µg/L at Tier I sites from 2016 through mid-2019, triggering corrosion control review. pH adjustment with caustic soda proved inadequate; orthophosphate dosing initiated in April 2019 reduced the 90th percentile Tier I concentration from 22 µg/L in June 2016 to 5.1 µg/L by June 2020. The study characterized both programs statistically to assess their representativeness and complementarity.
Key numbers
LCR Tier I 90th percentile concentrations (first-draw 1-L, µg/L):
- June 2016 (action level exceedance): 22 µg/L
- 2017: 17 µg/L
- 2018: 15.4 µg/L (borderline)
- 2019 (pre-orthophosphate): 12.8 µg/L
- June 2020 (post-orthophosphate): 5.1 µg/L
Historical LCR 90th percentiles (Tier I, µg/L): 2007: 9.0; 2010: 10.4; 2013: 14.8.
Customer-requested 90th percentile concentrations (µg/L): 2016: 15.8; 2017: 9.0; 2018: 11.0; 2019: 7.6; 2020: 4.2. Medians were below the reporting limit (1 to 4 µg/L depending on laboratory) in all years; 56 to 76 percent of customer-requested samples were non-detect.
IEUBK context (cited from Triantafyllidou et al. 2014, as reported in the paper): at 7 µg/L, 25 percent of 1 to 2 year old children drinking 500 mL per day would exceed a blood lead level of 5 µg/dL; at 4 µg/L, 5 percent of infants consuming 800 mL per day of reconstituted formula would exceed that threshold.
Methods (brief)
LCR sampling: first-draw 1-L cold tap water from kitchen or bathroom faucet after 6 h or more stagnation, Tier I sites selected per LCR (homes with lead service lines or lead interior plumbing). Customer-requested: resident-collected 1-L first-draw samples submitted voluntarily to the utility. Multiple contracted laboratories; reporting limits 1 to 4 µg/L (laboratory-dependent). Non-detect values imputed via semi-parametric regression on order statistics (ROS) using a log-normal distribution. Total lead by ICP-MS or graphite furnace atomic absorption.
Evidence fitness
Supports: tracking the effect of corrosion control treatment on 90th percentile compliance concentrations; illustrating the difference in population served by LCR high-risk sampling versus broader community sampling; quantifying lead concentration distributions across multiple monitoring years at city scale.
Does not support: individual home risk assessment; identification of lead service line locations; real-time exposure estimation; generalization to other utilities without comparable plumbing age profiles.
Limitations
LCR Tier I sites are selected for high plumbing-lead risk; they are not a random sample of city homes and overrepresent worst-case exposures compared to the customer-requested cohort. Reporting limits varied across laboratories and years, complicating direct inter-year comparison of non-detect frequencies. The customer-requested program was self-selected (residents motivated to test may differ systematically from non-testing households). The study covers one utility’s experience with one corrosion control intervention; orthophosphate efficacy depends on water chemistry and plumbing configuration.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.