What Happened?
EPA’s scientists, with support from colleagues at HUD and CDC, published an impressive study identifying the nation’s potential lead exposure hotspots that warrant a deeper analysis for targeting lead actions. The map below shows 30,208 census tracts identified as the highest potential lead exposure risk locations based on one of five indexes and two statistical methods.
By looking at those areas that show up on at least four of the five indexes, they identified the nine states (Illinois, Michigan, New Jersey, New York, Ohio, Pennsylvania, Massachusetts, California, and Texas) and eleven counties with the highest potential lead exposure risk. Four of those counties are in New York (Bronx, Kings, New York, and Queens). The others are Cook County, IL; Wayne County, MI; Philadelphia County, PA; Essex County, NJ; Los Angeles County, CA; Harris County, TX; and Milwaukee County, WI.
Why it Matters
Virtually every community in the United States has a lead exposure risk because of decades of use in gasoline, paint, manufacturing, and consumer products. While our country continues to make progress reducing the risk, much remains to be done.
An important aspect of that effort is identifying hotspots—geographic locations with a higher prevalence of children’s lead exposure—that can be targets for more aggressive interventions. When scientists at EPA, HUD, and CDC, agencies with expertise on lead exposure, collaborate to provide a map of potential hotspots, public health agencies and communities have a useful tool to set priorities.
Our Take
Unleaded Kids is a big fan of the data analysis efforts of EPA’s Office of Research and Development (ORD) because they take a fresh look at the evidence, clearly acknowledge bias and limitations, and are strategically focused on delivering useful tools. This study is no exception.
For these reasons, we encourage federal, state, and local agencies as well as community-based organizations and their allies to make full use of the maps and tables in the study.
However, it is important to recognize the limitations of the analysis. In particular:
- Not being a hotspot does not necessarily mean the risk is low in your community. The national analysis is focused on the areas at the top of the various lead exposure indexes and models based on publicly available housing age and sociodemographic data. But, as research by EPA and by CDC has shown, lead exposure is pervasive. With no safe level identified, the goal is to drive all exposures closer to zero.
- Analysis is biased to housing-related potential lead exposures and is weak on risks from industrial or other environmental lead sources. EPA’s analysis is based on five indexes, all of which rely heavily on age of housing. Only two consider blood lead levels. See below for details.
- Blood lead levels of adults are not considered. The relative contribution of different sources of lead exposure for adults is significantly different than for children, especially for lead-based paint dust in the home. As a result, the hotspots may be quite different if the focus is protecting adults.
- Results may be skewed to urban areas and populous states. EPA used three of the indexes in a Michigan case study and found that they predominantly identify urban hot spots.
Fortunately, the researchers say they will continue to “ground-truth” the approach so they can refine it and incorporate more data such as drinking water lead service line locations and commercial sources of lead. They will also adapt the approach to various purposes depending on available resources.
Five Indexes Used by EPA
- EPA EJSCREEN Lead Paint which is based on percent of housing built before 1960, percent low income, and percent people of color.
- HUD’s Deteriorated Paint Index which is based on sociodemographic data from the 2009–2013 American Community Survey and pre-1980 housing data with reported areas of peeling paint from the 2011 American Housing Survey.
- Vox U.S. Lead Risk Exposure score which was adapted in 2016 from a Washington State Department of Health’s model and is based on age of housing and on poverty.
- EPA ORD’s “Schultz Model” which estimates children’s blood lead levels using percent of housing built before 1960, percent below poverty line, and percent non-Hispanic African Americans. EPA developed the model based on blood lead testing results in Michigan validated with data from Massachusetts, Texas, and NHANES.
- EPA ORD’s Random Forest Regression EBLL Prediction Model which is based on blood lead testing results from Ohio from 2007-11 and a 2013 Ohio Department of Health model. It uses five variables: 1) percent of homes built prior to 1940; 2) percent of homes built prior to 1950; 3) percent of families whose income-to-poverty ratio was greater than 2; 4) percent of population with either high school or higher education; and 5) percent of non-Hispanic African Americans in each census tract.
Next Steps
Unleaded Kids will monitor EPA’s continuing efforts to enhance and refine the model and will promote its use as appropriate.