Housing age and poverty rank childhood lead risk for all 83,388 US census tracts

Each tract's score rests on two inputs. Lead paint for home use was banned in 1978, so the share of housing built before then stands in for paint that may still be on the walls, and higher poverty adds risk on top ^[3]. The EPA research behind the method suggests both predict childhood lead exposure ^[2]. Ritter published the national version in GeoHealth in 2026 ^[13].

Michigan and Ohio aside, nine states tested the score independently ^[15]. In all 11, higher predicted risk went with higher measured blood lead in children ^[9]. If the tract-level correlations are ordinary Pearson coefficients, squaring them puts the variation shared between score and blood lead at roughly 18 percent in Minnesota and 41 percent in Wisconsin ^[16]. For a score built from two inputs, those are respectable figures. In Minnesota, about four-fifths of the variation still sits outside the score ^[16].

The county results run higher, from 0.48 in New York to 0.77 in New Jersey ^[11]. A county is a coarse unit. The average US county holds about 26 tracts ^[14]. A strong county correlation says the score finds a state's worse counties. It says less about which of those 26-odd neighborhoods a testing team should visit first.

The thing a correlation doesn't tell a health department is how the top of the list performs. A team with a fixed number of testing days needs to know what share of children with elevated blood lead live in the highest-scoring tracts, because those are the tracts it can reach. The Eos summary of the paper reports only correlations.

Geography is the larger gap. All 11 validation states have older housing and sit in the Northeast, Midwest and Great Plains, so the map's low-risk predictions for the West and Sun Belt have not been tested ^[7]. The score sees only housing age and poverty ^[2]. An exposure route that follows neither would not raise a tract's score in any region.

Ritter founded the lead testing company Fluoro-Spec ^[1]. The paper presents the map as a neighborhood screening tool ^[12], and I think the evidence supports that use in the 11 states where it was checked ^[6].