Locating lead: mapping leads to intervention. (Science Selections).

Children's environmental health programs have been handicapped by the difficulty in identifying specific sites that pose the greatest risks and exposure potential within a larger geographic region. As a result, health programs such as those intended to reduce the threat of low-level lead exposure have had to focus on mitigating effects rather than implementing preventive steps. But a new mapping model developed by Marie Lynn Miranda and colleagues at Duke University's Nicholas School of the Environment and Earth Sciences identifies individual buildings and sites where the risk of lead exposure is statistically high, providing an intervention analysis tool for policy makers and public health officials [EHP 110:947-953]. The model looks at health data at the level of tax parcels (such as houses), which gives a degree of detail that should help health programs become better focused and more effective. The model uses a geographic information system, or GIS, to compile data the researchers gathered while studying childhood lead exposure risk in six North Carolina counties. The detailed spatial analysis incorporates blood lead screening results, county tax assessor records, and U.S. Census data.

Such detail is important because lead poses a complex threat to children's health that reflects many different risk factors, including the age of their homes, whether they live in rural or urban areas, their race or ethnicity, household income, and nutrition. For example, across the nation children from low-income families exhibit blood lead levels nearly eight times the rate of children from high-income families. However, unlike in other regions of the country, North Carolina's urban housing, much of which was built since the 1978 ban on lead in paint, generally poses less of a lead exposure risk than the state's older, rural housing.

Past studies of childhood lead exposure risk have captured only part of the picture. Some identified risk factors but didn't considered the relative weight for each factor. Others considered relative weights but didn't link analysis to geographic locations. Still others linked analysis to larger, aggregated areas without factoring in the relative weights.

Gathering the multiple types of data for this level of mapping was the biggest single challenge of the project. Two confidentiality agreements with the state's Childhood Lead Poisoning Prevention Program and the State Registrar's office provided the researchers with the data they needed for children born and screened between 1994 and 1999. Miranda says that such trust and cooperation with state and local officials was essential to the success of the project.

The resulting multilayer map of each county combined all screening data in a manner that allowed the researchers to analyze the relationship between observed blood lead levels and individual tax parcels, and compare that to data on the age of housing, median income, rental status, families in poverty, one-parent households, race, and other variables specific to each county. Color-coded maps were then generated showing which households or buildings were most likely to contain lead-based paint hazards, based on all the variables considered. The researchers are now in the process of collecting environmental samples to validate their models, and are applying this approach to other children's health issues such as asthma triggers, allergens, pesticides, and industrial contaminants.