Mapping Environmental Health

Marie Lynn Miranda Uses Geospatial Technologies to Protect Our Children p.3

Lead poisoning can lead to serious disease that doesn’t become apparent in children until long-term, irreversible effects already have set in. Children exposed to lead levels far below what was once considered safe may be asymptomatic, but they can eventually develop learning and behavioral disorders, hearing impairment, decreased IQ, and decreased attention span, she said.

In 1998 Miranda received $25,000 from the National Institute of Environmental Health Sciences (NIEHS) in seed money and then major funding from the Centers for Disease Control to use GIS to create a household predictive model of lead exposure risks across the state.

GIS analysis has been widely explored for environmental sciences as well as for public health purposes. It works because most data contain a geographic component that can be tied to a specific location—a zip code, a state, a county, a single address.

Users can then overlay data by location and expose trends that might not be readily available in traditional spreadsheet software. What’s more, they can use GIS to generate maps and reports that can serve as the basis for developing policies and for doing community outreach.

To accomplish what Miranda envisioned, she and her team needed to develop methodologies to bring together information at a higher geographic resolution than previous public health GIS analysis had done. It wasn’t enough to generate information at the block level. To develop a predictive model that would be useful for practitioners who aimed to create prevention programs, she needed to find a way to work at the individual house level.

“Because the work is done at this very high level of geographic resolution, you can have a much more carefully tailored program. And that means for every dollar that you spend in your environmental public health programs, you can go beyond the census track level and start thinking about placing your priorities on a house-by-house basis,” said Miranda.

To construct a predictive model with a risk index using GIS technology and spatial analysis, Miranda and her team drew from county tax assessor data, U.S. Census demographic data and North Carolina blood lead screening data for six North Carolina counties: Buncombe, in the western portion of the state, Durham and Orange in the central piedmont, Wilson and Edgecombe in the eastern coastal plain, and New Hanover on the southeast coast.

Then once they had built a preliminary model, they sent a group to do house-by-house environmental sampling to enable them to validate and calibrate the model with what they found in the field.

Her team members are research associates Alicia Overstreet, data manager; Michelle Abrams, project manager; Christine Bradshaw, GIS programmer; Jennifer Silva, community relations manager; Dana Dolinoy, who is now at Harvard University working on her masters in public health; and field research associates Lyle Whitney and Matthew Stiegel. Several team members hold degrees from Duke: Overstreet, Abrams and Dolinoy graduated with bachelor’s degrees in environmental science and policy; Overstreet received a bachelor’s in biology; and Whitney received the Nicholas School’s masters of environmental management degree.

“We’ve been collecting environmental samples from mid-April to mid-October this summer and last summer. Our initial samples collected from 500 houses indicate pretty tight model validation. I have a lot of confidence in the lead model right now, and a paper on our preliminary results came out in the September issue of Environmental Health Perspectives.”

Sampling is a labor-intensive exercise that involves sending out hundreds of letters to homeowners asking them to allow a team from the Children’s Environmental Health Initiative to take samples in their homes. Out of 300 letters they might get 20 positive responses.

This summer, research associates Whitney and Stiegel were joined by state environmental specialist Alan Huneycutt driving in a white, equipment-laden van for dawn-to-dusk sampling.

page 1 | 2 | 3 | 4