October 2008

Grant Results

SUMMARY

From 2002 to 2007, researchers at the Children's Environmental Health Initiative at Duke University in Durham, N.C., built a model based on geographic information systems (GIS) to integrate highly localized geographic data with other health, demographic and environmental data to analyze local health issues in six North Carolina counties.

Although the researchers used children's asthma as a representative illness, they designed the model to apply to other health conditions.

Key Results

  • The researchers built a GIS-based model with data from the counties of Chatham, Durham, New Hanover, Orange, Wayne and Wilson. The model integrated a base layer map of tax and census information with three overlays of information on health (including asthma statistics), environmental exposures (including samples of dust and other allergy-inducing substances from more than 500 households) and vital statistics on children.
  • Several local governments and health departments incorporated GIS-based modeling into daily activities such as environmental management, outreach and education campaigns, as well as program planning and evaluation.

Key Findings

  • Analyses of environmental samples indicated that the majority of homes had airborne and dust-borne allergens.
  • Through GIS-based modeling, the researchers determined that in the mostly rural areas surveyed, low-income neighborhoods or those with high numbers of minority residents did not suffer disproportionate exposure to airborne and dust-borne allergens.

Funding
The Robert Wood Johnson Foundation (RWJF) supported this project with a grant of $660,807 from January 2002 to June 2007.

 See Grant Detail & Contact Information
 Back to the Table of Contents


THE PROBLEM

Risk factors for childhood asthma include genetic predisposition and environmental exposures, according to the Centers for Disease Control and Prevention (CDC). Environmental exposures that cause asthma include house dust mite allergens and secondhand tobacco smoke. (An allergen is any substance that causes an allergic reaction.)

Geographic information systems (GIS) capture, store, manage, analyze and display data using specific geographical references. The technology offers the potential to understand the relationships among these factors, according to Project Director Marie Lynn Miranda, Ph.D., director of the Children's Environmental Health Initiative at Duke University. This initiative is a research, education and outreach program housed within the Nicholas School of the Environment and Earth Sciences at the university.

When this grant began in 2002, progress to explore fully the connections between geography — including neighborhood location, local topography and climate conditions and local zoning — and an individual's health risks had been limited because the data and systems available were not organized at an individual level of detail. For example, most databases were organized at the zip code, census tract or block level rather than the level of the individual tax parcel (i.e., housing lot).

Researchers at the Children's Environmental Health Initiative had begun developing such precise GIS-based modeling to support lead-poisoning prevention under a $415,000 grant from the CDC.

For that effort, they overlaid spatial data from North Carolina county assessors, the U.S. census and state blood lead-level screening databases to develop a model to predict — at an individual tax parcel level — the risk of children having high blood lead levels. County health departments then had the tools to target outreach, education and lead abatement efforts toward high-risk neighborhoods and residences.

Researchers at the Duke initiative saw another opportunity to use GIS-based models to identify locations where children might be at higher risk for asthma and where inadequate health care access might limit asthma treatment and management.

Childhood asthma was a serious health problem in North Carolina at the time: According to researchers at the Duke initiative, some 122,000 children had been diagnosed with asthma in 1997, and a 1999 North Carolina Adolescent Asthma Surveillance Study indicated that some 30 percent of children ages 13 and 14 had asthma-like symptoms with or without an asthma diagnosis.

Nationally, according to a 2001 study by the Trust for America's Health, some 90 percent of asthma health care costs in the United States go to treatment and not prevention. Project staff saw the use of GIS-based modeling as a tool to target and strengthen prevention efforts.

 Back to the Table of Contents


THE PROJECT

From 2002 to 2007, researchers at the Children's Environmental Health Initiative at Duke University developed a model based on geographic information systems (GIS) to integrate highly localized geographic data with other health, demographic and environmental data to analyze local health issues in six North Carolina counties.

Although the researchers used children's asthma as a representative illness, they designed the GIS-based model to apply to other health conditions as well. Throughout the project, project staff also worked with local health department staff and health care providers in all study counties both to develop their awareness and skills in using GIS-based modeling and to gather suggestions about how to improve the model and apply it to other health issues.

They also reached out to community-based partners such as churches and advocacy groups for their assistance in gathering information from the community, gaining access to residences to conduct environmental sampling and soliciting input about possible applications of the model to community health issues.

Their aim was to enable local health departments and medical providers to apply the model to other health problems, to target educational and outreach campaigns toward high-risk neighborhoods or households or to choose sites for new health care facilities.

The model integrated four layers of information:

  • A base layer map, including:
    • Tax parcel and tax assessor data on individual tax parcels, including age of housing and whether it was owner or renter occupied.
    • Street names and locations.
    • U.S. census demographic and socioeconomic data.
    • Public transportation routes, including schedules and stops, in those areas with public transportation systems.
    • City directory information, which helped identify commercial, industrial and residential properties by type and location.
  • A health overlay, including:
    • Data from the North Carolina Adolescent Asthma Surveillance study examining the prevalence of asthma, wheezing and environmental triggers in children ages 13 and 14.
    • North Carolina asthma hospitalization data.
    • Local health department and private practice asthma care data.
    • Locations, specialties and hours of physicians' offices.
    • Urgent care facilities.
    • Pharmacies.
  • An overlay on potential environmental exposures, including:
    • Tax assessor data on the presence of central air conditioning and slab or crawl space construction.
    • Floodplain data from the Federal Emergency Management Agency (FEMA) to help predict mold and moisture conditions in homes, as well as susceptibility to flooding.
    • Data from the researchers' own pilot project on allergen and asthma triggers in the home (conducted during the RWJF grant and funded by federal agencies).
    • Local air quality and pollution data.
  • An overlay of data on children, including:
    • Birth record information through an agreement with the North Carolina Vital Records Branch, including children's birth dates, addresses, race, gender and prenatal care history and parental age, race, education and marital status.

      (In combination with the tax assessor data, this information helped determine which houses were likely to be occupied by children.)
    • Locations of places that attract children, including schools, churches, recreation facilities, day care centers, parks and playgrounds.

Project researchers concentrated their data gathering on counties with large, medically underserved populations.

Other Funding

The U.S. Department of Housing and Urban Development also supported this project with a grant of $405,000 in 2002 that helped fund work to gather environmental samples from housing.

Communications

In addition to publications in scholarly journals, project staff made presentations at the 2004, 2005 and 2006 annual meetings of the American Public Health Association. (See the Bibliography.)

The project director had planned to create a Web site describing the project, with links to external resources. It was not created because staff found that the users of the materials at health departments were more comfortable with having them delivered directly to them and that the technology for a secured Web site was not as advanced at the time of the grant as it later became.

The staff also intended to create a manual outlining the framework for creating GIS-based applications for health care. This also was not produced.

Challenges

Acquiring patient data took much longer than project staff anticipated. Project Director Miranda attributed the delay to problems on the part of both project researchers and local clinicians in adjusting to the Health Insurance Portability and Accountability Act of 1996 (HIPAA). Part of this act aimed to strengthen the confidentiality of electronic medical records.

According to Miranda, the initiative conformed to all the standards and procedures outlined by HIPAA. All confidential electronic data resided on a completely private, password-protected network with no external access to ensure that data were inaccessible to unauthorized users.

Miranda reported that these security measures and project researchers' and clinicians' growing understanding of HIPAA regulations allowed the researchers eventually to gain access to enough patient data to complete the model.

 Back to the Table of Contents


RESULTS

Project staff described the results in reports submitted to RWJF:

  • The researchers built a GIS-based model to integrate highly localized geographic data with other health, demographic and socioeconomic data to analyze local health issues in the six North Carolina counties of Durham, Chatham, New Hanover, Orange, Wayne and Wilson. They made progress completing the data overlay maps as follows:
    • The base layer map and children's data overlay were fully constructed.
    • The health overlay included asthma cases and visits in Durham and Wayne counties as well as information on childhood obesity in Durham County.
    • To complete the overlay on environmental exposures, project staff visited some 350 homes, collecting interior bulk dust samplings (to detect allergens and asthma triggers from insects, rodents and other sources) and exterior bioaerosol samplings (to detect fungi, bacteria and viruses).

      Project staff also added data from 200 homes that were part of the allergens and asthma triggers pilot study.

      Staff completed the sampling, which included homes in the six counties, in 2004.
    • Project staff provided detailed sampling reports and educational information on asthma to participants whose homes were sampled.
  • Several local governments and health departments incorporated GIS-based modeling in daily activities such as environmental management, outreach and education campaigns and program planning and evaluation. For example:
    • The Alamance County Health Department used GIS-based modeling to prepare for emergency response situations and to plan and prioritize activities in areas ranging from environmental health to prenatal care.
    • The city of High Point used GIS-based modeling to target interventions to make homes more habitable and to gather information to support grant applications.
  • The researchers used GIS-based modeling to analyze student exposure to allergens and asthma triggers in an elementary school in Orange County. At the request of parents and teachers at the school, in spring and fall in both 2004 and 2005, project researchers collected bioaerosol samples from each room in the school. From the sampling, they built a GIS representation showing the variables that affected the school's indoor air quality.

    Project officials provided school officials with sampling results and recommendations on reducing student exposure to allergens and asthma triggers on school grounds.
  • Using the GIS-based model they developed, project staff members worked with Duke University Health Systems to determine the best locations for new fetal diagnostic centers across North Carolina.
  • As a result of this effort, Duke University Medical Center and Duke University Health Systems began a collaboration with the Children's Environmental Health Initiative to use spatial analysis and GIS to improve health outcomes in Durham County.

    As of June 2007, project staff was developing a system to assign a geographic location to every patient who interacted with Duke University Health Systems. Health system and project staff planned to link patient data and other health data with social and environmental factors in the Durham community.

Findings

The researchers reported the following findings in a July 2007 grant report to RWJF:

  • Analyses of bulk dust and bioaerosol samples indicated that a majority of homes visited had airborne and dust-borne allergens. This was of particular concern because approximately half the sampled homes had children living in them.

    Older homes and those located in forested areas were at higher risk of problematic bulk dust and bioaerosol levels.
  • Through the GIS-based modeling, the researchers determined that low-income neighborhoods or areas with high numbers of minority residents did not suffer disproportionate exposure to airborne and dust-borne allergens.

    Project Director Miranda attributed this unexpected finding to the impact of individual household behaviors on reducing exposure to allergens and asthma triggers. These behaviors included regular cleaning and leaving windows open.

    Miranda also conjectured that the results might have been different had the researchers done more sampling in densely populated areas where roach and rodent allergens are more prevalent rather than in mostly rural or semirural areas of North Carolina.

 Back to the Table of Contents


LESSONS LEARNED

  1. Start building partnerships from the beginning of the project. Involving local health departments, health care providers, community groups and other stakeholders and end users from the start of the grant period enabled the project to proceed more smoothly. In particular, this reaching out helped in gaining access to residences for sampling. (Project Director/Miranda)
  2. Build in sufficient time to comply with patient privacy regulations and policies. Project staff had anticipated some delays in gathering data on patients resulting from HIPAA's privacy requirements, but they found that the delays were greater than expected.

    The RWJF program officer reflected that the availability of public data from the state of North Carolina had facilitated the grantee organization's earlier work on lead-poisoning prevention. This project, in contrast, required patient information on asthma that local practitioners were reluctant to release. The project might also have benefited from more upfront data-sharing agreements with local health departments. (Project Director/Miranda, RWJF Senior Program Officer/Russo)
  3. Anticipate the need to make a case for new technological tools that can convince reluctant end users. Complete success of the project depended on widespread acceptance of GIS methods by medical practitioners and health departments. Although the grantee organization produced a viable model, both groups were slow to consider changing their methods of operation to integrate its use. RWJF Senior Program Officer Russo considered that the costs of gathering data might have outweighed what practitioners perceived as benefits of the modeling, especially in localities with tight budgets. Miranda noted that the perception about GIS has evolved since this project and more people are familiar with it and open to using it. In the end, she writes, "We found more interested users among larger organizations such as health systems." (Project Director/Miranda, RWJF Senior Program Officer/Russo)

 Back to the Table of Contents


AFTER THE GRANT

Project staff continued its work with Duke University Health Systems and local health departments on the application of GIS-based modeling to their operations. As of June 2008, project staff continued to analyze the sampling data and was preparing additional manuscripts for submission to peer-reviewed journals.

Russo cited the informatics programs RWJF has funded as the way the Foundation would make sure that the technological breakthroughs get into the public health arena. The programs are:

  • InformationLinks: Connecting Public Health with Health Information Exchanges. See Grant Results for more information.
  • Public Health Informatics Institute. See Grant Results for more information.
  • Public Health Informatics Fellows Training Program.

 Back to the Table of Contents


GRANT DETAILS & CONTACT INFORMATION

Project

Using Geographic Information Systems Approaches to Identify Asthma Risk Factors in Three North Carolina Counties

Grantee

Duke University, Nicholas School of the Environment and Earth Sciences (Durham,  NC)

  • Amount: $ 660,807
    Dates: January 2002 to June 2007
    ID#:  043524

Contact

Marie Lynn Miranda, Ph.D.
(919) 613-8023
mmiranda@duke.edu

 Back to the Table of Contents


BIBLIOGRAPHY

(Current as of date of this report; as provided by grantee organization; not verified by RWJF; items not available from RWJF.)

Articles

Miranda ML and Dolinoy D. "Using GIS-Based Approaches to Support Research on Neurotoxicants and Other Children's Environmental Health Threats." NeuroToxicology, 26: 223–228, 2005. Abstract available online. Article available for a fee.

Miranda ML, Hale BW, Anthopolos RA and Thomann WR. "Regional Variation in the Causes of Der p1 and Der f1." Unpublished.

Miranda ML, Hale BW, Overstreet MA and Thomann WR. "Crawl Spaces as Reservoirs for Transmission of Mold to the Livable Part of the Home Environment." Unpublished.

Xia G, Miranda ML and Gelfand A. "Approximately Optimal Spatial Design Approaches for Environmental Health Data." Environmetrics, 17(4): 363–385, 2006. Available online.

Presentations and Testimony

WR Thomann, ML Miranda, M Stiegel and MA Overstreet, "Shared Air: Examining the Contribution of Mold from Home Crawl Spaces to Home Interiors," at the Fifth International Conference on Bioaerosols, Fungi, Bacteria, Mycotoxins and Human Health, September 10–12, 2003, Saratoga Springs, NY. Proceedings available online.

ML Miranda, KL Atkinson, MA Overstreet and W Thomann, "Modeling Risk Factors for Childhood Respiratory Disease from the Physical and Home Environment," at the Annual Meeting of the American Public Health Association, November 6–10, 2004, Washington. Proceedings available online.

ML Miranda, W Thomann, BW Hale, M Stiegel, J Davis and MA Overstreet, "Co-Exposure to Bulk Dust and Bioaerosol Allergens and Asthma Triggers in the Home Environment," at the Annual Meeting of the American Public Health Association, December 10–14, 2005, Philadelphia. Proceedings available online.

J Davis, M Stiegel, W Thomann and ML Miranda, "Using a Multifaceted Assessment Tool to Determine Exposure Potential to Fungi for Children in an Elementary School in Orange County, N.C.," at the Annual Meeting of the American Public Health Association, November 4–8, 2006, Boston. Proceedings available online.

ML Miranda, MA Overstreet, JL Tootoo, AP Hull, D Kim, JA Davis, CN Wilson and S Doku, "Combining Geographic Information Systems and Web Based Mapping to More Comprehensively Inform Health Programs," at the Annual Meeting of the American Public Health Association, November 4–8, 2006, Boston. Proceedings available online.

 Back to the Table of Contents


Report prepared by: Paul Jablow
Reviewed by: Janet Heroux
Reviewed by: Molly McKaughan
Program Officer: Pamela G. Russo