Recorded presentation

Carrie M Kuehn, MA, MPH, Cancer Epidemiology Research Cooperative, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue N., Box M4-C308, P O Box 19024, Seattle, WA 98109, 206-667-6171, ckuehn@fhcrc.org, Beth Mueller, DrPH, Department of Epidemiology, University of Washington, 1100 Fairview Avenue N., Box M4-C308, P O Box 19024, Seattle, WA 98109, Harvey Checkoway, PHD, Environmental Health, University of Washington, Box 357234, Seattle, WA 98195, and Craig Erickson, Information Resource Management and GIS, Washington State Department of Health, 1102 SE Quince Street, Olympia, WA 98504.

Geographic information systems (GIS) software was used in a population based case-control study examining maternal residential proximity to nearest hazardous waste site and malformation occurrence among offspring. Subjects were identified from Washington State birth records. Hazardous waste site locations were identified using information from the WA Department of Health. The straight-line distance between geo-coordinates (latitude/longitude) of the maternal residence (per birth certificate) and the nearest hazardous waste site was estimated for all subjects as a proxy for potential exposure to teratogens. No coordinates were available for 7,231 (11%) of 63,006 cases and 43,706 (14%) of 315,030 controls. Subjects with and without coordinates were similar with respect to maternal smoking, alcohol consumption, marital status, age, and parity. Differences between those with and without coordinates included rural residency (29% and 46%, respectively) and maternal college education (33% and 19%, respectively), however, no case-control differences in these characteristics were observed. Subjects without coordinates were also more likely to have missing data for insurance status (14% of those with geo-coordinates vs. 3% of those without). Excluding non-geocoded cases and controls may bias results if differences related to lack of geo-codeable address are also associated with the exposure or outcome. Estimating environmental exposure in case-control studies by this method is limited not only by the accuracy of the GIS procedure in measuring distances, and use of proxy measures of exposure to chemicals or combinations of contaminants, but also by the availability of geo-coordinates in historical data.

Learning Objectives:

• Understand how GIS can be used to estimate exposure to potential teratogens in case-controls studies.
• Recognize how subjects with and without geographic coordinates differ with regard to important study variables.
• Describe how differences between those with and without geographic coordinates can potentially bias results in case-control studies.

Keywords: Geographic Information Systems, Environmental Exposures

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.

Recorded presentation