Geospatial Approaches to Environmental Determinants of Breast Cancer in the California Teachers Study

With the increasing availability of georeferenced data about people and places there are expanded opportunities to assess risk relationships for disease outcomes associated with living in areas impacted by adverse environmental insults. For the most part, particularly for cancer outcomes, estimates of exposures have been based upon a single address and single point in time, and for which traditional epidemiologic methods such as interview may be impractical for assessing earlier residences in large scale studies. Likewise environmental data are often not available for specific time periods of interest or at an appropriate geographic scale for epidemiologic studies. Using an example from the California Teachers Study (CTS), a large statewide prospective study of women, we sought to develop strategies to assess temporal and spatial changes in participant residences, temporal changes in ambient air pollutants, and the intersection between the two in time and space to form a basis for future risk analyses. Key findings suggest the importance of careful attention to data quality, to assumptions about data gaps, and the influence of modeling decisions. Nonetheless, these efforts show promise for future environmental health research.