Using GIS and historical records to reconstruct residential exposure to large-scale pesticide application

Investigation of pesticide impacts on human health depends on good measures of exposure. Historical exposure data are needed to study health outcomes, such as cancer, that involve long latency periods, and other outcomes that are a function of the timing of exposure. Environmental or biological samples collected at the time of epidemiologic study may not represent historical exposure levels. To study the relationship between residential exposure to pesticides and breast cancer on Cape Cod, Massachusetts, historical records of pesticide use were integrated into a geographic information system (GIS) to estimate exposures from large-scale pesticide applications between 1948 and 1995. Information on pesticide use for gypsy moth and other tree/vegetative pest control, cranberry bog cultivation, other agriculture, mosquito control, recreational turf management, and rights-of-way maintenance is included in the database. Residents living within or near pesticide use areas may be exposed through inhalation due to drift and volatilization and through dermal contact and ingestion at the time of application or in later years from pesticides that deposit on soil, accumulate in crops, or migrate to groundwater. Procedures were developed to use the GIS to estimate the relative intensity of past exposures at each study subject's Cape Cod addresses over the past 40 years, taking into account local meteorological data, distance and direction from a residence to a pesticide use source area, size of the source area, application by ground-based or aerial methods, and persistent or nonpersistent character of the pesticide applied. The resulting individual-level estimates of relative exposure intensity can be used in conjunction with interview data to obtain more complete exposure assessment in an epidemiologic study. While the database can improve environmental epidemiological studies involving pesticides, it simultaneously illustrates important data gaps that cannot be filled. Studies such as this one have the potential to identify preventable causes of disease and guide public policies.