Validity of geographically modeled environmental exposure estimates

Area of Science:

• Environmental Epidemiology
• Geographic Information Systems (GIS)
• Exposure Science

Background:

• Geographic modeling is widely used to estimate long-term environmental exposures in chronic disease epidemiology.
• The accuracy of these models is often unverified against measured concentrations, personal exposure, or biologic doses.

Purpose of the Study:

• To discuss epidemiologic associations using geographically modeled exposures.
• To identify critical issues affecting the reliability of geographically modeled exposure assessments.

Main Methods:

• Review of three case studies: air pollution (fine particulate matter), magnetic fields, and pesticide exposure (paraquat).
• Analysis of validation challenges and potential biases in geographic exposure modeling.

Main Results:

• Air pollution models show moderate correlation between ambient and personal exposures.
• Magnetic field modeling has high error at short distances.
• Pesticide exposure validation studies yield inconsistent and modest results.

Conclusions:

• Geographic exposure models require rigorous construction and validation for credible scientific results.
• Current models, without thorough validation, may not reliably detect causal associations between environmental exposures and chronic diseases.