Individual exposure to air pollution and lung function in Korea: spatial analysis using multiple exposure approaches

Area of Science:

• Environmental Health
• Epidemiology
• Geospatial Analysis

Background:

• Accurate individual-level air pollution exposure estimation is crucial for health studies.
• Existing methods for estimating exposure from ambient monitors vary, with limited evaluation of their impact on health risk assessments.

Purpose of the Study:

• To evaluate how different spatial interpolation methods for estimating air pollution exposure influence health effect estimates.
• To compare exposure estimation methods using lung function data (FEV1 and FVC) in a Korean cohort.

Main Methods:

• Applied multiple exposure estimation methods: all monitors average, nearest monitor, inverse distance weighting, and kriging.
• Utilized data from 13 monitors for PM(10), ozone, NO2, SO2, and CO in Ulsan, Korea (2003-2007).
• Assessed associations with lung function (FEV1, FVC) using linear regression, controlling for covariates.

Main Results:

• Kriging demonstrated the highest accuracy in cross-validation for exposure estimation.
• All air pollutants showed associations with Forced Vital Capacity (FVC) across all methods.
• Ozone exposure, estimated using kriging, was significantly associated with decreased FVC and Forced Expiratory Volume in 1s (FEV1).

Conclusions:

• Spatial interpolation methods, especially kriging, provide more accurate individual-level air pollution exposure estimates than monitoring values alone.
• These methods improve health risk assessments by capturing spatial variability and enabling exposure estimation in unmonitored areas.
• The choice of exposure estimation method can influence the magnitude and significance of observed health effects.