Presentation Abstract

Addressing the Uncertainties in the Epidemiologic Studies-Exposure Assessment
L.-J. Sally Liu
Northwest Center for Particulate Air Pollution and Health,
University of Washington, Seattle, WA

One of the major criticisms on PM epidemiological studies is exposure misclassification. Since people spend the majority of time indoors and total exposure often exceeds indoor and outdoor measurements, it was well documented in earlier studies that personal exposure to PM_2.5 was not correlated with those measured at centrally located monitoring sites. Thus, time-series studies using central site measurements for health assess ment may contain large uncertainties in exposure estimates. Recent panel studies have shown correlations between personal exposure and ambient measurements within individuals; these correlations do not differ by groups with various health conditions (e.g., normal vs. susceptible populations). Total personal exposure to PM_2.5 has been further broken down into exposure to ambient generated particles (E_ag, such as secondary aerosol, vehicle exhaust, wood smoke, road dust, that are regulated by the EPA) and non-ambient generated particles including particles produced by indoor and personal activities (cooking, vacuuming, etc). E_ag accounts for more than 50% of total personal PM_2.5 exposure, explaining the observed correlations between personal exposure and ambient measurements within individuals. E_ag is dominated by home ventilation and can be predicted by a recently developed model that accounts for time-activity pattern and estimated particle infiltration efficiency. Although secondary aerosol tend to be spatially homogenous, PM_2.5 and ultrafine PM from local combustion sources have been shown to distribute unequally throughout the city. Although central site measurements are good surrogates for personal exposure to total PM_2.5 mass, they may not predict well personal exposure to various PM sources. Thus, challenges lay ahead for studies trying to further our knowledge by using source estimates at central site to predict exposure and health effects across the city.

Top of page