Estimation of ambient and non-ambient components of particulate matter exposure from a personal monitoring panel study

To provide additional insight into factors affecting exposure to airborne particulate matter and the resultant health effects, we developed a method to estimate the ambient and nonambient components of total personal exposure. The ambient ( or outdoor) component of total personal exposure to particulate matter ( PM) ( called ambient exposure) includes exposure to the ambient PM concentration while outdoors and exposure while indoors to ambient PM that has infiltrated indoors. The nonambient component of total personal exposure to PM ( called nonambient exposure) refers to exposure to PM generated by indoor sources and an individual's personal activity. We used data collected from a personal monitoring study in Vancouver, Canada to demonstrate the methodology. In this study, ambient PM2.5 exposure was 71% of the measured ambient PM2.5 concentration and was responsible for 44% of the measured total personal PM2.5 exposure. Regression analysis of the pooled data sets for ambient and total exposure against outdoor concentrations yielded similar slopes (0.76 for ambient and 0.77 for total) but a higher coefficient of determination for ambient exposure (R-2 = 0.62) than for total exposure (R-2 = 0.072). As expected, the nonambient exposure was not related to the ambient concentration (R-2 < 10(-6)). For longitudinal analyses of the relationship between measured personal exposure and ambient concentrations for individual subjects, the correlation of total personal exposure with ambient concentration yielded values of Pearson's r from 0.83 to - 0.68 with an average of 0.36. The relationship was statistically significant for only five of the 16 subjects. In contrast, the correlation of the estimated ambient exposure with ambient concentration yielded values of Pearson's r from 0.92 to 0.77 with an average of 0.88; 14 were significant. An example, taken from an epidemiologic analysis using the exposure data from this paper, demonstrates the usefulness of separating total exposure into its ambient and nonambient components.