Table 2.

Summary of TRAP exposure assessment approaches

Category	Approach	Overview	Strengths	Limitations
Modeling	Proximity	Distance to nearby roads serves as surrogate of exposure	Straightforward, minimal data requirements, potential to incorporate traffic volume/mix	Crude estimates of exposure may result in misclassification
	Dispersion	Emission, atmospheric, meteorological, and topographic data are used to mathematically model pollutant concentrations from traffic sources	Consideration of emissions, meteorology, source specific, high-resolution	Computationally intensive, requires assumptions regarding pollutant transport, extensive data including traffic characteristics, source emissions, fleet mix, meteorology, and topography often not available in study areas
	Land-use regression	Geographic variables measured in buffer regions surrounding multiple air sampling sites are used to estimate measured air pollutant variability, resulting regression equation used to predict pollutant concentrations at unsampled locations	Straightforward approach, able to capture high spatial variability of TRAP, can be applied to varying measured air pollutants, potential to incorporate temporal information	Requires air sampling at sufficient density to capture TRAP variability and geographic predictors, limited transferability of developed model
	Hybrid	Combination of multiple approaches, i.e. LUR with additional spatial interpolation models or personal monitoring	Improved exposure estimates	Requires data / expertise for multiple approaches
Personal	Personal monitoring	Individuals’ exposure to air pollutants is measured using wearable sampling devices	Direct measure of exposure	Requires wearable devices capable of accurately measuring exposure over long time periods, participant burden, cost of measurement devices, requires study personnel time
	Biomarkers	Concentration of internal biomarker, frequently in blood or urine, measured as marker of exposure or effect	Individual exposure assessment, potential use as measure of effect	Lack of specific markers, difficulty differentiating markers of effect from exposure, cost