Exposure factors: How to characterize the data?
Öberg, T., Filipsson M, Bergbäck B.
Posterpresentation vid the Annual Meeting of the Society for Risk Analysis, San Antonio, Texas, December 9-12, 2007

Abstract
The exposure to chemicals is, apart from level of contamination and release, controlled by human behavior, physiological characteristics and environmental factors. The quantifiable statistics used to describe this information are called exposure factors, and are instrumental for any risk assessment. Evaluated data on exposure factors are often accessed from the literature or compilations such as the Exposure Factors Handbook from the U.S. EPA. However, these factors may differ between countries and regions, and we have recently collected and evaluated a similar range of country-specific data from Sweden. Our compilation includes the population age distribution, moving patterns, body weights, skin surface areas, activity patterns, dietary intakes and home produced items, drinking water consumption and private wells, residential building characteristics, and soil parameters. One problem that we faced in this study was the choice on how to best describe and present the available data. Some examples from previous compilations are arithmetic means, standard deviations, percentiles, and specific parametric distributions. We have in earlier work highlighted the uncertainty in choosing distributions, and the overconfidence that is conveyed in assigning single numbers or specific distributions. In the present study, we have instead chosen to estimate and present both parameters (mean, standard deviation, kurtosis, and skewness) and multiple percentiles (1-99%) with uncertainty ranges (95% confidence intervals). These uncertainty estimations were performed by resampling (bootstrapping) from the primary data. This tabulated compilation can serve as a basis for estimating exposure with several different approaches; deterministic, intervals, Monte Carlo-simulations, or probability bounds analysis. In this poster we will exemplify and discuss how the compilation and presentation of data on exposure factors affect the methodological choices and outcome of an exposure assessment.

Poster as a PDF-file, 78 kb