EPA-Expo-Box (A Toolbox for Exposure Assessors)
Since assessments using probabilistic methods are less common than those using deterministic ones, several probabilistic assessment examples are described below.
A Probabilistic Exposure Assessment for Children Who Contact CCA-treated Playsets and Decks Using the Stochastic Human Exposure and Dose Simulation Model for the Wood Preservative Scenario (SHEDS-Wood) (Zartarian et al., 2005).
EPA refined the SHEDS model (i.e., SHEDS-Wood) to estimate children’s exposure to chromium and arsenic residues from playsets and decks treated with chromate copper arsenate (CCA) (Zartarian et al., 2005). The SHEDS model is a probabilistic model that can be used to assess aggregate exposures and multimedia, multipathway chemicals of interest.
Using the probabilistic SHEDS-Wood model EPA was able to:
- address both variability and uncertainty in model inputs and outputs;
- generate a time series of exposure with high resolution of results;
- conduct comprehensive sensitivity analyses to identify critical model inputs and factors contributing the most to model predictions; and
- derive a more realistic estimate of exposure by incorporating specific information on activities related to exposure (e.g., diary surveys related to deck and playset contact) and addressing the effect of scenarios that reduce exposure (e.g., hand washing after play).
Details on the methods, assumptions, and results are provided in the report. The absorbed doses, including average daily doses (ADDs) and lifetime average daily doses (LADDs) presented in this report were used to generate estimates of children’s risk from exposure to chromium and arsenic residues in a separate report (Dang et al., 2003).
Comparison of Four Probabilistic Models (CARES, Calendex, ConsExpo, SHEDS) to Estimate Aggregate Residential Exposures to Pesticides (Young et al., 2012)
Young et al. (2012) compared the results of six exposure estimates including those derived from two deterministic approaches and four probabilistic approaches for predicting children’s indoor exposure to pesticides through inhalation, dermal, and non-dietary pathways for selected scenarios. The two deterministic approaches evaluated in this study were EPA’s Residential Standard Operating Procedures and the Draft Protocol for Measuring Children’s Non-Occupational Exposure to Pesticides by all Relevant Pathways (Draft Protocol). More information on the probabilistic models (i.e., CARES®, Calendex™, ConsExpo, and SHEDS) used in this study are found in Tools for Conducting Probabilistic Assessments.
The authors’ objective in comparing multiple approaches was to provide an indication of each model’s reliability, shed light on their underlying assumptions, and compare the range of exposure estimates that resulted from the different approaches (i.e., deterministic and probabilistic) and models. The paper provides a discussion of each approach and model including how the algorithms vary.
Young et al. (2012) found that the differences identiﬁed can be explained by activity assumptions, input distributions, and exposure algorithms. Smith (1994) published results comparing risk estimates for exposure to volatile solvents via ingestion and dermal contact. Smith (1994) compared risk estimates generated using probabilistic methods to those calculate using the deterministic “reasonable maximum exposure” approach and found that the deterministic estimate was in the high end of the range of probabilistic estimates.