You are here:
Fusing Exposure Data and Models to Reduce Uncertainty
Citation:
Wambaugh, J. Fusing Exposure Data and Models to Reduce Uncertainty. Presented at ToxForum Exposure for Safety Assessments of Consumer Products, Brussels, N/A, BELGIUM, May 20 - 22, 2019. https://doi.org/10.23645/epacomptox.8178833
Impact/Purpose:
This is a platform presentation as part of the ToxForum Workshop Determining Relevant Low-Level Chemical Exposures for Safety Assessments of Consumer Products in Brussels, Belgium in May 2019. This talk will lead off the session on "Experimental and/or Modeling Approaches."
Description:
Prioritizing the thousands of chemicals that occur in commerce with respect to potential risk posed to the public health requires tools that can estimate exposure from limited information. Human exposure biomarker data indicates that consumer products are a primary source of chemical exposures. Multiple exposure models exist for quantitatively predicting exposure from consumer products. To use high throughput consumer product models, additional data on consumer products are needed. While databases do exist to characterize the chemical content of consumer products, analytical chemistry surveys of products on the market have identified the presence of many additional compounds. Machine-learning models indicate that these chemicals are largely related to the compounds already known to be present in the products. The U.S. Environmental Protection Agency makes use of exposure models in a consensus, meta-model derived using the Systematics Empirical Evaluation of Models (SEEM) framework. The human exposure SEEM meta-model reconciles exposure predictions for various pathways (near-field (residential), dietary, far-field industrial, and far-field pesticide) with inferred chemical intake rates for the median U.S. population from the National Health and Nutrition Examination Survey (NHANES). Initial SEEM analyses identified the need for high throughput consumer exposure models, which have now reduced the uncertainty of the meta-model. The most recent SEEM analysis, using a dozen models and predictors of exposure, can explain 80% of the chemical-to-chemical variance in median intake rates. This abstract may not reflect U.S. EPA policy.
URLs/Downloads:
DOI: Fusing Exposure Data and Models to Reduce Uncertainty
TOXFORUMLLE-WAMBAUGH-DATAFUSION-052019.PDF (PDF, NA pp, 2498.474 KB, about PDF)