Publications : 2022

East, A, Dalton C, Egeghy P, Vallero D. Estimating exposure across chemical, individual, and media using sparse summary statistics with the Lorber-Egeghy-East Method R Package. Poster presented at Society of Toxicology Annual Meeting, San Diego, CA, March 2022. DOI: 10.23645/epacomptox.20110814, open access online.



Limited data are available to assess potential chemical risks to humans from manufacture, use, and disposal of consumer products and articles.  Tools are needed to access and leverage available data on chemical manufacture, use, and occurrence for important chemical exposure scenarios and pathways across the product lifecycle.  Scientific workflows are designed to execute a series of computational or data manipulation steps.  The simplest automated scientific workflows are scripts that call in data, models, and other inputs and produce outputs that may include analytical results and visualizations.  The value of using this approach is that domain-specific data types and tools can be made available to the exposure scientist and easily accessible to the exposure assessor for specific decision contexts. This product provides regulatory scientists, students and researchers with the ability to effectively access and exploit the many in silico data streams to support different regulatory purposes and supports current Agency efforts to reduce mammal study requests by 30% by 2025, and completely eliminate all mammal study requests and funding by 2035


Deterministic exposure estimates across media and chemicals are the product of pollutant concentrations and a series of pathway-specific exposure factors. These estimates are strengthened with greater data resolution and more data sources. However, published measurement studies commonly report only sparse summary statistics, with inconsistent measures of central tendency and variance. To reconcile this issue, the Lorber-Egeghy-East Method (LEEM) R package was developed. The model accepts a spreadsheet of available summary statistics as its input and produces synthetic lognormally distributed points for each medium, individual, and chemical entered to the model. Studies may also be ‘weighted’ (for example, by sample size) to be more influential. To demonstrate, summary statistics for 12 per- and polyfluoroalkyl substances (PFAS) – 10:2 FTOH, 6:2 FTOH, 8:2 FTOH, PFDS, PFHpA, PFPeA, PFUdA, 4:2 FTOH, PFDoA, PFHpS, PFNS, and PFPeS – were extracted from scientific literature published between 2011 and 2020. Intake was estimated from dust, water, and air, and soil for population groups, i.e. children, women, and adults. Results are sample-size weighted. The output contains raw data generated for each individual (by PFAS and pathway), summary statistics, the data used, the data not used, and an inventory of how many datasets were used to create each lognormal distribution of points. Abstract does not necessarily reflect EPA policy.