HUMAN EXPOSURE MODELING FOR CUMULATIVE RISK
Impact/Purpose:
The goal of this research is to develop a research plan for the development of computational tools and approaches for characterizing exposure to facilitate assessment of cumulative risks. To achieve this goal, the following research questions will initially be addressed:
1) What information exists and is needed to characterize spatial and temporal patterns of exposure to multiple chemicals for vulnerable populations?
2) What approaches are available or are needed to collect, interpret, and present exposure information for cumulative risk assessment?
3) How can we stratify existing data to identify exposure metrics to classify individual and populations for epidemiological studies, public health tracking, and cumulative risk assessment?
4) What available methods are best for characterizing variability and uncertainty in exposure models for cumulative risk assessment?
5) What available methods and approaches are best for evaluating human exposure models for cumulative risk assessment?
Description:
US EPA's Office of Research and Development (ORD) has identified cumulative risk assessment as a priority research area. This is because humans and other organisms are exposed to a multitude of chemicals, physical agents, and other stressors through multiple pathways, routes, and sources under a variety of temporal and spatial conditions. In order to assess cumulative risk, the traditional paradigm of assessing risk from a chemical or facility basis will need to shift to a community/population or individual basis. Typically, when estimating population exposures, risk assessors do not simultaneously track an individual's exposure to multiple chemicals and do not encompass factors other than chemical exposure that can contribute to greater adverse health effects such as individual-specific factors (e.g., compromised health), exposure to non-chemical stressors, and exposures accumulated over time. Many different types of data are required and these data need to be collected, interpreted, and presented in new ways to classify populations and characterize exposures.
A key role of science at EPA is to reduce uncertainties in the information used for environmental decision-making (EPA/600/9-91/050). Precisely capturing and interpreting both the variability inherent to population exposures and the quantitative evaluation of uncertainty within exposure estimates is required. Few methods exist to quantitatively evaluate both variability and uncertainty of exposure and cumulative risk estimates, and, to date, those methods have not been extensively used. This is mostly related to issues of computational efficiency and reliability, but partly due to a general lack of understanding of the basic concepts and how, when, and where to apply them.
EPA's National Exposure Research Laboratory (NERL) human exposure modeling research program has conducted research over the last 6 years to develop person/population-oriented exposure models (e.g., Stochastic Human Exposure and Dose Simulation (SHEDS) model). In this task, we develop a research plan to build on this strong foundation to address important needs for quantitatively assessing cumulative risks to potentially vulnerable populations. The ultimate outcome following research plan implementation will be computational tools and approaches to perform appropriate sensitivity analyses, model evaluations, and exposure modeling. These tools and methodologies will help identify important model parameters, improve model structure, and provide reliable exposure estimates for use by regulatory decision-makers in assessing cumulative risks.
Record Details:
Record Type:PROJECT
Start Date:10/01/2004
Projected Completion Date:09/01/2006
OMB Category:Other
Record ID:
135584
Keywords:
SHEDS, CUMULATIVE, SENSITIVITY, UNCERTAINTY, EXPOSURE, MODELING,
Project Information:
Progress
:Over the last 6 years significan research has been conducted to develop the SHEDS human exposure models. SHEDS models are used to evaluate population exposure and dose to environmental pollutants (i.e., pesticides, particulate matter, air toxics). SHEDS models use a physically-based probabilistic approach to predict population exposures to these pollutants. There is strong emphasis on the human element of exposure since sequential time-location-activity data from diaries contained in the Consolidated Human Activity Database (CHAD) are combined with concentrations of pollutants in various exposure media, as well as other physical factors, to generate time series of exposure for simulated individuals. A two-stage Monte Carlo simulation technique is used to produce distributions of exposure for various cohorts (e.g., age/gender groups) that reflect both the uncertainty and variability in the input parameters. The SHEDS model and other similar models currently estimate exposure and resultant dose from a single chemical only. The SHEDS models will serve the primary research tool to develop and evaluate the newly developed methodologies.
Research on developing approaches for exposure modeling for cumulative risk has been initiated in FY04 as part of task 3948:
In the SHEDS-Wood model, bootstrap methods were used to generate the inputs for uncertainty distributions to solve the issue of correlation between the parameters (shape factor, scale factor, etc.) for given distribution types, particularly for beta, gamma, and Weibull distributions. In addition, correlation, regression, and variance analysis were used to interpret the SHEDS model two-stage Monte Carlo simulation technique to identify uncertainty attributed by input parameters, key inputs, and input data gaps.
A module is currently being developed that can impose rank correlation structure on input variables. This module is to be employed in population-based exposure models to assist in reliably estimating exposures to multiple high priority pollutants whose emission/usage is correlated (e.g., active ingredients in various pesticide formulations and other pollutant mixtures in consumer products). Note that the module can also be applicable to other models and other model applications where correlation of input parameter distributions is desired (e.g., PBPK models-heart rate, respiration rate, blood flow, metabolism)
An investigation of efficient computational methods to retain annual exposure profiles to multiple chemicals was conducted to evaluate how simultaneous exposures can be tracked on an individual-by-individual basis. Report submitted - "Storage and Retreival of Human Exposure Simulation Data in Scenarios with Multiple-Simultaneous Exposures". Advantages and limitations of data organization and storage, data reduction, and the formation of metadata based catalogue systems as applied to this issue were described.
Relevance
:NERL's human exposure modeling program has already positioned considerable effort on the human element in exposure modeling. The research conducted under this task will now address needs identified in the Framework for Cumulative Risk Assessment (EPA/630/P-02/001F) and at the Region/ORD Workshop on Cumulative Risk Assessment (November 2002). In this task we will develop a framework and strategy for addressing the most important research needs in human exposure modeling to increase the scientific basis for conducting cumulative risk assessment. Following implementation of the research plan, risk assessors will be able to more accurately estimate both aggregate and cumulative risk. Improved and new approaches to assess variability and uncertainty will be useful in identifying data gaps and producing more reliable exposure distribution estimates, thus assisting risk managers in regulatory decision-making. Newly developed techniques are also anticipated to be applicable to other exposure and dose model development efforts inside and outside of EPA.
The need for this type of research is incredibly high given the novelty of exposure modeling tools and techniques, and the high demand for their use. Assessing cumulative risk through complex exposures is one of EPA's highest priorities, particularly considering the Food Quality Protection Act mandates, and it is germane and of great interest to all EPA program and regional offices. Previous and ongoing discussions regarding this research has included both research scientists and risk assessors within ORD, the Office of Air Quality Protection and Standards (OAQPS), the Office of Pesticide Programs (OPP), the Office of Pollution Prevent and Toxics (OPPT), the Office of Transportation and Air Quality (OTAQ), interagency partners (e.g., Lawrence Berkeley National Laboratory) and the scientific community.
Clients
:OPP, OPPT, OAQPS, OTAQ
Project IDs:
ID Code
:20597
Project type
:OMIS