In the early days of risk assessment, the end result was typically a single risk estimate for the maximally exposed individual. Furthermore, the estimate for the maximally exposed individual was often an overestimate of any actual exposure. As risk assessment has evolved, statistical methods have been developed to characterize exposure and risk across the full distribution of the exposed population. These methods are referred to as probabilistic risk assessment (PRA). Understanding the full distribution of exposures and risks provides risk assessors with substantially more information to make decisions. It also allows risk assessors to more accurately estimate aggregate risks (i.e., combined risks across different exposure pathways), avoiding the inaccurate summing of high-end exposure and risk estimates across each pathway. EPA and other regulatory agencies have moved increasingly toward probabilistic risk estimates across all its programs.
Exponent has been a leader in applying probabilistic risk assessment methods. Our contributions in this area include:
- Development of mathematical models to estimate dietary and aggregate risks for pesticides, including DEEM™, FARE™, and Calendex™. These models are widely applied for EPA pesticide regulatory applications.
- Development of the Probabilistic Exposure and Risk model for FUMigants (PERFUM), which is a mathematical model for estimating bystander exposure following fumigant applications. The EPA Pesticide Program applies this model.
- Routine use of EPA’s probabilistic risk assessment methods for the evaluation of risks at Superfund sites and other hazardous waste sites.
- Evaluation of human health risks to anglers who consume fish from contaminated water bodies, by characterizing their fishing and consumption behavior using probabilistic methods, including event-by-event exposure assessments, instead of simple point estimates.
- Exponent staff stay current on EPA’s guidelines regarding probabilistic assessments and the tools used to perform such assessments (e.g., @Risk, CrystalBall).