Comparison of Approaches for Developing Distributions for Carcinogenic Slope Factors. C. P. Petito Boyce, PTI Environmental Services, 15375 SE 30th Place, Suite 250, Bellevue, WA 98007
In recent years, risk assessors have increasingly been moving away from deterministic risk assessment approaches (i.e., approaches in which risk calculations are based on single point estimates for each necessary input parameter) and are moving towards probabilistic approaches (e.g., Monte Carlo simulation) that incorporate distributions of possible values for each input parameter. To date, most efforts in developing parameter distributions for use in probabilistic risk assessments have focused on exposure parameters (e.g., distributions of body weights or fish consumption rates in a population of concern). By contrast, efforts to develop approaches for incorporating variability and uncertainty in toxicity factors into risk calculations are in a preliminary state of development. As one component of exploring the use of probabilistic risk assessment approaches in calculating human health sediment quality criteria (HHSQC) in the state of Washington, available approaches for developing distributions of carcinogenic slope factors (CSFs) were compiled and evaluated. The following three frameworks for developing distribution of CSFs were identified and reviewed: those approaches operating primarily within a statistical framework, those applying a decision analysis framework, and those structured within a biological framework (e.g., physiologically based pharmacokinetic [PB-PK] models). The evaluation of these approaches included consideration of the technical and statistical validity of the available approaches. The evaluation also compared the results obtained using the alternative approaches and identified issues that would be encountered in applying the available approaches to derive distributions of CSFs for use in developing HHSQC.