A Decision Support Software Package
RiskFX is a comprehensive human health risk assessment tool intended for use in environmental investigations and cleanup. The package is designed to provide the following capabilities:
Human Health Risk Assessment for Hazardous and Radioactive Contaminants
RiskFX is the commercial outgrowth of the DOE code, the Probabilistic Risk Evaluation and Characterization Investigation System, or PRÉCIS, developed by Sandia National Laboratories. RiskFX is a computer code designed to aid operable unit leaders, responsible parties, stakeholders, and regulators in establishing concentration thresholds for contaminants in soils, evaluating the potential human health risk a site poses to the public, as well as prioritizing the site characterization needs. A probabilistic risk assessment methodology is the underlying foundation for the RiskFX code, and its use is in the quantification of endpoints in the decision making process (e.g., risk, action levels, Data Quality Objectives or DQOs). The code is based on several key elements:
RiskFX has been developed for both the PC Windows and Macintosh platforms. RiskFX is intended to be used to quantify the uncertainty in human health risk. Nine environmental pathways are available for simulation of contaminant transport to a receptor. Both hazardous (150+ chemicals) and radioactive (48 nuclides) constituents are accounted for in the analysis. Many realizations are performed to translate the uncertainty in the input parameters in the code into the uncertainty in risk. Each realization also yields an estimate of the soil concentration threshold based on a specified acceptable risk. In other words, if the acceptable risk is 10-6 increased incidence of contracting cancer, then the code performs an inverse calculation to predict the maximum allowable soil concentration that would yield that risk. The combined effects of multiple constituents are taken into account also.
Concentration thresholds can be used to guide the adequacy of the field screening equipment (i.e., sensitivities), to allow an early determination of whether a site is safe (resulting in a No Further Action proposal early in the assessment process), and to determine which constituents are of concern. The sensitivity analysis option allows the user to prioritize the waste site characterization needs based on risk. Another benefit of this tool, when used early in the assessment process, is the identification of sites that pose an unacceptable risk. These sites are candidates for remedial alternatives selection, and therefore characterization needs would be for optimizing the preferred alternative. RiskFX may also be used to evaluate the risk associated with the potential remedial alternatives. This code is a must for persons planning site investigations.
RiskFX was designed for ease-of-use. It has an intuitive graphical user interface (GUI) and real-time graphical display of results. You can view many of these GUI elements in a RiskFX tutorial within this web site. The following graphic illustrates the graphical nature of the interface, showing the concept of selecting environmental pathways through a point and click selection method.
Additional environmental pathways are under development for inclusion in RiskFX. These include dermal contact and vapor-phase contaminant migration/exposure (potentially important for hydrocarbon and solvent contaminant analyses).
Other GUI constructs include popup menus and dialog boxes. One of the software features that DecisionFX prides themselves on is the utilization of uncertainty concepts in their software packages. In RiskFX , as in other of our packages, we utilize Latin Hypercube Sampling (LHS) to perform sampling functions for Monte Carlo analysis methods. The short of this functional attribute is that uncertainty analyses can be performed much more efficiently than with conventional methods. The computational burden is therefore greatly reduced. The following graphic illustrates some of the user-friendly concepts employed in RiskFX while implementing the LHS techniques for defining parameter uncertainties and correlations amongst variables.
Analysis of the results of a risk analysis with the RiskFX package is essentially instantaneous following completion of the simulations. The user can choose from many combinations of plot types to evaluate in order to evaluate total risk, risk from specific exposure pathways, risk from specific chemicals, cumulative effects, estimated soil cleanup concentration limits, or practically any combination thereof. The following graphic illustrates some of the options available.
The results of a probabilistic risk analysis are quite useful in determining the ultimate disposition of a waste site. With a quantitative expression of uncertainty one can look at the probability of meeting a given regulatory threshold in order to support decision making. The following graphic illustrates this concept. Each curve on this graph represents a different set of data, or assumptions, or sites in a hypothetical results to illustrate possible outcomes. The blue curve suggests that for the data, parameters and assumptions used in the analysis there is a 100% likelihood that the risk is greater than 10-8, but a zero percent probability that it is greater than 10-6. The EPA has set the 10-6 risk limit as the point of departure for risk management decision making. Any site with a less than 10-6 risk does not require cleanup. Therefore, this analysis shows that the site in question should be proposed for No Further Action (NFA). Alternatively, the magenta curve on the far right represents a set of conditions where there is a 100 percent probability that the risk is greater than 10-4, the upper bound on allowable EPA risks, given the assumptions and data used in the analysis. No additional amount of data collection could show that this site is safe, if a reasonable set of data and assumptions were employed. Therefore, this site should be considered for cleanup, possibly on a fast track. The green curve represents a case where there is significant probability that the site is above the point of departure risk, 10-6, but still less than the upper bound, 10-4. One would probably negotiate with the regulators for an NFA option. The real challenge to decision making comes with the red curve, wherein there is a significant probability that the risk may be greater than 10-4. However, it may be possible to collect additional characterization data to refine the input distributions and data to reduce uncertainties. The curve might then display less than the current two orders of magnitude range in possible risk results. The problem is that predicting whether the curve compresses to the left or the right side of its current range when reducing uncertainty is difficult to predict apriori. The decision maker needs to balance the costs and possible benefits of spending additional time and money on characterization versus committing to a cleanup action.
Another benefit is realized from this type of probabilistic analysis relative to prioritization of waste site investigations. If a probabilistic analysis is performed on all sites early in the investigation phases at a facility one could determine the likelihood of whether sites will pursue an NFA proposal, need detailed characterization, or jump right to cleanup. This has the potential of saving millions of dollars in unnecessary investigations compared to blindly following standard processes in compliance with RCRA or CERCLA stages.
DecisionFX plans to release the RiskFX package for commercial sale at the end of 1999. Pricing will be reasonable for an environmental analysis package (probably about $500). We hope to offer the package free-of-charge to environmental regulators, in the hope that they will evaluate RiskFX for appropriateness of use. DecisionFX would appreciate an email inquiry from prospective buyers of this product to help in marketing, packaging, training, and support considerations. If you express interest, we will keep you apprised of the pending release and planned training activities. email address: info@decisionfx.com
The DOE version of this risk analysis package, PRÉCIS, has been used for site analyses at Sandia National Laboratories. It has also been used for prioritization of the 100+ waste sites that Sandia has responsibility for under RCRA Corrective Action. Some of these applications have been summarized as case studies in this web site. Click on Case Studies to view these applications.
Return to Software Products page