Deterministic Risk Assessment Research Articles

Uncertain Numbers and Uncertainty in the Selection of Input Distributions—Consequences for a Probabilistic Risk Assessment of Contaminated Land

Risks from exposure to contaminated land are often assessed with the aid of mathematical models. The current probabilistic approach is a considerable improvement on previous deterministic risk assessment practices, in that it attempts to characterize uncertainty and variability. However, some inputs continue to be assigned as precise numbers, while others are characterized as precise probability distributions. Such precision is hard to justify, and we show in this article how rounding errors and distribution assumptions can affect an exposure assessment. The outcome of traditional deterministic point estimates and Monte Carlo simulations were compared to probability bounds analyses. Assigning all scalars as imprecise numbers (intervals prescribed by significant digits) added uncertainty to the deterministic point estimate of about one order of magnitude. Similarly, representing probability distributions as probability boxes added several orders of magnitude to the uncertainty of the probabilistic estimate. This indicates that the size of the uncertainty in such assessments is actually much greater than currently reported. The article suggests that full disclosure of the uncertainty may facilitate decision making in opening up a negotiation window. In the risk analysis process, it is also an ethical obligation to clarify the boundary between the scientific and social domains.

Field evaluation of an avian risk assessment model

We conducted two laboratory subacute dietary toxicity tests and one outdoor subacute dietary toxicity test to determine the effectiveness of the U.S. Environmental Protection Agency's deterministic risk assessment model for evaluating the potential of adverse effects to birds in the field. We tested technical-grade diazinon and its D.Z.N 50W (50% diazinon active ingredient wettable powder) formulation on Canada goose (Branta canadensis) goslings. Brain acetylcholinesterase activity was measured, and the feathers and skin, feet, and gastrointestinal contents were analyzed for diazinon residues. The dose-response curves showed that diazinon was significantly more toxic to goslings in the outdoor test than in the laboratory tests. The deterministic risk assessment method identified the potential for risk to birds in general, but the factors associated with extrapolating from the laboratory to the field, and from the laboratory test species to other species, resulted in the underestimation of risk to the goslings. The present study indicates that laboratory-based risk quotients should be interpreted with caution.

PCB contamination from polysulphide sealants in residential areas—exposure and risk assessment

From the late 1950s to the early 1970s elastic polysulphide sealants were used in outdoor seams between concrete blocks in prefabricated buildings. The sealants contained 5–30% polychlorinated biphenyls (PCBs). Due to the weathering of sealants in general and the replacement of seams with new PCB-free materials in the 1990s, PCBs have drifted to the soil adjacent to buildings. The objectives of this study were to survey PCB contamination in the surroundings of former PCB-containing buildings and to evaluate the risks to human health. Samples from soil, and also from blood serum of residents, were collected to obtain data for exposure assessment. The health risk assessment was based on deterministic and probabilistic calculations for cancer and non-cancer risks. Soil ingestion and dermal contact were considered the main routes of exposure and children the most important exposed group. The mean total PCB concentration was 6.83 mg/kg within 2 m of the buildings and 0.52 mg/kg within 3–10 m from of the buildings. The deterministic risk assessment with conservative parameters resulted in lifetime cancer risk estimates on the order of 10 −6–10 −7. The lifetime average daily dose (LADD) for PCBs was less than 10% of the reference dose (RfD) 0.02 μg/kg day, which is based on immunosupression in monkeys. The LADD corresponding to the total site attributable exposure was less than 10% of the estimated average dietary PCB intake in Finland. Children can, however, in worst cases be exposed to daily doses near the level of the RfD. Low cost measures are recommended to reduce possible exposure of children.

“Public Health Consequences of Macrolide Use in Food Animals: A Deterministic Risk Assessment,” A Comment on:J. Food Prot. 67(5):980–992 (2004)

macrolide use in food animals: a deterministic risk assessment. J. Food Prot. 67:980–992. 4. McDermott, P. F., S. M. Bodeis, and L. L. English. 2002. Ciprofloxacin resistance in Campylobacter jejuni evolves rapidly in chickens treated with fluoroquinolones. J. Infect. Dis. 185:837–840. 5. U.S. Food and Drug Administration, Center for Veterinary Medicine. 2001. Risk assessment of the human health impact of fluoroquinolone resistant Campylobacter associated with the consumption of chicken. Available at: http://www.fda.gov/cvm/antimicrobial/Riskpasses.htm. Accessed 17 June 2004. 6. U.S. Food and Drug Administration, Center for Veterinary Medicine. 2003. Guidance for industry: evaluating the safety of antimicrobial new animal drugs with regard to their microbiological effects on bacteria of human health concern. Document no. 152. Available at: http:// www.fda.gov/cvm/guidance/fguide152.pdf. Accessed 17 June 2004. 7. U.S. Food and Drug Administration, Center for Veterinary Medicine. 2004. Initial decision on proposed withdrawal of Baytril poultry NADA. Available at: http://www.fda.gov/cvm/antimicrobial/FQWithdrawal. html. Accessed 17 June 2004. 8. World Health Organization (WHO). 2003. Impacts of antimicrobial growth promoter termination in Denmark. The WHO international review panel’s evaluation of the termination of the use of antimicrobial growth promoters in Denmark. Ref. no. WHO/CDS/CPE/ZFK/ 2003.1. WHO, Geneva. 9. Zhao, C., B. Ge, J. De Villena, R. Sudler, E. Yeh, S. Zhao, D. G. White, D. Wagner, and J. Meng. 2001. Prevalence of Campylobacter spp., Escherichia coli, and Salmonella serovars in retail chicken, turkey, pork, and beef from the greater Washington, D.C., area. Appl. Environ. Microbiol. 67:5431–5436.

Risk characterisation in direct toxicity assessment of the River Esk and the Tees Estuary.

This paper illustrates the use of deterministic and probabilistic approaches to the ecotoxicological risk assessment of effluents through two simple examples taken from the UK Direct Toxicity Assessment Demonstration Programme. In this study, the direct protection objective for the River Esk in Scotland was the prevention of short-term lethal toxicity to invertebrates from exposure to effluent containing insecticides. Both a deterministic and a probabilistic risk characterisation show that such toxicity is very unlikely to occur. The protection objectives for the Lower Tees Estuary were more complex, due to a variety of discharges and potentially sensitive species. A deterministic risk assessment using the most sensitive species suggests that toxicity will be high at all times. In contrast, a probabilistic risk assessment suggests that, although acute toxicity is considerable, it is not likely to occur at all times or affect all species, and discharge scenarios do exist in which acute toxicity could be relatively low. These two case studies illustrate the use of probabilistic risk assessment in achieving more realistic assessments of effluent toxicity through use of all available data.

Australia’s CO 2 geological storage potential and matching of emission sources to potential sinks

Within the GEODISC program of the Australian Petroleum Cooperative Research Centre (APCRC), Geoscience Australia (GA) and the University of New South Wales (UNSW) have completed an analysis of the potential for the geological storage of CO 2. The geological analysis assessed over 100 potential environmentally sustainable sites for CO 2 injection (ESSCIs) by applying a deterministic risk assessment based on the five factors of: storage capacity, injectivity potential, site details, containment and natural resources. Utilising a risked storage capacity suggests that at a regional scale Australia has a CO 2 storage potential in excess of 1600 years of current annual total net emissions. Whilst this estimate does give an idea of the enormous magnitude of the geological storage potential of CO 2 in Australia, it does not account for various factors that are evident in source to sink matching. If preferences due to source to sink matching are incorporated, and an assumption is made that some economic imperative will apply to encourage geological storage of CO 2, then a more realistic analysis can be derived. In such a case, Australia may have the potential to store a maximum of 25% of our total annual net emissions, or approximately 100–115 Mt CO 2 per year.

Public Health Consequences of Macrolide Use in Food Animals: A Deterministic Risk Assessment

The potential impact on human health from antibiotic-resistant bacteria selected by use of antibiotics in food animals has resulted in many reports and recommended actions. The U.S. Food and Drug Administration Center for Veterinary Medicine has issued Guidance Document 152, which advises veterinary drug sponsors of one potential process for conducting a qualitative risk assessment of drug use in food animals. Using this guideline, we developed a deterministic model to assess the risk from two macrolide antibiotics, tylosin and tilmicosin. The scope of modeling included all label claim uses of both macrolides in poultry, swine, and beef cattle. The Guidance Document was followed to define the hazard, which is illness (i) caused by foodborne bacteria with a resistance determinant, (ii) attributed to a specified animal-derived meat commodity, and (iii) treated with a human use drug of the same class. Risk was defined as the probability of this hazard combined with the consequence of treatment failure due to resistant Campylobacter spp. or Enterococcus faecium. A binomial event model was applied to estimate the annual risk for the U.S. general population. Parameters were derived from industry drug use surveys, scientific literature, medical guidelines, and government documents. This unique farm-to-patient risk assessment demonstrated that use of tylosin and tilmicosin in food animals presents a very low risk of human treatment failure, with an approximate annual probability of less than 1 in 10 million Campylobacter-derived and approximately 1 in 3 billion E. faecium–derived risk.

Australia?s CO2 geological storage potential and matching of emission sources to potential sinks

Within the GEODISC program of the Australian Petroleum Cooperative Research Centre (APCRC), Geoscience Australia (GA) and the University of New South Wales (UNSW) have completed an analysis of the potential for the geological storage of CO2. The geological analysis assessed over 100 potential environmentally sustainable sites for CO2 injection (ESSCIs) by applying a deterministic risk assessment based on the five factors of: storage capacity, injectivity potential, site details, containment and natural resources. Utilising a risked storage capacity suggests that at a regional scale Australia has a CO2 storage potential in excess of 1600 years of current annual total net emissions. Whilst this estimate does give an idea of the enormous magnitude of the geological storage potential of CO2 in Australia, it does not account for various factors that are evident in source to sink matching. If preferences due to source to sink matching are incorporated, and an assumption is made that some economic imperative will apply to encourage geological storage of CO2, then a more realistic analysis can be derived. In such a case, Australia may have the potential to store a maximum of 25% of our total annual net emissions, or approximately 100–115 Mt CO2 per year.

Model Intercomparison for the Uptake of Organic Chemicals by Plants.

Currently, a variety of models are available for predicting the uptake, translocation, and elimination of organic contaminants by plants. These models range from simple deterministic risk assessment screening tools to more complex models that consider physical, chemical, and biological processes in a mechanistic manner. This study evaluates the performance of a range of such models and model types against experimental data sets. Three dynamic, three regression-based, and three steady-state and equilibrium models have been selected for evaluation. These models differ in terms of their scope, methodological approach, and complexity. Data from nine published experiments were used to create scenarios to test model performance. These experiments consider plant contamination via both soil and aerial exposure pathways. A total of 19 different organic chemicals were used in the experiments along with 7 different plant species. Model predictions of chemical concentrations in the relevant plant compartments were compared with the experimentally recorded values. The results indicate that dynamic models offer performance advantages for acute exposure durations and for rapidly changing environmental media. Equilibrium/steady-state and regression-based models perform better for chronic exposure durations, where stable conditions are more likely to exist. The selection of an appropriate plant uptake model will therefore be dependent on the requirements of the assessment, the nature of the environmental media, and the duration of the source term. The results generated by the regression-based models suggest that in their current form these models are unsuitable for evaluating the uptake of organic chemicals from the air into plants.

Probabilistic and deterministic risk assessment for extreme objects and ecologically hazardous systems

Probabilistic and deterministic risk assessment for extreme objects and ecologically hazardous systems

Uncertainty in toxicological risk assessment for non-carcinogenic health effects

Uncertainty in risk assessment results from the lack of knowledge on toxicity to the target population for a substance. Currently used deterministic risk assessment methods yield human limit values or margins of safety (MOS) without quantitative measurements of uncertainty. Qualitative and quantitative uncertainty analysis would enable risk managers to better judge the consequences of different management options. This article discusses sources of uncertainty and possibilities for quantification of uncertainty associated with different steps in the risk assessment of non-carcinogenic health effects. Knowledge gaps causing uncertainty in risk assessment are overcome by extrapolation. Distribution functions for extrapolation factors are based on empirical data and provide information about the extent of uncertainty introduced by these factors. Whereas deterministic methods can account only qualitatively for uncertainty of the resulting human limit value, probabilistic risk assessment methods are able to quantify several aspects of uncertainty. However, there is only limited experience with these methods in practice. Their acceptance and future application will depend on the establishment of evidence based distribution functions, flexibility and practicability of the methods, and the unambiguity of the results.

Integrated Human and Ecological Risk Assessment: A Case Study of Organophosphorous Pesticides in the Environment

This study was chosen as an example of integrated risk assessment because organophosphorous esters (OPs) share exposure characteristics for different species, including human beings and because a common mechanism of action can be identified. The “Framework for the integration of health and ecological risk assessment” is being tested against a deterministic integrated environmental health risk assessment for OPs used in a typical farming community. It is argued that the integrated approach helps both the risk manager and the risk assessor in formulating a more holistic approach toward the risk of the use of OP-esters. It avoids conclusions based on incomplete assessments or on separate assessments. The database available can be expanded and results can be expressed in a more coherent manner. In the integrated exposure assessment of OPs, the risk assessments for human beings and the environment share many communalities with regards to sources and emissions, distribution routes and exposure scenarios. The site of action of OPs, acetylcholinesterase, has been established in a vast array of species, including humans. It follows that in the integrated approach the effects assessment for various species will show communalities in reported effects and standard setting approaches. In the risk characterization, a common set of evidence, common criteria, and common interpretations of those criteria are used to determine the cause of human and ecological effects that co-occur or are apparently associated with exposure to OPs. Results of health and ecological risk assessments are presented in a common format that facilitates comparison of results. It avoids acceptable risk conclusions with regard to the environment, which are unacceptable with regard to human risk and vice versa. Risk managers will be prompted to a more balanced judgement and understanding and acceptance of risk reduction measures will be facilitated.

Application of a Probabilistic Risk Assessment Methodology to a Lead Smelter Site

Exposure of children to lead in the environment was assessed at the Murray Smelter Superfund site using both a deterministic risk assessment approach, the Integrated Exposure Uptake Biokinetic (IEUBK) model, and a probabilistic approach, the Integrated Stochastic Exposure (ISE) model. When site-specific data on lead in environmental media were input as point estimates into the IEUBK model, unacceptable risks were predicted for children living within five of eight study zones. The predicted soil cleanup goal was 550 ppm. Concentration and exposure data were then input into the ISE model as probability distribution functions and a one-dimensional Monte Carlo analysis (ID MCA) was run to predict the expected distribution of exposures and blood lead values. Uncertainty surrounding these predictions was examined in a two-dimensional Monte Carlo analysis (2-D MCA). The ISE model predicted risks that were in the same rank order as those predicted by the IEUBK model, although the probability estimates of exceeding a blood lead level of 10 µg/dl (referred to as the P10) from the ISE model were uniformly lower than those predicted by the IEUBK model. The 2-D MCA allowed evaluation of the confidence around each P10 level, and identified the main sources of both uncertainty and variability in exposure estimates. The ISE model suggested cleanup goals ranging from 1300 to 1500 ppm might be protective at this site.

Application of a Probabilistic Risk Assessment Methodology to a Lead Smelter Site

Exposure of children to lead in the environment was assessed at the Murray Smelter Superfund site using both a deterministic risk assessment approach, the Integrated Exposure Uptake Biokinetic (IEUBK) model, and a probabilistic approach, the Integrated Stochastic Exposure (ISE) model. When site-specific data on lead in environmental media were input as point estimates into the IEUBK model, unacceptable risks were predicted for children living within five of eight study zones. The predicted soil cleanup goal was 550 ppm. Concentration and exposure data were then input into the ISE model as probability distribution functions and a one-dimensional Monte Carlo analysis (ID MCA) was run to predict the expected distribution of exposures and blood lead values. Uncertainty surrounding these predictions was examined in a two-dimensional Monte Carlo analysis (2-D MCA). The ISE model predicted risks that were in the same rank order as those predicted by the IEUBK model, although the probability estimates of exceeding a blood lead level of 10 µg/dl (referred to as the P10) from the ISE model were uniformly lower than those predicted by the IEUBK model. The 2-D MCA allowed evaluation of the confidence around each P10 level, and identified the main sources of both uncertainty and variability in exposure estimates. The ISE model suggested cleanup goals ranging from 1300 to 1500 ppm might be protective at this site.

Seismic reevaluation and upgrading of nuclear power facilities outside the US using US developed methodologies

Several seismic licensing and safety issues have emerged over the past fifteen years for commercial US Nuclear Power Plants and US Government research reactors, production reactors and process facilities. The methodologies for the resolution of these issues have been developed in numerous government and utility sponsored research programs. The resolution criteria have included conservative deterministic design criteria, deterministic seismic margins assessments criteria (SMA) and seismic probabilistic risk assessment criteria (SPRA). The criteria for SMAs and SPRAs have been based realistically on considering the inelastic energy absorption capability of ductile structures, equipment and piping and have incorporated the use of earthquake and testing experience to evaluate the operability of complex mechanical and electrical equipment. Most of the applications to date have been confined to the US, however there have been several applications to Asian, Western and Eastern Europe reactors. This paper summarizes the major issues addressed, the development of reevaluation criteria and selected applications to non US reactors including VVER reactors of Soviet origin.

Comparison of Approaches for Developing Distributions for Carcinogenic Slope Factors

In recent years, risk assessors have increasingly been moving away from deterministic risk assessment approaches and are applying probabilistic approaches that incorporate distributions of possible values for each input parameter. This paper reviews several approaches that are being used or that could potentially be used to develop distributions for carcinogenic slope factors (CSFs). Based on the primary tool or framework that is applied, these approaches have been divided into the following three categories: the statistical framework, the decision analysis framework, and the biological framework. Work that has been done on each approach is summarized, and the aspects of variability and uncertainty that are incorporated into each approach are examined. The implications of the resulting distributional information for calculating risk estimates or risk-based concentrations is explored. The approaches differ in their stage of development, the degree to which they diverge from the U.S. Environmental Protection Agency's (EPA) current practice in establishing CSF values, their flexibility to accommodate varying data sets or theories of carcinogenicity, and their complexity of application. In some cases, wide ranges of potential potency estimates are indicated by these approaches. Such findings suggest widely divergent risk assessment implications and the need for additional evaluation of the goals of developing CSF distributions for use in risk assessment applications and the types of information that should be reflected in such distributions. Some combination of the features offered by these approaches may best support risk assessment and risk management decisions.

The need for new methods to backcalculate soil cleanup targets in interval and probabilistic cancer risk assessments

Abstract Ordinary algebra may be used to backcalculate health‐based cleanup targets in deterministic risk assessments, but it does not work in interval or probabilistic risk assessments. Equations with interval or random variables do not follow the rules of ordinary algebra. This paper explains the need for more sophisticated methods to backcalculate soil cleanup targets when using interval or random variables.

A discussion on the use of probabilistic risk assessment in human health impact assessment

Risk assessment provides a methodology for estimating human exposure and risk from contaminated sites as part of a human health impact assessment (HHIA). All risk assessments follow a common outline and require the use of site-specific data and modeling at each of the steps in the process. Assumptions are made regarding the conditions under which exposure to contamination will take place. Deterministic risk assessment quantifies each of these assumptions as point estimates, whereas probabilistic risk assessment incorporates probability distributions for each variable, providing a context for the point estimates that shows the full range of risks likely to be experienced under specific exposure conditions. Once the risks have been estimated, risk managers communicate this information to decision-makers, site developers, and interested members of the public. Risk assessors must answer the question, “what is the chance that a hypothetical individual will develop cancer over the course of a lifetime as a result of living in the vicinity of a given site?” This article presents an overview of the individual steps in a risk assessment, with an emphasis on the use of probabilistic methods. Risk assessment is distinguished from risk management, and a brief discussion of computer desktop/video presentation tools is provided.