Extrapolation in Risk Assessment: Improving the Quantification of Uncertainty,and Improving Information to Reduce the Uncertainty

Risk assessments inevitably extrapolate from the known to the unknown. The resulting calculation of risk involves two fundamental kinds of uncertainty: uncertainty owing to intrinsically unpredictable (random) components of the future events, and uncertainty owing to imperfect prediction formulas (parameter uncertainty and error in model structure) that are used to predict the component that we think is predictable. Both types of uncertainty weigh heavily both in health and ecological risk assessments. Our first responsibility in conducting risk assessments is to ensure that the reported risks correctly reflect our actual level of uncertainty (of both types). The statistical methods that lend themselves to correct quantification of the uncertainty are also effective for combining different sources of information. One way to reduce uncertainty is to use all the available data. To further sharpen future risk assessments, it is useful to partition the uncertainty between the random component and the component due to parameter uncertainty, so that we can quantify the expected reduction in uncertainty that can be achieved by investing in a given amount of future data. An example is developed to illustrate the potential for use of comparative data, from toxicity testing on other species or other chemicals, to improve the estimates of low-effect concentration in a particular case with sparse case-specific data.     

Axes of Extrapolation in Risk Assessment

Extrapolation in risk assessment involves the use of data and information to estimate or predict something that has not been measured or observed. Reasons for extrapolation include that the number of combinations of environmental stressors and possible receptors is too large to characterize risks comprehensively, that direct characterization is sometimes impossible, and that the power to characterize risk in a particular situation can be enhanced by using information obtained in other similar situations. Three types of extrapolation are common in risk assessments: biological (including between taxa and across levels of biological organization), temporal, and spatial. They can be thought of conceptually as the axes of a 3-dimensional graph defining the state space of biological, temporal, and spatial scales within which extrapolations are made. Each of these types of extrapolation can introduce uncertainties into risk assessments. Such uncertainties may be reduced through synergistic research facilitated by the sharing of methods, models, and data used by human health and ecological scientists     

A Probabilistic Analysis of Regional Mercury Impacts on Wildlife

We investigate the uncertainties associated with modeling the potential health effects on piscivorous animals of mercury released to the atmosphere. The multimedia modeling system combines an atmospheric fate and transport model, an aquatic cycling model, and a terrestrial food web model. First, the modeling system is used to calculate point values of the animals' hazard quotients (i.e., measures of toxic dose). Next, we use a simplified version of the modeling system to conduct a probabilistic analysis for the Great Lakes region that takes into account input uncertainty, variability, and uncertainty and variability combined. The use of two different software packages for the combined uncertainty/variability analysis led to similar results except for high values (>90th percentile) where some differences were evident. A sensitivity study was performed on the combined uncertainty and variability analysis. Regional variability caused more than 70% of the variance in the results, with the fish bioaccumulation factor accounting for the majority of the variability. The major sources of uncertainty were the speciation of the mercury emissions, the lake pH, and the sediment burial rate.     

Extrapolation in Human Health and Ecological Risk Assessments: Proceedings of a Symposium

A symposium was conducted in April 1998 by the U.S. Environmental Protection Agency's National Health and Environmental Effects Research Laboratory (NHEERL) to explore issues of extrapolation in human health and ecological risk assessments. Over the course of three and one half days, leading health and ecology experts presented and discussed research methods and approaches for extrapolating data among taxa and across levels of biological organization, through time, and across spatial scales. The intended result of this symposium was enhanced interaction among a diverse array of scientists, policymakers, and risk assessors to promote identification of approaches for reducing the uncertainties of extrapolation in risk assessment.     

Extrapolation from Individual-Level Responses to Population Growth Rate Using Population Modeling

Ecological risk assessments of chemicals are often based on simple measurements of toxicity in individuals. However, the protection goals are often set at the population and community levels. Population models may be a useful tool to extrapolate from individual-level measurements to population-level endpoints. In the present study, the population growth rate (λ) was calculated for three sets of full life-cycle data (Tetranychus urticae exposed to agrimek, and Daphnia pulex exposed to spinosad and diazinon). The results were compared to λ from population models, where survival and/or reproduction were adjusted according to 4 d of data from the same life-cycle data. This was done to determine whether truncated demographic data can give results similar to that obtained with full life-cycle data. The resulting correlations were strong when both effects on survival and reproduction were included in the model (p < .001, 0.93 < R₂ < 1.00). There were also strong correlations in several cases when only effects on survival or reproduction were considered, although the total risk to the population tended to be underestimated. The results of the present study show that population models can be useful to extrapolate truncated data on the individual level to more ecologically relevant population-level endpoints.     

Uncertainty Is Part of Making Decisions

Advances in computer technology and applied statistics have provided the opportunity for the non-statistician to investigate uncertainty in a quantitative manner. The following discussion argues, notwithstanding the possible misuse of uncertainty analysis, that uncertainty is always present and that decisions based on human or ecological risk assessment would benefit from disclosure of uncertainty in the estimated risks.     

Intra-Species Extrapolation in Risk Assessment of Different Classes of Antimicrobials

The risk assessment process for non-carcinogens incorporates all available scientific information, including toxicokinetic and toxicodynamic data. A 10-fold uncertainty factor (UF) is most commonly used to account for underlying variability within the human species. The purposes of this investigation are to evaluate whether the magnitude of the 10X-UF can be reduced when pharmacokinetic and pharma-codynamic data are incorporated to characterize interindividual variability and whether another UF is needed for the children group. An extensive literature search was conducted on seven antimicrobials in order to incorporate information on kinetics and dynamics to allow extrapolation among susceptible humans. The drugs are cefaclor, cefuroxime, erythromycin, clarithromycin, ampicillin, gentamicin and amikacin. The composite factor was calculated using the highest ratio for appropriate parameters and default subfactor. According to the data, we concluded that when relevant kinetic and dynamic data are available, replacing the default factors with actual data-derived values was possible for the antimicrobials evaluated and that there is no need to add another UF to the children group.     

Adequacy Conditions for Reporting Uncertainty in Chemical Risk Assessments

Uncertainty may influence decision-making. A prerequisite for a decision to be well founded is thus that scientific experts inform decision-makers about all decision relevant uncertainty. A set of conditions is provided for adequate characterization of scientific uncertainty for the purposes of regulatory decision-making. These conditions require specification of (1) the character and degree of uncertainty about the assessment variables, (2) the possibility of reducing the uncertainty, and (3) the degree of agreement among experts. Furthermore, it is required that (4) the information covered by the previous conditions is presented in a clear and comprehensible way. The point of departure is that characterizing scientific uncertainty conceptually means specifying all potentially important possibilities that are consistent with the state of scientific knowledge. The conditions are intended to be applied to human health risk assessment of chemicals. However, the basic approach, to consider potentially important possibilities, should be useful also to environmental, and site-specific risk assessment.     

Treatments of Uncertainty and Variability in Ecological Risk Assessment of Single-Species Populations

The selection of the most appropriate model for an ecological risk assessment depends on the application, the data and resources available, the knowledge base of the assessor, the relevant endpoints, and the extent to which the model deals with uncertainty. Since ecological systems are highly variable and our knowledge of model input parameters is uncertain, it is important that models include treatments of uncertainty and variability, and that results are reported in this light. In this paper we discuss treatments of variation and uncertainty in a variety of population models. In ecological risk assessments, the risk relates to the probability of an adverse event in the context of environmental variation. Uncertainty relates to ignorance about parameter values, e.g., measurement error and systematic error. An assessment of the full distribution of risks, under variability and parameter uncertainty, will give the most comprehensive and flexible endpoint. In this paper we present the rationale behind probabilistic risk assessment, identify the sources of uncertainty relevant for risk assessment and provide an overview of a range of population models. While all of the models reviewed have some utility in ecology, some have more comprehensive treatments of uncertainty than others. We identify the models that allow probabilistic assessments and sensitivity analyses, and we offer recommendations for further developments that aim towards more comprehensive and reliable ecological risk assessments for populations.     

Uncertainty Analysis of Pesticide Residues in Drinking Water Risk Assessment

Pesticide residues in drinking water can vary significantly from day to day. However, water quality monitoring performed under the Safe Drinking Water Act (SDWA) at most community water systems (CWSs) is typically limited to four data points per year over a few years. Due to this limited sampling, likely maximum residues may be underestimated in risk assessment. In this work, a statistical methodology is proposed to study two types of uncertainties in observed samples and their propagated effect in risk estimates. The methodology was demonstrated using data from 16 CWSs that have three independent databases of atrazine residue to estimate the uncertainty of risk in infants and children. The results showed that in 85% of the CWSs, chronic risks predicted with the proposed approach may be two- to four-folds higher than that predicted with the current approach, wheras intermediate risks may be two- to three-folds higher in 50% of the CWSs. In 12% of the CWSs, however, the proposed methodology showed a lower intermediate risk. A closed-form solution of propagated uncertainty was developed to demonstrate the number of years (seasons) of data and sampling frequency needed to reduce the uncertainty of risk estimates. In general, this methodology provided good insight into the importance of addressing uncertainty of observed water quality data and the need to predict likely maximum residues in risk assessment.     

Test systems to determine the ecological risks posed by toxin release from Bacillus thuringiensis genes in crop plants

Procedures for the selection of species for ecotoxicological risk assessment of Bacillus thuringiensis (Bt) gene products in the epigeal and hypogeal environments are proposed. Although species can be selected on the basis of ecological realism and functional importance, the number of organisms requiring testing and the nature of the test procedures remain uncertain with such a selectively toxic material. The heterogeneity of the soil environment, the stratification of plant material at different stages of breakdown and decomposition and the aggregation and patterns of movement of the soil fauna and flora impose problems for the design of ecologically relevant test methods. Similarly, the impact upon beneficial invertebrates, if toxic effects are detected, will be mediated by the scale and pattern of transgenic plant release in the fragmented agricultural landscape. To properly assess the ecological risks posed by a widely released toxin with a narrow spectrum of effects, a combination of laboratory tests, field experiments and longer-term monitoring will be required.     

Harmonizing the Incorporation of Uncertainty and Variability Into Risk Assessment: An International Approach

International harmonization of risk assessment approaches affords a number of opportunities and advantages. Overall, harmonization will lead to more efficient use of resources, but also will lead to better understanding amongst scientists and regulators worldwide. It is with these goals in mind that in 1994 the International Programme on Chemical Safety (IPCS) initiated its Project on the Harmonization of Approaches to the Assessment of Risk from Exposure to Chemicals (Harmonization Project). An ongoing activity under this project addresses uncertainty and variability in risk assessment. The goal of the overall activity is to promote harmonization of risk assessment methodologies for noncancer endpoints. However, given the common links in uncertainty and variability that apply across a range of end-point-specific activities, these links are identified wherever possible. This paper provides an overview of the IPCS Harmonization Project and reviews the activity and future plans related to uncertainty and variability.     

An ecological risk assessment paradigm using the Spatially Integrated model for Phosphorus Loading and Erosion (SIMPLE)

Ecological risk assessments provide a probabilitistic approach to analyzing and predicting ecosystem responses to stress. We are evaluating the relationship between nonpoint source (NPS) phosphorus loading and the trophic status of the aquatic ecosystem. We are using SIMPLE (the Spatially Integrated Model for Phosphorus Loading and Erosion) to identify probable phosphorus sources in a watershed, simulate the phosphorus loading to streams, and analyze the relationships between input variables and their ecological impact. The objective of this paper is to describe a risk-based paradigm using SIMPLE to characterize the probability of exceeding a critical phosphorus loading to a lotic ecosystem. We have characterized the risk of exceeding a threshold loading of 0.5 kilogram total phosphorus per hectare per year from a 2238 hectare watershed. Two-hundred-fifty random SIMPLE simulations were performed to estimate annual total phosphorus, dissolved phosphorus, and sediment-bound phosphorus loading to a lotic ecosystem from the watershed. Simulation results were analyzed statistically to determine the probabilities of exceeding the critical loadings. Based on the current land use practices in the Battle Creek watershed, the probability of exceeding the total phosphorus critical loading rate of 0.5 kg/ha/yr was approximately 11 percent, or one year in nine the total annual loading will exceed the critical loading rate. The 95 percent confidence intervals for the total phosphorus loading occurring on average once in nine years were relatively close (0.45 to 0.60 kg/ha/yr), assuming the only variability from year to year was due to natural variability in weather.     

Quantifying Uncertainty in Predictions of Invasiveness

Using the Australian weed risk assessment (WRA) model as an example, we applied a combination of bootstrapping and Bayesian techniques as a means for explicitly estimating the posterior probability of weediness as a function of an import risk assessment model screening score. Our approach provides estimates of uncertainty around model predictions, after correcting for verification bias arising from the original training dataset having a higher proportion of weed species than would be the norm, and incorporates uncertainty in current knowledge of the prior (base-rate) probability of weediness. The results confirm the high sensitivity of the posterior probability of weediness to the base-rate probability of weediness of plants proposed for importation, and demonstrate how uncertainty in this base-rate probability manifests itself in uncertainty surrounding predicted probabilities of weediness. This quantitative estimate of the weediness probability posed by taxa classified using the WRA model, including estimates of uncertainty around this probability for a given WRA score, would enable bio-economic modelling to contribute to the decision process, should this avenue be pursued. Regardless of whether or not this avenue is explored, the explicit estimates of uncertainty around weed classifications will enable managers to make better informed decisions regarding risk. When viewed in terms of likelihood of weed introduction, the current WRA model outcomes of ‘accept’, ‘further evaluate’, or ‘reject’, whilst not always accurate in terms of weed classification, appear consistent with a high expected cost of mistakenly introducing a weed. The methods presented have wider application to the quantitative prediction of invasive species for situations where the base-rate probability of invasiveness is subject to uncertainty, and the accuracy of the screening test imperfect     

Approaches to Ecological Risk Characterization and Management: Selecting the Right Tools for the Job

Chemical-specific hazard quotient (HQ) risk characterization in ecological risk assessment (ERA) can be a value-added tool for risk management decision-making at chemical release sites, when applied appropriately. However, there is little consensus regarding how HQ results can be used for risk management decision-making at the population, community, and ecosystem levels. Furthermore, stakeholders are reluctant to consider alternatives to HQ results for risk management decisions. Chemical-specific HQs risk characterization should be viewed as only one of several approaches (i.e., tools) for addressing ecological issues; and in many situations, other quantitative and qualitative approaches will likely result in superior risk management decisions. The purpose of this paper is to address fundamental issues and limitations associated with chemical-specific HQ risk characterization in ERA, to identify when it may be appropriate, to explore alternatives that are currently available, and to identify areas that could be developed for the future. Several alternatives (i.e., compensatory restoration, performance-based ecological monitoring, ecological significance criteria, net environmental benefit analysis), including their limitations, that can supplement, augment, or substitute for HQs in ERA are presented. In addition, areas of research (i.e., wildlife habitat assessment/landscape ecology/population biology, and field validated risk-based screening levels) that could yield new tools are discussed.     

Chemical Contamination of Aquatic Organisms from an Urbanized River in the New York/New Jersey Harbor Estuary

The lower 6 miles of the tidal Passaic River, part of the New York/New Jersey (NY/NJ) Harbor Estuary system, are contaminated with a variety of organic and inorganic chemicals as a result of more than 150 years of heavy industrialization and urbanization. The River's ecology is substantially degraded due to habitat removal/alteration, and the organisms that reside in or utilize the River are exposed to and bioaccumulate chemicals from sediments and food web interactions. We quantify in this study the extent and magnitude of chemical contamination in several fish species (representing a range of trophic levels) and blue crab (Callinectes sapidus). In addition, the concentration of several contaminants of concern are compared to concentrations in similar organisms from other areas of the NY/NJ Harbor Estuary, as well as available tissue-based toxicological effects benchmarks that are reported in the literature. The results suggest that a variety of contaminants are present at elevated levels in each of the species collected from the River. Several contaminants, including 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD), total DDT (2,2-bis[4-chlorophenyl]1,1-dichloroethene), copper, and mercury are present at average concentrations that exceed those from other waterways in the NY/NJ Harbor Estuary. However, the concentrations of contaminants in the River, with few exceptions do not exceed available toxic effects levels as reported in the literature for these or similar fish and crustaceans. This suggests that toxicological risks from bioaccumulative contaminants in the lower Passaic River are limited to select contaminants and species.     

Analysis and Portrayal of Uncertainty in a Food-Web Exposure Model

The results of quantitative risk assessments are key factors in a risk manager's decision of the necessity to implement actions to reduce risk. The extent of the uncertainty in the assessment will play a large part in the degree of confidence a risk manager has in the reported significance and probability of a given risk. The two main sources of uncertainty in such risk assessments are variability and incertitude. In this paper we use two methods, a second-order two-dimensional Monte Carlo analysis and probability bounds analysis, to investigate the impact of both types of uncertainty on the results of a food-web exposure model. We demonstrate how the full extent of uncertainty in a risk estimate can be fully portrayed in a way that is useful to risk managers. We show that probability bounds analysis is a useful tool for identifying the parameters that contribute the most to uncertainty in a risk estimate and how it can be used to complement established practices in risk assessment. We conclude by promoting the use of probability analysis in conjunction with Monte Carlo analyses as a method for checking how plausible Monte Carlo results are in the full context of uncertainty.