2
PRIORITY-SETTINGPROCESSES
BASIS FOR DEVELOPING A PRIORITY-SETTING APPROACH
The United States faces the challenge of environmental restoration of thousands of contaminated hazardous-waste sites across the nation. Although it is difficult to estimate accurately the total resources required to address this challenge, current projections in terms of the dollars and person-years needed are enormous (see Chapter 1). Given the situation of resources limited by natural catastrophes, federal budget deficits, and other demands, it has become critical that scientifically credible estimates be developed to help choose sites for remediation, determine the extent to which each should be remediated, and set the priority in which remediations occur.
A system to help set priorities for the restoration of hazardous-waste sites could benefit greatly a
wide range of individuals and groups involved in environmental restoration, waste management, and public health. The primary goal of such a system is to provide a formal, systematic, consistent, and scientifically based framework to catalog and compare factors to assist in decision making and resource allocation. There are many factors that can affect a priority-setting outcome. These include health, safety, and ecological risks; social and economic values and policies; regulatory requirements; technical considerations, and variation of all these over time. A properly designed priority-setting system would aid decision-makers in (1) designing strategies for minimizing human health and ecological damage; (2) enhancing the sound use of natural resources; (3) promoting the efficient allocation of remediation resources; (4) increasing the efficiency of administrative processes associated with the restoration programs; and (5) strengthening the credibility and acceptance of the priority-setting process.
Most of the discussions in this report focus on site-ranking processes rather than on priority-setting processes. That distinction is an important one, because most analytic systems developed to date are used only to rank sites according to some numerical score (Halfon and Reggiani, 1986; Halfon, 1989). The ranked scores are then considered along with other factors to arrive at actual remediation priorities (see Chapters 4-7 for specific examples). The site-remediation priorities are, therefore, subject to being different from the numerical rankings. This chapter discusses some of the desirable features of a priority-setting system including the analytic models used in such a system. This chapter also discusses some of the evaluation criteria applied to methods discussed in later chapters.
Although a priority-setting process should focus on individual sites and the feasibility of remediating such sites, the incorporation of such evaluations into a nationwide scoring system and subsequent budget analysis requires careful consideration. For exam-
ple, even though a ranking based upon the reduction of human-health risk can be used as the basis for national priority setting, the inclusion of issues that pertain to societal impacts might not always have a common denominator nationwide. Different communities or states might place different values on such factors as the loss of wildlife, diminished air quality, or the decline in local real estate values. Thus, the locally affected communities must be involved in the evaluation of sites being considered under any ranking or priority-setting system. An appropriate format for soliciting and explicitly incorporating public input into the priority-setting process is essential.
DESIRABLE FEATURES OF A PRIORITY-SETTING SYSTEM
Overview
A priority-setting system should be designed with an a priori knowledge of the purpose and process by which it can affect decision-making. The system should consider the possible solutions in the evaluation process rather than just the severity of environmental impact. For example, a site restoration might have a simple solution, such as the removal of a small amount of contained waste (e.g., in barrels). That might not require a large allocation of resources, and therefore, could be completed in a relatively short time.
Numerous, and often competing, objectives enter into environmental restoration and decision-making. These include the direct and indirect impacts of the hazardous-waste sites on human health and the environment, as well as social and economic effects, at the local level. However, ramifications at the national level must also be considered, especially from the viewpoints of the economic and
political impacts of site remediation. The various effects of hazardous waste sites can be categorized as presented in Table 2-1, following the classification of Greenberg and Anderson (1984). Clearly, priority setting is rooted in a multi-objective decision-making process, and thus, it is necessary to have appropriate measures developed for each of the relevant objectives. Although it is tempting to provide a comparative assessment of contaminated sites based on an overall single score that encompasses all factors, such an approach might be unrealistic. A priority-setting method might have to be designed that provides a range of scores that might not be necessarily additive, but might be sufficiently informative to present decision-makers with a clearer view of the problem that they are facing.
Finally, a priority-setting system must have scientific credibility. That is, such a system must be objective and replicable, so as to strengthen its acceptability and effectiveness. The credibility of the system also depends on the accuracy of the data that are available for the sites being considered in the priority-setting process. Uniform requirements for technical data and cost estimates must be established, which will ensure that all sites are evaluated and compared on a consistent basis. At the same time, given that priority setting might be required at many sites in the early stages of investigation—when detailed site information is lacking—the priority-setting process must be flexible enough to handle information at different levels of detail and accuracy, along with the associated uncertainties. The process should (1) allow isolation of those areas of uncertainty that affect site scoring and (2) suggest what additional data should be acquired. That is, readers should be able to tell which information used in the process is of high quality and which information is not. The process should have a mechanism for updating the rankings or priorities as more information becomes available.
TABLE 2-1 Potential Impacts of Hazardous-Waste Sites
Health |
Environment |
Social |
Economic |
Immediate death |
Ecosystem elimination |
Disruption of existing communities |
Severe damage of human-made structures |
Life-shortening exposure |
Elimination of species |
Disruption of a few families |
Extreme devaluation of property |
Acute Illness |
Reduction in abundance within species |
Disruption of a single household |
Reduced appreciation of property values |
Severe disability |
Reduction in biomass productivity |
|
Loss of productivity of the land |
Chronic illness |
Loss of use of resource |
|
Local taxpayers pay for cleanup, security, and other site maintenance |
Chronic disability |
Reduction in use of resource |
|
|
Minor or temporary illness |
|
|
|
Emotional illness |
|
|
|
Source: Adapted from Greenberg and Anderson, 1984. |
General Issues in Model Development and Application
Because ranking and priority-setting models are designed to be influential components of various environmental restoration programs, it is essential that such models meet high standards. Professionally accepted protocols for proper model development and application should be followed (Gass and Thompson, 1980; Gass and Joel, 1981; Gass, 1983; GAO, 1987).
The following issues concerning model development and application should be considered:
-
Purpose: A clearly defined and explicitly stated purpose for the model including a defined user population;
-
Credibility and acceptability: The model's development must include scientific peer review, public participation, and public comment;
-
Appropriate logic and implementation of the model's mathematics: The equations for evaluating and combining factors must be consistent, scientifically valid, and well chosen for numerical execution;
-
Documentation of the model's development: Documentation must be provided not only on how to use the model, but also on how the model was developed, i.e., why the model components were chosen over other plausible alternatives;
-
Validation of the model: The model must have been shown to produce a ranking of site risks or threats reliable enough to fulfill the purpose for which it was designed; and
-
Appropriate sensitivity and uncertainty analyses: Evaluations must be performed to determine the uncertainties in model scores and the resulting implications for site ranking and setting priorities; appropriate quality control and quality assurance procedures must be incorporated and emphasize quality for input data to which model scores are most sensitive.
Specific Technical Features of a Hazardous-Waste Site-Ranking and Priority-Setting Model
In addition to the general issues and features of the ranking or priority-setting process discussed above, the following technical requirements should be addressed during the development of a computation model used in the process:
-
Model should formally incorporate information regarding uncertainty into its various algorithms when input parameters are unavailable and therefore must be estimated, or when there is lack of confidence in the data. It is also important that the effects of uncertainties on the final ranking process are clearly identified and reported in a format that is usable by decision-makers.
-
Model should be applicable to all hazardous-waste sites. The process should be sufficiently flexible to handle all types of hazardous waste sites including, but not limited to, landfills, surface waters and sediments, and contaminated groundwater plumes.
-
Model should allow for dynamic tracking and updating of information. As new data are obtained about the site and its potential impacts on the surrounding communities and environment, such information should be incorporated into the model; the model should be able to accommodate new information, keep track of the change in priority or rank, and provide a quantitative comparison with prior rankings of the site.
-
Model should discriminate between immediate- and long-term risk to human health and environment. Long-term risk refers to the potential for harmful effects that might take more than 20 years to be manifest or to site contaminants that pose a risk for long periods (e.g., centuries). A special algorithm may be needed for indicating risk beyond several generations.
-
Model should include cost estimates of remediation alternatives. These should include considerations of timing related to immediate remediation versus delay in remediation.
-
Model structure should be "transparent." The various components of the ranking or priority-setting modal should be clear, logical, and thoroughly explained; despite possible complexity of the model, the scoring algorithms should be clearly documented and articulated for easy understanding by the users. The model output also should be transparent in terms of providing an overall score and additional information that would allow a person to readily determine why a score is high, and which contaminants, environmental pathways, and receptor populations, etc., are of concern. This additional information is important because different sites could receive high scores for very different reasons.
-
Model should be user-friendly. The model should be constructed so that nonscientists and individuals who are not computer experts can operate the model; it should be constructed as an interactive system that allows detailed system interrogation and maximum flexibility in generating various scenarios; it should have sufficient on-line help to guide the user through the process of data input and analysis.
-
Model should include appropriate security features to prevent unauthorized changes in site data, model parameters, and model outputs. Only the system designers and maintenance group should be allowed to make changes to the data bases inherent in the model (e.g., data bases of physicochemical properties or unit risk factors); the system should be protected against tampering and the input of meaningless data (e.g., negative concentration values or values outside certain defined upper and lower limits).