Optimisation of radiological protection

Draft document: Optimisation of radiological protection
Submitted by Andrew Wallo, Office of Air, Water and Radiation Protection Policy and Guidance, U.S. Department of Energy
Commenting on behalf of the organisation

Comments on the 11 April 2005 Draft ”The Optimisation of Radiological Protection” General Comments: We appreciate the opportunity provided by the Commission to comment on the subject report. Unlike prior ICRP recommendation reports (ICRP 2, ICRP 26, and ICRP 60) on radiation protection, the current draft document does not offer substantial changes in dose assessment methodology, terminology, or limits on radiation exposures. This draft recommendation document appears to be much easier to read, follow, and understand than many prior ICRP publications. However, with a few exceptions, the optimization report is not likely to aid in a better understanding or implementation of the optimization process and it may in fact, further complicate implementation. Although those very familiar with optimization (and how it is used in the radiation protection system) will be able to understand and appropriately use the information contained in the document, others may not. The primary issues include: - The lack of a clear discussion of the objectives of the optimization process in support of decision-making relative to the radiation protection system. - The lack of clear distinctions between the purpose and goals of each of the 3 radiation principles which results in: o the blurring of the optimization and justification principles and o the blurring of dose limitation and optimization principles. - Inadequate discussion of the graded approach to optimization -- implementation of the process should not cost more than the potential benefits associated with the possible dose reductions to be gained by the process. - The document does not adequately address all possible processes by which optimization may be addressed and suggests that optimization is a “never ending” process which is not the case in every situation (cost and resources for optimization should not be exceed potential value. Hence, you should not continue it if there is not value.). Several terms such as “intergenerational issues”, “equity” and “disaggregating” make the text somewhat unclear. Since optimization is one of the more difficult concepts to properly understand and because it plays such a central role in radiation protection, greater clarity is needed in this document. Specific Comments: Executive Summary (pages 6- 8): Because the Executive Summary is the portion of the report that will be read by most individuals, some of whom may not read the entire document, to the extent possible it should be able to stand on its own. At a minimum, the Executive Summary should briefly indicate the goal of optimization relative to the general purpose of the other principles: - Justification ensures that any activity causing exposures are necessary to the national interests or beneficial to society. - Dose limitation ensures that no individual or group of individuals get a disproportionate share of the dose from exposures related to the justified action. - Optimization supports the assessment of alternatives for implementing the justified action in a manner that either maximizes the benefit or minimizes the detriment at or below the dose limit or constraint. Page 8 – It should be clear that the first paragraph is talking in general terms. Optimization can be done by the regulatory authorities such that they define designs, operations or closure requirements in the regulations and no further application of the principle (other than meeting the regulatory requirements) is necessary on the part of the operator. Introduction (Chapter 1, pages 9 and 10): Although the first paragraph recognizes that optimization is one of the fundamental principles of the system of radiological protection, its discussion is far too general. Before beginning the discussion of the historical ICRP guidance on optimization, there should be a brief description in Chapter 1, and then more complete discussion of the three fundamental principles in Chapter 2. It should be clear that optimization, as used in radiation protection, is the optimization of protection, so that the action being taken is implemented in a manner that produces the greatest benefit or least detriment to society. Care should be taken not to inappropriately blur the general goals or objectives of the three fundamental principles. In simple terms: Justification ensures that an action taken that will cause an exposure is overall beneficial or necessary to society and/or to the national interests; Dose Limitation (including constraints) provides assurance that no individual or small group of individuals can receive a disproportionately large share of the dose from exposure to a justified source for the benefit of society in general; Optimization is a process to assist in decisions that ensure that exposures from a justified action are as far below the dose limits or constraints as practical giving due consideration to social and economic factors. The new ICRP position allows for broader considerations when implementing optimization. Hence, the description of optimization now should be: optimization is a process to assist in decisions that ensure implementation of a justified action (in a manner that meets dose constraints) will provide the greatest benefit or the least risk (detriment) to society. This expansion is a welcome improvement and is consistent with many applications of the As Low As is Reasonably Achievable (ALARA) concept in the U.S. and other risk management processes used elsewhere but, to maintain the effectiveness of ICRP radiological protection system, it needs to be clear that: 1) Optimization does not replace or override justification. For example, the government may determine that a certain facility is essential to the national interests and must be built. The goal of optimization is to identify the most effective means of designing, constructing and operating the facility (i.e., the alternative that has the most benefit or least detriment). Also, it should not be implied that Optimization is in any way an affirmation for Justification. 2) Optimization is not used to replace or override dose limits. Because the optimization process selects an alternative that controls exposures to doses that are 10% of the dose limit does not mean the dose limit should be reduced. (Also see additional discussion on use of optimization in regulations developed in comments on Chapter 3.) The History of the Optimisation Principle (Chapter 2, pages 11-14): Although we recommend some general discussion of the three principles and their general goals early in the Introduction (first or second paragraph), more detailed discussion to address the comment on Chapter 1 above should be included in the first section (“Foundation of the Principle”) of Chapter 2 and clarification of the role of justification versus optimization should be considered throughout the “Evolution of the Concept” section. On page 13, paragraph (17), the draft report raises the equity issue and refers the reader back to ICRP 60 for further discussion. However, with the expansion of optimization into benefits and detriments other than dose reduction and exposure, the discussion in ICRP 60 is too limited to address the expanded equity issues. Optimization now needs to consider not only who is exposed and who benefits from the exposure but also, for example, who pays for the reduction or benefits. This is hinted at on page 14, paragraph (19), but not discussed sufficiently. On page 14, paragraph (20), the authors discuss the importance of stakeholder input to the optimization process. Although we agree that consideration of stakeholder concerns and values must be addressed and are important to an effective optimization process, care should be taken not to confuse the need to seek and consider stakeholder input in the decision making process with stakeholder involvement in the decision. For optimization to work and the safety culture to thrive, it must be clear that the decision maker (the operator or in some cases the regulatory authority) is both responsible and accountable for the decision. The decision makers cannot waive that responsibility or explain away their accountability by saying that the decision is what most stakeholders wanted. Therefore, we suggest changing “decision-making” to “optimization” in the 4th line of paragraph (20) so that it reads “…involvement in the radiation protection optimization process” and that a brief discussion consistent with Chapter 6 on who the decision maker is needs to be added here. The Role of the Optimisation Principle in the System of Protection (Chapter 3, pages 15-17): The discussion on page 15, paragraphs (21) through (23) further blurs or confuses the difference between optimization and dose limitation. The ICRP apparently is attempting to broaden the scope of optimization to optimize protection beyond just dose such that more than dose reduction needs to be considered in the analysis. However, the first three paragraphs (instead of explaining the role of optimization in balancing risks and benefits below the dose limits or constraints) proceed to try to explain optimization’s goal or objective as supplementing the dose limitation principle to achieve some enhanced level of protection via dose reduction (paragraph (22)). Where dose limits are applicable the goal of optimization should be to select alternatives that at least meet the dose limits or constraints but overall optimize (lowest detriment or greatest benefit) the system. This discussion seems to indicate that the optimum system always produces the lowest dose – that is not true. There are cases where the optimum (most beneficial system) would exceed the dose limit but the next best optimum system that meets the dose limit will be selected in order to achieve all three principles. This is an important difference between those situations where the ICRP proposes that dose limitation is applicable and those where it is not. Trying to explain the expanded optimization principle in terms of supplementing the dose limitation principle is wrong and will not be effective in improving its implementation or strengthening the safety culture. Paragraph (24) - We support the Commission’s recommendation to base optimization analyses on realistic assumptions. As noted, assessing and comparing the many attributes important to an optimization is difficult and adding varied levels of conservatism to them will only make that comparison more difficult. All attributes should be best estimates of the metric used to measure or evaluate the detriment or the benefit. To have varied conservatism for different attributes is just as likely to lead the decision maker to a bad decision as a good decision. Page 16, Figure 1, and paragraphs (25) through (28) – The figure should be supported by additional discussion so as not to confuse the reader with generalities. First, as noted elsewhere in the report, the Optimization principle calls for the application of a process that will result in doses below the dose constraint. Although the process can result in “authorized levels,” the principle does not require the development of authorized levels below the dose constraint that must be achieved. For example, given a dose constraint of 0.3 mSv/y, the operator may conduct an analysis of several control technologies and find technology X to be the optimum technology because it is the most cost effective, minimizes waste and will result in exposures that will limit doses to on the order of 10% to 30% or less of the dose constraint. The optimization requirement is complied with by the installation, operation and maintenance of that technology. There is no authorized level other than the dose constraint. Hence, it is suggested that the figure eliminate the “authorized level” bar from “Planned Situations” and just have the arrow showing an undefined reduction below the constraint or provide a more detailed explanation of the figure and the authorized level noting that it need not be a specific level. Similarly, the Controllable Existing Situation figure seems to suggest that the existing situation is only considered for mitigation if a constraint is exceeded. That should be clarified as well. We were under the impression that one may choose to intervene in an existing situation if optimization indicated it was beneficial irrespective of the dose constraint. Likewise, if existing levels are below optimization process derived intervention levels ICRP recommended no action. This should be discussed and explained in the text along with what constraints are to be applied. The effective institution of the optimization process depends on the existence of a safety culture. On page 17, paragraph (28), the text states that the role of optimization is to foster a safety culture. Although the document addresses the individual attributes of a safety culture, it does not specify how to achieve this state – which would seem to be more of a behavioral-based safety issue. The current document should provide more guidance on how to achieve the expected safety culture. As noted in comments that follow, some of the discussion which leads one to believe accountability and responsibility are being distributed to a larger group may detract rather than improve the safety culture. The very general discussion in this chapter and the simplistic charts do not seem to meet the goals of the title of the chapter. In addition to the comments above, the section does not adequately discuss the various implementation possibilities for satisfying the optimization requirements of the system. For example, the chapter does not discuss the possibility of the regulator conducting the optimization analyses and dictating design and operational requirements, requiring best available technology be used or simply defining optimization-based “authorized levels” (concentration- or dose-based) that the operator must implement with no further requirements for optimization by the operator. Although such an approach is not necessarily consistent with current “management systems” approaches to environmental protection the approach has and continue to be used. Are we to conclude from this chapter that ICRP finds these other approaches unacceptable or just that ICRP did not choose to address them? In general, the optimization process can be implemented in the following ways: 1. Regulator directs operator to meet dose constraint and optimize 2. Regulator optimizes and defines: o a level (dose, concentration or condition) o Best available technology o Design, operation and maintenance requirements that meet dose constraints and operator simply complies with requirements 3. Regulator optimizes and defines: o a level (dose, concentration or condition) o Best available technology o Design, operation and maintenance requirements but also requires operator to optimize. In the first case above the regulator ensures that the dose constraint has been met and that the operator has a functioning optimization process that enhances performance to the extent practicable given social and economic factors. However, the regulator only evaluates the process not its out come (so long as the constraint is met). In the second case, the regulator oversees the operator to ensure it complies with the level, technological or operating requirements and enforces only those requirements. In the third case, the regulator enforces against the level, technological requirements or operating requirements and ensures that there is a functioning optimization system in place as well. The ICRP’s discussion is too general to determine if any or all of these approaches are acceptable and meet the Commissions goals. The Optimisation Process (Chapter 4, pages 18-25): Page 19 presents a list of sample “attributes” but is not very useful. Some of the listed items could be formulated into an attribute in a multi-attribute analysis but many of the items listed are elements or factors that would be considered or addressed by an attribute. For example, gender and age inequities or the varied risks to different genders or age groups may be factored into the health impacts attribute but how is gender an attribute to be valued and rated? Fairness and equity seem redundant and could be factored into numerous attributes. Who is exposed, who benefits, who pays to mitigate exposure may all be different groups and might need to be assessed and compared when assessing fairness. The payer may not be willing to expend as much to reduce dose as the exposed and the individuals benefiting may have little concern for either. The decision maker, therefore, must fairly balance all of these factors. It is simply not clear what we learn from the list. Is the Commission proposing that we value (weight) doses by age, gender and health status of the target group for example? It is easy to create “non-exhaustive lists” but understanding how to use them is far more difficult. Page 20, “Characteristics of the Process” is a good discussion of the process but it fails to emphasize the fact that the optimization process should be tailored to the need (commensurate with the potential benefits). That is, the level of effort and resources placed on optimization should be proportional to the potential benefit derived from conducting the process. For example, if the maximum collective dose from the project is on the order of 1 person-Sv and the maximum individual dose is a few 10s of microSv per year, there is little to be gained by reducing dose and a costly optimization process cannot be justified. This section should clearly indicate that optimization is a graded approach and the cost of implementing the optimization process should not exceed the potential benefits that may be derived from its implementation. Page 22, paragraph (43), correctly recognizes the importance of relevant stakeholder’s input in identifying and prioritizing attributes that affect the decision. However, so as not to confuse involvement in the process with the actual decision making it should be made clear that the decision maker (regulator or operator) is responsible for fairly evaluating and comparing the attributes to select the most beneficial or least detrimental alternative. The statements in paragraphs (46) and (47) can also be read to confuse the issue as to who is responsible and accountable for the decision (e.g., “…all those involved in the optimisation process are accountable…” and “…requires a commitment from all relevant parties…”). Without question participants in the process should be committed to being productive and not disruptive of the process and they are responsible for communicating their values to the decision maker but to imply that the stakeholder somehow has to be accountable for implementation of the process is extreme, not likely to be agreed to, and has the potential to negatively affect the safety culture by confusing responsibility and spreading accountability across multiple individuals or groups so as to make no one truly accountable. Page 23, paragraph (46), implies that it is necessary to accept the linear non threshold hypothesis to effectively implement the optimization process. Such a statement seems more philosophical than scientific and seems inappropriate for this document. In any case, given the broadened context of optimization exposure may be irrelevant. If, in the comparison of alternatives for implementing an action, one results in a projected three traffic accident-related deaths and another zero deaths, and alternative two caused an additional person-sv of dose, it would not optimum to select option one because of the higher option two dose. Page 24, paragraph (54), recognizes the importance of accountability for the decision suggesting that the government is responsible for the final decision. As discussed in previous comments this is only the case in some situations. In many situations, it is the operator that should and must be responsible and accountable for the decision as to what is optimal, giving due consideration to stakeholder input and regulatory requirements. In many cases the government’s or regulator’s responsibility is only to ensure that the process is implemented and not to “second guess” the result. However, the decision is not “shared” someone must be responsible for assessing the various positions and values and appropriately weighting and comparing the alternative solutions. It is not a vote or election – the decision maker cannot simply count the votes but rather, must evaluate the positions and select the optimum alternative. The key is that the decisions must be open, explained and fully documented. Pages 24-25, Paragraph (55) finally does express the decision-maker’s responsibility and need to consider stakeholder input appropriately. However, if these two lines are missed by the reader, much of the rest of the report can lead the reader to misconstrue that the stakeholders are responsible for the decision. This is worth repeating a few more times throughout the report given the multiple times stakeholder involvement is discussed. Page 25, Paragraphs (56) through (59) provide some good advice that should be integrated into earlier discussion – they would have been useful in trying to explain or frame Figure 1. Page 27, Collective Dose. These paragraphs correctly identify some of the problems with some previous applications of collective dose in past optimization exercises. Discussions to follow go on to discuss how collective dose might be presented to make it more useful to the decision process. The problem with this discussion is that it never clearly identifies the root cause of previous problems associated with global, temporal and spatial integration of doses. The problem is that the decision process did not adequately define the data quality objectives for the analyses. Dose should only be integrated over those space, time or dose quantities that produce data that will be useful to support good decisions. For example, the United States National Academy of Public Administration (NAPA) completed a study in 1997 that addresses key criteria for evaluating risks across generations. The NAPA study delineated a process which indicated the current generation, as the steward for future generations, needs to prioritize its criteria. As a first priority it is essential to protect the current and next few generations then efforts should be made to ensure that the current and next few generations do not take actions that foreclose options that would allow future generations equivalent life styles and to prevent actions that might cause irreversible harm and catastrophic damage in the distant future. These criteria and associated priority lead to the recognition that it is not deemed appropriate to use the same quantitative metrics (either costs, doses or risks) in comparing impacts to the near future (a few hundred years) to impacts in the distant future (e.g. 1000 years or more). Comparing quantitative estimates of integrated doses over thousands of years is of little value to the decision process. Comparisons of collective doses from alternatives over a few 10s to a few hundreds of years may be of value in selecting optimum alternatives but comparisons of quantitative doses or cost estimates out thousands of years are not typically. For long times in the future uncertainties (not only in the estimates but also in the importance of the value) are so great as to make the comparison of quantitative values meaningless. Footnote for NAPA report: Deciding for the Future: Balancing Risks, Costs and Benefits Fairly Across Generations, June 1997, NAPA. What is needed in this chapter is not just a discussion of how one disaggregates collective dose but also how one defines the relative importance of the various groupings. Before calculating any metric, collective dose, cost or whatever, the optimization team must identify how it will be used and what its value will be. When one is looking at a number of potential alternatives that may be implemented, typically, the local or at most the regional collective dose provides sufficient data to demonstrate which of the alternatives have the lowest or highest potential for collective dose and the alternatives can be ranked on that basis. Similarly, where doses occur in the environment and to the general population, little value for the decision process results from numerous gender, age or national groupings of the doses such disaggregation merely complicates the process. The key is to “disaggregate” the data enough to support a good decision but not overly complicate the comparison of attributes of the alternatives. Dose Matrix: The use of the dose matrix will require more effort associated with data analysis. What is not clear is if the increased effort will improve optimization and radiation protection. For occupational radiation protection dose measurements are available, and in many cases, dose registries exist. Thus, it would be relatively easy to increase the level of effort needed to examine radiation dose distributions and other more complex approaches for analyzing worker dose. However, because occupational work is reasonably well defined, collective doses of work groups seem, in most cases, to be sufficient for optimization. For environmental radiation protection, increased analysis may introduce increased uncertainties because of the use of more complex models. The increased uncertainty would cancel the benefit obtained using the more complex models. What would be useful is guidance on when at what times and distances temporal or geographical integration of collective dose should be used for quantitative, or qualitative comparisons or not at all based on the confidence in the data and the outcome of the impact.