Register for Updates | Search | Contacts | Site Map | Member Login

ICRP: Free the Annals!

View Comment

Submitted by Richard V. Osborne, Consultant; former member of ICRP Committee 4
   Commenting as an individual
Document Recommendations
Comments on the Draft Recommendations of the International Commission on Radiological Protection, 02/275/06 – 5 June 2006

The current draft document mostly reads well and is a much clearer presentation of the principles of radiological protection and their implementation. There are some topics though that could be treated more clearly such as the discussions of LNT, dose constraints, optimization and controllability. These four topics are discussed in the sections below. Detailed editing points are listed at the end.


The report variously refers to the linear non-threshold hypothesis and the linear non-threshold relationship as the best approach for managing risk from radiation. It would be much clearer, and possibly less controversial, to refer consistently to a linear non-threshold model. A hypothesis is a limited statement regarding cause and effect in specific situations. A model is a description of a relationship when it is known that the hypothesis has limited (or undeterminable) validity. The latter is a more accurate description of the situation here and would be consistent with the application of models elsewhere in radiological protection – biokinetic models, or phantoms for example. Such terminology would not lend itself as easily to the interminable debate on the validity of LNT. Also, the use that ICRP makes of such a model for protection purposes, while at the same time arguing that it should not be applied for estimating health effects in an exposed population, as in paragraphs (147), (147) for example, would appear more reasonable and persuasive.

Dose constraints

The document correctly draws attention to the difference between individual-related dose limits and source-related dose limits. It also points to the inappropriateness of applying the limits defined for particular situations (occupational exposures, public exposures from practices) to other situations. Dose constraints are presented as the way of avoiding the misapplication of the well-established individual-related dose limits and of achieving the apropriate levels of protection in all situations. Though the distinction of source-related and individual-related quantities is quite logical and is important, the presentation of dose constraints in this document as being the “fundamental” means of control in all situations has led to misunderstandings as well as justifiable concern that their application leads to another layer of restrictive regulation. The desired ends can be achieved more simply in a less confusing way.

There are five distinct applications recommended for dose constraints:
(1) Provide a level of protection for an individual from a single defined source (para.184).
(2) Take other sources under regulatory control into account (paras. 198, 216).
(3) Reduce inequity in the distribution of doses when optimising protection (para. 199). (4) Reflect general practice, benefits, international guidance. . . (para. 210).
(5) Indicate when action is almost always warranted (para 235).

These are quite disparate applications and lumping them together under the banner of dose constraint application is not helpful. It would be much better to emphasize the principles and considerations needed in the situations reflected by the five suggested applications, which may or may not be helped by the introduction of a dose or annual dose criterion specific to that situation. Application (1) reflects the basic distinction between individual-related and source-related quantities. The key point here is that the individual limits defined for occupational exposures and for exposures of the public from practices are for the doses to individuals from the totality of all occupational or public exposures for regulated practices. Clearly, other dose criteria should be applied for other types of exposure – for example, application (5). The concept of dose constraint is applicable here.

The application of the same quantity for adjusting protective standards to reflect multiple sources (application (2)) and for leveling distributions of dose in optimising (application (3)) is confusing and is unnecessary. In application (2), there is no reason to use a value different from the numerical value of the individual annual dose limit unless there are (or are likely to be) individuals who are the most highly exposed from more than the one source – which is a rare occurrence. Such determination should be left to the national authorities. If the national authority were to decide that there are, in fact, likely to be most-exposed individuals common to two or more practices, then the derived limit(s) associated with the practices being considered could be reduced from those that correspond to the numerical value of the individual dose limit. The reduced limit would reflect the needs of that particular situation – it would not be the ab initio application of a more restrictive dose constraint that is arbitrarily half or one third the individual limit, and it should not be called a dose constraint. (Optimisation of protection may well lead to a further reduction in any derived limits anyway.)

There is also the confusion in the present text stemming from paragraph (197) where it is noted that the dose constraints are to be used only prospectively and not for “retrospective dose limitation”, which appears to mean not for compliance purposes. It is difficult to see how this is reconciled with the suggested application of a dose constraint in application (2) above or even some of the others. Avoiding applying dose constraints, as suggested here, when there is an individual-related dose limit that can be adapted to the source-related application would avoid this confusion.

In application (3) where the aim is to level out the distribution of doses when protection is being optimised, the imposition of an arbitrary constraint in an optimising is counter to the idea of optimising. Any inequity in the distribution of doses can be taken into account by a weighting that reflects the importance ascribed to that particular factor.

In application (4), the reflection of general practice, international guidance etc., then for those situations that are not associated with defined individual limits, the application of a dose criterion called a dose constraint makes sense – as with medical practice for example. With regulated practices though, general practice, BAT, or whatever, is best considered as part of optimising, not, as suggested in paragraph (210), by applying a dose constraint a priori to the optimising but as some appropriately weighted input factor in the optimising analysis.

In summary, the use of “dose constraints” should be restricted to situations other than those involving regulated practices. The individual-related versus source-related issue can be resolved by adjustment of the individual-related limit to reflect the particular situation being considered. Such an approach would likely be much more acceptable to the radiological protection community.

A further comment: Part of the problem stems from the early (though laudable) attempt to reduce to a few values the multiplicity of numerical values of dose or annual dose that had been introduced over the years as various decision-aiding criteria or limits. Table 4 (which should be Table 5) reflects the broadened rationalisation but it is still something of a force fit of current values, ex post facto, into ranges defined by the individual annual dose limit for members of the public from practices, the individual annual dose limit for occupational exposures and an introduced upper value of 100 mSv (/a). The inclusion of doses together with annual doses in Table 4 is confusing. As well, it would be a mistake to see the Table as definitive and useful prospectively. This caveat should be noted in the text.

Optimisation. Section 5.8.7, paragraphs (221) – (233)

This section is far too wordy, has far too many generalities and has phrases that are difficult to understand: the key recommendations could be presented much more clearly.

An example of the repetition is in the explanation of what optimisation is. The various descriptions that are spread throughout the section are shown below (paragraph numbers are shown in parenthesis).

"The optimisation of protection is a forward-looking iterative process aimed at preventing exposures . . ."(221)
"It is continuous, taking into account both technical and socio-economic developments and requires both qualitative and quantitative judgements." (221)
"It is the result of an evaluation, which carefully balances the detriment from the exposure (economic, human, social, political, etc.) and the resources available for the protection of individuals." (227)
"The requirement to optimise protection…. is a continuing process. "(224)
"The process should be systematic and carefully structured to ensure that all relevant aspects are taken into account. "(221)
"Optimisation is a frame of mind, always questioning whether the best has been done in the prevailing circumstances, and if all that is reasonable has been done to reduce doses." (221)
"The Commission wishes to emphasise that optimisation is not minimisation." (227)
" . . .optimisation is more an obligation of means than of results." (233)
"The best option is always specific to the exposure situation and represents the best level of protection that can be achieved under the circumstances." (226)

Another example of repetition concerns the roles and responsibilities:

"An active safety culture supports the successful application of optimisation by both the operational management and by the authority. "(232)
"This means that the responsible agencies should periodically ask whether all that is reasonable has been done to reduce doses." (224)
" . .both the operators and the appropriate national authorities have responsibilities for applying the optimisation principle. "(232)
"The implementation of the process of optimisation of protection is the responsibility of the operating management, subject to the requirements of the authority. "(232)
"It also requires the commitment at all levels in all concerned organisations as well as adequate procedures and resources. "(221)

These could be presented much more succinctly and clearly.

Redundancies and wordings that are difficult to understand are noted in the paragraph-based listing at the end of this commentary.

The discussion in paragraph (223), concerning the selection of a constraint, indicates that the value selected comes out of the optimisation, rather than is an input into it as is suggested elsewhere in the document. This is part of the confusion noted above. By avoiding the application of a constraint as such, as suggested above, this confusion could be avoided.

The question of disaggregation of collective dose and its application in optimising is dealt with in paragraphs (228), (229), (231). In paragraphs (228) and (229) the basis is not given for the assertions that the aggregated collective dose is not useful and may even be misleading. Accordingly, there is no basis presented that would guide the consideration of the separate components. If the basis is uncertainty as discussed in paragraph (231) then this topic would be much more clearly presented by starting with the ideas in paragraph (231) which lead logically to the need for disaggregation as discussed in (228) and (229).


The attribute of “controllability” and the adjective “controllable” are treated as if they are self-defining (see, for example, paragraphs (21), (31) ). However, they are not and the text itself indicates this; in paragraph (31) we find “The Commission defines controllable dose as the dose…….that can reasonably be controlled by whatever means.” and a similar definition in paragraph (35). What is reasonable? The implication is that optimisation is involved. Indeed, in paragraphs (41) and (42) there is an excellent discussion of exclusion and exemption that sets the scene in a far more logical manner. The earlier paragraphs could be adjusted to reflect that the Commission’s recommendations do cover all sources and to avoid using controllability as a meaningful description. The optimum regulatory action required for some sources and in particular circumstances may be to exclude or exempt but it is not that such sources have been ignored. This is a much clearer approach than trying to make controllability a criterion by itself.

Detailed comments by paragraph number

(58) “The Commission also notes that since the estimation of nominal cancer risk coefficients is based upon direct human epidemiological data, any contribution from these cellular phenomena would be included in that estimate.” If the processes such as adaptive response were have an impact only at doses less than 50 mSv say, then they would not have influenced the estimates of risk since these are determined primarily by data in higher dose ranges.

(139) “Effective dose is not based on data from individual persons (see Annex B). In its general application, effective dose does not provide an individual-specific dose but rather that for a reference person.” These are particularly important points – they should be emphasized.

(162) The word “situation(s)” is used extensively in the report and is generally redundant. Here, for example, it would be better to have “planned operations”, “emergencies”, and “existing exposures”.

(176) “The representative individual, who may be hypothetical, receives a dose that is representative of the more highly exposed individuals in the population. It is important that the individual habits . . . are typical habits . . . and not the extreme habits of a single member of the population.” This is another recommendation that needs emphasizing.

(188) line 7: “it” rather than “they”

(203), (210) The references here to Table 4 should be to Table 5. The Table labeled Table 4 on page 61 should be Table 5.

(221) line 2: “before they occur” is redundant.
line 3/4: What are “qualitative and quantitative judgements”?
line 4: “carefully” is redundant.
line 5: How can optimisation, which is a process, be “a frame of mind”?

(227) line 1: Optimisation of what? Minimisation of what?
line 2: Optimisation is a process – it is not a “result”.

(252) The introduction of diagnostic reference levels could be much more direct – the first three sentences are not needed.

(269) Why are insurance companies singled out as having ICRP’s permission to require people to receive medical exposures? What is the basis for this? We can have major public hearings to decide if a doses of a fraction of a millisievert from a nuclear practice are acceptable but insurance companies can order hundreds of doses of “a few millisieverts” at will?

(279) “Below the relevant constraint, action may be undertaken if technically easy and not entailing excessive cost, to reach an optimised level of exposure.” “if technically easy .not entailing excessive cost” is yet another definition of optimisation and is redundant here.

(294) The first sentence is terribly convoluted and somewhat circular. “Many natural radiation sources are not amenable to control in that they are unavoidable or uncontrollable, at least without inordinate effort, or while theoretically controllable are not feasible to control.” There is the circularity “not amenable to control . . in that they are uncontrollable”. Some rewriting is needed to avoid this and clarify generally.

(295) Section 2.2 is referred too; another section must have been intended.

(323) line 8: “together” would be better than “an integrated approach” . Is the last rather platitudinous sentence really needed?

(326) “stochastic” and “deterministic” are no longer the preferred descriptions according to Section 3.

(331) – (333) These paragraphs need restructuring; as it is writen, quite different accidents are mixed up. It would be better to start by noting the two general kinds of accidents with an example for each:

(1) Accidents involving patients
Incorrect calibration (e.g., Costa Rica)
Machine malfunctioning (e.g., Poland)
The comment from paragraph (331) could go here

(2) Accidents involving staff and the public
Inadequate prevention of access to high radiation fields (e.g., Vermont)
Inadequate control of radiotherapy sources (e.g., Thailand)
The comment from paragraph (333) could go here. Leave out the analogy to military strategy – it is not unique to the military.

(334) A comment is needed on existing exposures in this introductory paragraph. The material of paragraph (340) could be placed here.