NDAWG comments on ICRP foundation documents The National Dose Assessment Working Group (NDAWG) was established in the UK in 2002. The aim of NDAWG is to bring together people and organisations with responsibility for, and/or an interest in, the assessment of radiation doses from the operation of the nuclear industry and from minor users of radioactivity. Further information on the work of NDAWG, including the current membership, is available on the website, www.ndawg.org. Members of NDAWG have considered the draft Foundation document of the International Commission on Radiological Protection (ICRP) as published on the ICRP website. The following set of comments represent the views of the majority of members but not necessarily the organisations that they represent. Assessing Dose of the Representative Individual for the Purpose of Radiation Protection of the Public NDAWG members welcome the publication of a document on assessing doses to individuals for protecting members of the public. This is an important area and it is a good idea to bring together and clarify all previous ICRP guidance on the topic. We support the basic premise of the representative individual and much of the guidance given in this draft document. However, it would be helpful if the report made the relationship between the old critical group concept and the new approach of the representative individual clear from the outset. It would be sufficient to state that the representative individual is effectively an average member of the critical group and that many of the principles associated with critical group therefore also apply to the representative individual. Here the term “representative individual” is simply replacing the term “critical group”. The advice that generally it is sufficient to consider three age groups in assessing doses to the public is welcomed. This reflects what is done in practice in the UK and advice given in both the UK and the EU. We also welcome the advice on taking account of exposure of the fetus, although this needs to be expanded. There are only a few lines at the end of paragraph 79 on fetal dose and there is nothing on the fetus in the executive summary which mentions the three age groups only. We support the recognition of the role of stakeholders in the process of identifying characteristics of the representative individual. A definition of the possible roles for stakeholders would be useful, along the lines given in the document on optimisation. The suggested approach where calculations are carried out in such a way that uncertainty in dose estimates and the variability in dose estimates due to the different habits and locations of individual are kept separate is supported. In several places in the document it is stated that retrospective doses are for ‘specific (real) individuals’ (e.g Abstract and paragraphs S4, S18, 28). This is not necessarily the case. For example, a retrospective assessment might be required to consider the significance of measured activity concentrations in the environment and this might be done on the basis of an assessed dose to a hypothetical individual. An assessment of radiation doses from past practices might also be carried out to compare the impact of different sources and again hypothetical individuals might be considered. Although the draft document discusses the concept of homogeneity it would be helpful if further guidance could be given. In particular guidance would be welcomed from ICRP on the extent to which homogeneity in habits/characteristics is important as opposed to homogeneity in dose. If a distribution of doses is estimated in a population taking account of the different habits it is possible to identify individuals with similar doses but received in very different ways (e.g for one individual from consuming seafood and for another from living close to the source of an atmospheric discharge and consuming terrestrial foods.) Can such doses be deemed to be homogenous for the purpose of defining a representative individual? Paragraph S11 states that homogeneity in habit data is important while paragraph 14 can be interpreted that it is homogeneity in dose that is important. Specific comments on the text: The Commission are unclear whether the representative individual should be hypothetical or specific. The abstract admits both but paragraph S9 says just hypothetical; paragraphs 23 and 28 again say both. Paragraph S3 says just hypothetical. The material in Section B.3 could be brought forward to earlier in the main text to clarify the issues for defining distributions of habits and behaviour for hypothetical populations. Para S7. We welcome the clarification of the status of dose coefficients with respect to uncertainty. Para. S8 (and third paragraph of abstract). Although in general three age groups are sufficient paragraph 79 does raise the issue of exposure to the fetus. It would be worth adding a caveat to these paragraphs to cover the rare cases when the fetus needs to be considered. Para S12. Opens with ‘If specific habit data for a local population are not available, values may be derived from national or regional population data’. What should be done if specific habit data for a local population actually are available? Statistical treatment of data (use of habit data percentiles for example) derived from local population are only likely to be meaningful if the local population is sampled randomly or an 100% survey in the locality is carried out. Para S17. States that ‘care must be taken to include all hypothetical individuals whose dose could possibly be representative of persons receiving the highest dose’. The only way to be sure that such individuals and their habits have been identified/found would be to carry out a 100% survey of all individuals in the locality. This is not considered to be a reasonably practicable approach to collecting information on habits. Para S18. It is unlikely to be acceptable to modify the habits of the critical group as a means of reducing doses in routine conditions. If doses are unacceptable then generally discharges should be reduced. Para 23. It would help to specify here that the representative individual replaces the critical group. Para 25. Is the last part of this paragraph consistent with the main ICRP recommendations? The draft issued for consultation in 2004 specified going beyond the ALARA principle. Para. 41, It is implied that site-specific data are always preferred to regional or national data on human habits or behaviour. However, it is useful to audit the site-specific data against regional or national data, as site-specific data may be limited in scope and may only reveal a snapshot of habits and behaviour. There is also the possibility of using (say) a 5 year rolling average approach to remove annual one-off situations. Para 63. See comment on Para 86. The use of a percentile approach to derive habits for a deterministic assessment is useful for analysing national data sets, where a large number of individuals can be randomly selected. However, a more practicable approach to collecting information on habits around sources has been to target the most highly exposed individuals including high consumers. This has the added benefit of providing confidence that some of the most exposed individuals have been included. Unfortunately, data collected in this manner is not amenable to a percentile approach for selection of habits as it is not randomly sampled. It would be helpful if the ICRP recognised this approach. Para 65, If the environment and population differ, leading to varying habits and hence doses to the public, there is justification for radiation protection reasons for different operating regimes. Para 69. It is not clear whether the statement in this paragraph with regard to homogeneity is still appropriate if use of the 95th percentile of distributions is being recommended. Para 74, It is not always the case that a period of 70 years is considered for accumulation of radionuclides in the body. For adults a period of 50 years is used. It would be clearer to say to age 70 y. Para 78, It would be clearer if in the 4th line ‘0-to-5-year age category’ was replaced by ‘infant category’. Para 79, The idea that infant doses will also be high for these radionuclides (4th sentence) does not really apply to doses from phosphorus-32 and 33, particularly when considering doses from the ingestion of fish following discharges to freshwater. This sentence could therefore be misleading for these potentially important radionuclides where it will be important to consider the fetus. Para 82, It is not clear why all available age-specific dose coefficients should be used in relation to accidental releases. Assessed doses from accidental releases are particularly uncertain and so the increased sophistication of considering 7 different age groups does not seem worthwhile. Para 86. It is true that it is important to identify the region over which the assessment is conducted and no doubt each situation will be unique. However, if the 95th percentile is to be used to define habit data for a deterministic assessment or to assess compliance against the dose constraints for a probabilistic assessment, the outcome of the assessment could be highly sensitive to the region selected. ICRP should consider providing some additional guidance, such as an approximate number of people whose habits are to be assessed. The 95th percentile may be appropriate when analysing the habits for a group of several tens or a few hundred individuals. However, in the UK, historically, the 97.5th percentile has been seen as a better measure of critical habits when analysing a sample of many hundreds of individuals to derive UK generic habits. Equally, if the group selected is the critical group (as per the old approach) then the habits selected should be the 50th percentile. Para, 88, The definition and explanation of the reference individual for probabilistic assessments needs to be made much clearer. One reading of this paragraph suggests that the representative individual should be defined post hoc once the assessment has been done. In this case the dose is almost certain to be lower than the constraint - because by selection there is only a 5% chance of exceedance. Another reading suggests that the reference individual should be rather like the current critical group - in which case the criterion of 5% probability for exceedance of the constraint, taking account of all sources of uncertainty, is likely to be always more restrictive than the results of a deterministic assessment. Finally there seems to be a contradiction in any case in saying the dose should be less than the constraint - in probabilistic assessment the output will always be a distribution, not a single value. Para 89, This paragraph is rather obscure and probably unnecessary. Para 90, This comes back to the comment on paragraph S18. If we have a past practice and it is continuing into the future, then if we exceed the constraint we can do something about the source. If we have an existing situation, we do not really have a source to control - the activity is somewhere in an environment and therefore not totally under our control. We can of course reduce doses via intervention. Is this what is meant by changing habits? If it is perhaps the scope of the term habits needs to be clearer.