PART 3 OF 3 Per Wikman, Philosophy Unit, Royal Institute of Technology, Stockholm, Sweden. perwi@infra.kth.se 3. OPTIMISATION 3.1. SUMMARY The new draft states that the collective dose “aggregates information excessively” in cases where a large population is exposed to small individual doses. It is not clear what this means. If the risk is without a threshold and proportional to the dose, the collective dose gives an adequate measure of the expected detriment. In such cases, the aggregation of small individual doses does not suffer from any serious excessive aggregation. It is therefore suggested that a better explanation is given in what sense the collective dose “aggregates information excessively” in cases of exposure of small individual doses to a large population. It is another matter how we should evaluate the estimation of the detriment. It is important to separate the two issues. First an estimation of a matter of fact should be made and then a value judgement should be made. This difference ought to be expressed in the new Recommendations and any value judgements the ICRP makes ought to be explicit and should be based by reasoning. The ICRP has previously discussed such value judgement, in form of the principle: “If the risk of harm to the health of the most exposed individual is trivial, then the total risk is trivial—irrespective of how many people are exposed.” It is a value judgement to the effect that very small doses to a large number of people may be ignored. However, the principle yields absurd consequences if applied strictly. It can hardly be justified by a change from a “utilitarian approach” to any reasonable rights-based or a duty-based ethics. If the goal of the new Recommendations should be “more concern for the individual”, the principle fails its purpose, as the likely result of applying the principle is less protection of the individual. 3.2. AGGREGATE INFORMATION EXCESSIVELY? Section 7.2 on optimisation in the draft claims that ”[...] the integral of low individual exposures over large populations, large geographic areas, and over large periods of time is generally not a useful tool for decision aiding because this may aggregate information excessively.” (ICRP 2004, Paragraph 198, 200) It is not clear what this really means, but the base of this claim seems to be that “a large dose to a small number of people is not equivalent to a small dose to many people, even if the two cases correspond to numerically equal collective doses.” (ICRP 2004, Paragraph 200). But this is not a relevant argument in the context where a large population is exposed to small individual doses. In these cases the aggregation of small individual doses does not suffer from any serious excessive aggregation, instead the collective dose gives an adequate measure of the expected detriment. This is due to the fact that the collective dose is a measure of the expected detriment if the risk from ionizing radiation is proportional to the dose and without threshold (Lindell 2001). This was also explicitly recognised in the previous Recommendations: “The collective effective dose is an adequate representation of the collective detriment” (ICRP 1991, Paragraph 118). If the Commission does not think that the collective dose is not an adequate representation of the detriment anymore, then this ought be stated in the new Recommendations, and the reasons for this changed view ought to be explained. The present reference to “excessive aggregation” is just not plausible in the cases under discussion. In fact, as long as the individual dose poses an increased risk of deleterious effects, a large enough population may yield a considerable total effect. It is a mathematical fact that exposing ten million people to an individual risk of a one in a million probability of getting cancer means that the expected detriment will be 10 cancers. Also, in this case the mathematical “law of large numbers” yields that there is a large probability that the actual outcome will be somewhere around that figure. It might be countered that we do not know that there really is a risk for very small doses of ionising radiation. The outcome may be zero extra cancers, or may even be fewer cancers. This is a relevant objection. However, the best available scientific evaluations of the risks yields that it is likely that there is no threshold for the risk (e.g. UNSCEAR). Also, the ICRP still acknowledges the linear, no threshold model for radiation. Therefore, the Commission ought to draw the consequences of this and accept that a large collective dose composed of small individual doses should be considered as a measure of the expected detriment. It is another matter how we evaluate this. It is important to separate the two issues. The estimation of the detriment is not a value judgement – it is an estimation of a matter of fact. After making such an estimation we have to make a value judgement about the importance of the estimation. We could for example make the value judgement that “the uncertainties are so large that the measure ought not to be used in a decision”, or we could make another value judgement that “even if there are uncertainties, it is likely that there is a risk even for small doses and this is something we ought to consider”. Whatever path we take it is a value judgement, and as such it ought to be explicit. The ICRP has previous made a value judgement related to the collective dose, reflected in what I have called “the disaggregation principle”. The principle is not present in the new draft; however, it is likely that this principle is still in effect. The reason for suspecting this is that it has been present for a long time in the communication from the ICRP on the new Recommendations and that the effect of the disaggregation principle is precisely that the collective dose for large populations and small individual doses is not important. This is discussed in the next section. 3.3. THE DISAGGREGATION PRINCIPLE This particular principle, has until very recently been explicit in the communication regarding the new Recommendations. ICRP’s chairman Roger Clarke introduced it in his “controllable-dose” paper where he laid out the basic framework of the new Recommendations: “The protection philosophy for controllable dose is based on the individual. If the individual is sufficiently protected from a single source, then that is a sufficient criterion for the control of the source. The principle is ‘If the risk of harm to the health of the most exposed individual is trivial, then the total risk is trivial—irrespective of how many people are exposed.’” (Clarke 1999, p.111). Slightly different formulated, the principle was still present in the ICRP communication as recently as 2003: “For each source, basic standards of protection are applied for the most exposed individuals, which also protect society – constraints. If the individual is sufficiently protected from a source, then society is also protected from that source. However, there is a further duty to reduce doses, so as to achieve a higher level of protection when feasible and practicable. This leads to authorised levels — optimisation.“ (ICRP 2003, p. 133-134). The principle is not explicitly formulated in the present draft, but given the fact that it has until recently has been present in communication from the ICRP, it is plausible that it has guided the current presentation of the new Recommendations. I have elsewhere (Wikman 2004a and 2004b) discussed this principle which I denoted the “disaggregation principle”. It can be written as: (D) If the individual is sufficiently protected from X, then society is also protected from X. The disaggregation principle may at first sight seem very reasonable, but it may lead to absurd consequences depending on what we mean by “sufficiently protected”. One idea that has been proposed is that an individual is ”sufficiently protected” if the probability of being harmed for an individual is small enough. A popular value is a probability of harm below one in a million. What are the implications of this interpretation? Very strange it turns out. To see this, just assume that X means ”terrorist acts”. Then, it follows from principle (D) that society is sufficiently protected from terrorist acts, as long as the individual probability of being the victim of a terrorist act is less than one in a million. This is of course absurd, and can hardly be an acceptable political view. Another idea is that an individual is “sufficiently protected” as long as the dose to that individual is very small (say 1/100th of the average annual contribution to an individual from the natural background radiation). Also assume that this (surely trivial?) level of dose implies a probability of, say one in a million, of a acquiring a lethal cancer. But from these assumptions it follows that if 10 million people each are exposed to this level of radiation, the expected detriment is that 10 individuals are likely to get a lethal cancer. But according to principle (D) this also implies that society should be considered sufficiently protected. It is a value judgement what weight we should give to the outcome (expected detriment = 10 cancers). The principle states that we need not consider the fact that ten individuals are likely to be harmed. That we may disregard from the likely harm of people is to say the least a highly controversial ethical position. Someone might object that the principle is not intended for public policy against terrorists, but for the regulation of hazards from ionising radiation. Also, it can be argued that no victims of terrorism are ever acceptable, while we have to accept that people are exposed to radiation. The problem is that the principle does not take these kinds of considerations into account. The principle is “absolutist” in this way, and this is the reason why it is so controversial. Instead, a more reasonable principle ought to consider other things such as benefits of allowing exposure to radiation and the cost of avoiding exposure. 3.4. JUSTIFICATION OF THE DISAGGREGATION PRINCIPLE? In the present draft the Commission states that “[in societal development] there has been a move from the utilitarian approach of the ‘greatest good for the greatest number’, to one with more concern for the ‘individual’, … which have inevitably led to some changes in the formulation of the recommendations.” (ICRP Draft 2004, p 10) It is not clear what “more concern for the individual” means here, but in ethical theory so-called duty-based and rights-based ethics are often set in opposition to utilitarian ethics. It could be that the ICRP thinks that the disaggregation principle follows from a rights-based or a duty-based ethical point of view. The thesis that the disaggregation principle follows from rights-based or duty-based ethics have been examined by me elsewhere (Wikman 2004a). My conclusion in that paper was that the disaggregation principle would hardly be allowed by any reasonable rights-based or duty-based ethics. The main reason for my conclusion is that the principle implies that individuals’ right not to be harmed (or exposed to risk of harm) will be violated, as the principle allows exposure to individuals, as long as the risk is small. Also, the duty towards individuals not to harm them or expose them to unnecessary risks would be violated by the principle. Furthermore, if the societal development now has ”more concern for the individual” we should foremost ask what the consequences for individuals would be if we were to follow the ethical value judgement represented by the disaggregation principle. It is likely that an application of the principle would lead to less rather than more protection for individuals, and thus also less “concern for the individual”. Finally, I want to emphasise that this does not mean that we always should consider the collective dose as the most important factor in a decision making process. However, as argued above, there are good reasons that the collective dose should not be disregarded, even if it means the exposure of a large population to small individual doses. At least it should be a factor to take into consideration, appreciating the consequences of the linear, no threshold model – that it is a measure of the expected detriment. Of course, this does not mean that the expected detriment is necessarily unacceptable, only that it should always be given a proper weight in the decision. 3.5. SUGGESTIONS My suggestion is that 1) A better explanation is given in what sense the collective dose regarding exposure of small individual doses to a large population “aggregates information excessively”. 2) If the Commission does not think anymore that the collective dose is not an adequate representation of the collective detriment this ought be stated in the new Recommendations, and what the reasons for this changed view are. 3) The distinction between estimation of the matter of fact and value judgement is made. 4) Any value judgements made for disregarding the collective dose in such cases is made explicit. 5) Reasons for any value judgements ought to be given. 4. REFERENCES Clarke, R, 1999, “Control of low-level radiation exposure: time for a change?” Journal of Radiological Protection, Vol. 19, No 2, p 107-115. ICRP, 2001, “A report on progress towards new recommendations: A communication from the International Commission on Radiological Protection”, Journal of Radiological Protection, Vol. 21, p 113-123. ICRP, 2003, “The evolution of the system of radiological protection: the justification for the new ICRP recommendations”, Journal of Radiological Protection, Vol. 23, p 129-142. ICRP, 2004, ”Draft for consultation: 2005 Recommendations of the ICRP”, Available at http://www.icrp.org (2004-09-29) Lindell, B, 2001, “Logic and ethics in radiation protection”, Journal of Radiological Protection, Vol. 21,p 377-380 Shrader-Frechette K and Persson L, 2002, ”Ethical, logical and scientific problems with the new ICRP proposals”, J. Radiol. Prot. 22 149–61 Wikman, P, 2004a, ”Trivial Risks and the New Radiation Protection System.”, Journal of Radiological Protection. Vol 24(1):3-11. Wikman, P, 2004b, ”The Philosophy of the New Radiation Protection System”, in Proceedings of the 11th International Congress of the International Radiation Protection Association, available at http://www.irpa11.com