Collective dose: kill or cure?

Abstract

The term `collective dose' was used in ICRP Publication 22 [1] to mean the product of the number of individuals in a group and the average dose to those individuals. To allow for a continuous variation of the size of the group with time, the definition becomes the integral over time of the product of the mean dose rate in a group and the size of that group. This quantity has sometimes been called the collective dose commitment, but it is still a simple collective dose. The usefulness of collective dose needs to be re-examined. In general terms, a quantity that is expressed as the product of two components is useful whenever there is reciprocity between the two components. The significance of the magnitude of the product will then not change when one component is increased by a factor and the other is decreased by the same factor. The first requirement for reciprocity in collective dose is a proportional dose-effect relationship over the relevant ranges of dose and dose rate (a linear, non-threshold relationship). This is not the place to discuss dose-effect relationships, beyond saying that ICRP still believes that a proportional relationship is the one most likely to remain consistent with the biological data. Of course, a collective dose can be specified without a proportional relationship, but it will be of limited use. It could indicate comparisons or trends of exposure, but would not indicate their significance. For protection purposes, the reciprocity should also apply to society's responses to radiation exposures. This reciprocity is widely assumed, but it may be an error. The existence of the universal background of radiation from natural sources does not provide any justification for adding doses from manmade sources, but it does provide a basis for comparisons. If the artificial additions are trivial compared with natural background, the public reaction may well be lower per unit added dose than it would be for larger added doses, i.e. the response may not be proportional to the added dose. If this is so, a correction could be introduced by applying a dose-related weighting factor to the dose in the definition of collective dose. This concept has not been widely adopted. If there were a universal acceptance of reciprocity, would there still be doubts about the usefulness of collective dose? Yes, I believe there would and, yes, I believe the doubts would be valid. In the first place, collective dose conceals information about the distribution of the individual doses in both magnitude and time. In ICRP Publication 77 [2], the Commission recommended that the collective dose should be partly disaggregated to give the decision maker a wider basis for choice. A further problem with collective dose is that it encourages the user to extend the predictive models to regions of dose and time where the uncertainties are overwhelming. Finally, the collective dose encourages the use of automatic, rather than well considered, procedures for taking protection decisions. It discourages the use of common sense. I see these problems as flaws that are serious, but not fatal. I do not want to kill collective dose, but I would like to cure it. I would like to see a changed quantity, but a `User's Manual' might be sufficient. The most obvious change is to limit the timescale over which collective dose is defined. ICRP Publication 77 provided some guidance as follows: `In general, ... forecasts of collective dose over times longer than several thousand years and forecasts of health detriment over times longer than several hundred years should be examined critically'. For defining collective dose, the time limit might be selected from this range. A more controversial change would be to exclude small doses from the definition of collective dose. One objection to this proposal has been due to a fear of a serious build-up of small doses from many sources. In fact, such a build up is not threatened at present and could easily be prevented in future. In short, collective dose is a useful tool in the hands of a sensible practitioner, but its functions are limited. The present misuses could be avoided by additional safeguards, either in the definition or in a User's Manual.