INVITED EDITORIAL

The limits of epidemiology

Published under licence by IOP Publishing Ltd
, , Citation Sir Richard Doll 1999 J. Radiol. Prot. 19 1 DOI 10.1088/0952-4746/19/1/001

0952-4746/19/1/1

Abstract

ICRF/MRC/BHF Clinical Trial Service Unit and Epidemiological Studies Unit, Harkness Building, Radcliffe Infirmary, Oxford OX2 6HE, UK

The National Registry for Radiation Workers, which has enabled Muirhead and his colleagues to provide important new data on the risks associated with chronic exposure to very low doses of ionising radiation in this issue, owes its origin to recommendations made by the Royal Commission on Environmental Pollution (1976) under the Chairmanship of Sir Brian Flowers, as Lord Flowers then was. The Commission was concerned, inter alia, that a report on the mortality of Windscale workers who had been exposed to plutonium included only the numbers of deaths observed and expected among those who were actively employed (Dolphin 1976). There was a need, the Commission believed, for a complete national register of radiation workers to be kept by the National Radiological Protection Board, with a section comprising those exposed to plutonium and the higher actinides, which would include those who had retired or left the industry.

The Board differed then from the one we have had for the last 15 years, and when the register was set up on 1 January 1976, it was limited to workers who opted to be included on it, a limitation that, to my knowledge, has never been imposed on any other industrial cohort that was the subject of research to establish the extent of the hazards to which its members were exposed. That the research should be known to and approved by representatives of the workers is important, but that it should require workers to opt in individually could seriously reduce the numbers available for study and might also introduce a potential source of bias into the findings. The restriction was subsequently changed to the provision of an opportunity for individuals to opt out - still an undesirable and unnecessary condition - but one which, as it has been operated, has generally resulted in such a low refusal rate that only a few units, in which the rate was greater than 10%, have had to be excluded from the study.

The register has now been in use for over 20 years and includes over 120 000 men and women, some of whom were first employed as long ago as 1955, but even so follow-up to the end of 1992 has recorded less than 13 000 deaths (10.4%) and, in view of the small doses to which the workers were exposed (the mean cumulative dose having been 30.5 mSv), this is far too few to establish the precise effects that low doses received at low dose rates are likely to cause. This is not a criticism of the study. On the contrary, it is a good reason for supporting its continuation and its participation in the international study of radiation workers that the International Agency for Research on Cancer (Cardis et al 1995, Cardis and Martuzzi 1998) is now carrying out.

With the current findings it is perhaps understandable that some might argue that those who were conducting these studies were pursuing a will-o'-the-wisp, that the ultimate object will always be just out of reach, and that the conclusion will always be that larger numbers and more research are needed. In a sense this is true. Sooner or later we have to base our estimate of risk at very low doses, such as those (say) from natural background radiation, on some model, whether derived from laboratory experiment, epidemiological evidence of the dose-response relationship at higher doses, or, better, a combination of the two. But we have not reached that point yet. The new data from the registry of radiation workers, reported on pages 3-26, give some assurance that the limitations of epidemiology are less restrictive than the French Academy of Sciences would imply (Académie des Sciences 1997, Tubiana 1998) and they strengthen our confidence that estimates of the effect of doses less than 200 mSv, based on the dose-response relationships observed with higher doses in the survivors of the atomic bomb explosions, are not far out. They don't in themselves provide a good estimate of what the effect is; but it is reasonable to expect that they will, when combined with the data from other countries that are being collected by the International Agency for Research on Cancer (Cardis and Martuzzi 1998) and the experience of a few more years.

References

Académie des Sciences 1997 Problems Associated with the Effects of Low Doses of Ionizing Radiation Rapport de l'Académie des Sciences No 38 (Paris: Technique and Documentation)

Cardis E and Martuzzi M 1998 Improving the estimates of radiation induced cancer risk Seguridad Nuclear 6 31--7

Cardis E et al 1995 Effects of low doses and low dose rates of external ionizing radiation: cancer mortality among nuclear industry workers in three countries Radiat. Res. 142 117--32

Dolphin G W 1976 A Comparison of the Observed and Expected Cancers of the Haematopoietic and Lymphatic Systems among Workers at Windscale NRPB Report R54 (Chilton: NRPB)

Royal Commission on Environmental Pollution 1976 Sixth Report: Nuclear Power and the Environment (London: HMSO)

Tubiana M 1998 The report of the French Academy of Sciences: `Problems associated with the effects of low doses of ionising radiation' J. Radiol. Prot. 18 243--8

Export citation and abstract BibTeX RIS

10.1088/0952-4746/19/1/001