In his News and Views article “A dose of the bomb” (Nature 424, 495–496; 2003) about the impact of the new A-bomb dosimetry on radiation-risk estimates, Mark Little argues that the risks of γ-ray exposure should become more reliable. Two other factors suggest that the new dosimetry is likely to have little or no impact on risk estimates for doses that are occupationally or environmentally relevant.

The US National Council on Radiation Protection and Measurements (NCRP Report No. 126, NCRP, Bethesda, 1997) has determined the dose-rate effectiveness factor to be the main source of uncertainty in risk estimates, accounting for about 40% of total uncertainty. Overall, dosimetry uncertainties account for less than 10% of the total.

Dose-rate effectiveness factors must be applied to A-bomb-derived risk estimates to account for differences in temporal dose delivery. The bomb exposed people to very high dose rates (essentially instantaneous exposure), whereas risk estimates are usually applied to occupational or environmental situations where the dose is delivered at a significantly lower rate and risks are assumed to be lower because of biological repair mechanisms.

The second source of uncertainty is dose extrapolation. A-bomb risks are based on excess cancer deaths in people exposed to doses greater than 200 mSv (D. A. Pierce et al. Radiat. Res. 146, 1–27; 1996). Below this dose the total excess number of cancer deaths is too small to be used reliably in risk estimation. Accordingly, risks calculated from high-dose data are extrapolated using the linear non-threshold theory to predict risks at small doses. Assuming that occupational and environmental doses are around 2 mSv per year, dose extrapolation is significant and the theoretically derived risks are highly uncertain.