After any serious nuclear incident, there are liable to be predictions of so many thousand cancers or deaths as a result of radiation doses to the public. Concern is justified, but such claims show a complete if pardonable misunderstanding of the situation.
Radiation and contamination
There are two kinds of hazard due to radioactive material. One is an escape from containment and attachment to skin or other surfaces, that is contamination, posing a continuing risk to the health of people who may not be otherwise involved. The other is radiation from a source that once removed or avoided adds nothing to the risk already incurred. There is also an intermediate case of irradiation with neutrons that can induce radioactivity in previously stable materials, but this cannot normally happen with alpha, beta or gamma radiation which outside specialised facilities are the only forms usually met.
High radiation doses can certainly cause cancer, or in extreme cases almost immediate death. However, everyone receives low doses in the normal course of life. Except for workers occupationally exposed, most comes from purely natural sources (including a tenth of it in Britain from the potassium that forms an essential part of our own bodies), and all but a minute proportion of the rest from medical procedures. Outside the vicinity of an accident, any medical cases due to it will be indistinguishable from those occurring naturally and probably in much greater numbers, so cannot be counted. Only estimates can be made, based on probabilities, and the basis of calculating them is disputed.
Risk can be directly related to dose only at rather high values under unusual conditions, and then only roughly since there are complications. At zero dose, were it possible, the risk due to it would also be zero. What happens in between is unknown.
Calculations for the purpose of precautionary regulation assume a direct proportionality, that is a straight line as in Fig. 1a. Doing so ignores some mitigating factors, such as that a dose is less dangerous if spread out in time than when given at once, and so is held to be pessimistic; however, without this or even more dubious assumptions, no estimate would be possible at all.
Other relationships between dose and risk have been proposed, such as in Fig. 1b or even Fig. 1c which shows a slight benefit from low doses, on the analogy of promoting resistance to bacterial infection. Such relationships imply a threshold below which risk is virtually nil, and there are some grounds for belief that this may be so. Since most of the doses concerned are at the low end of the scale, it implies that casualties might be very few.
The important point is that published figures, purporting to predict how many victims to expect, are in fact what might perhaps happen at worst. The actual outcome, if it could be determined, might be anything less down to practically zero.