Radiation protection approach to assessing population risk for threshold-type radiobiological effects

The potential harm to a population exposed to radiation from a nuclear accident, such as the one that recently occurred at the Chernobyl plant in the Soviet Union, is of concern to many individuals. The average dose to a population is a useful index of harm (IH) only for linear, nonthreshold-type, quantal (i.e., all-or-none) effects. For such radiobiological effects, the expected harm to the population is linearly related to the average dose. However, for nonstochastic effects, it is not. An IH is proposed for threshold-type nonstochastic effects which is based on a form of the Weibull model where, at low to moderate doses, the individual risk at dose X = D/D50 is given by the approximation Risk = ln(2)XV; where D50 is the absorbed radiation dose that produces the specified effect in one-half the population, D is the absorbed radiation dose, and V is a positive parameter. The dose, X, is in units of the D50. Use of this form of the Weibull model is limited to doses such that X is small in comparison to 1. An IH for the population can be obtained by defining a new variable, P = XV, in dimensionless units, because the individual risk is linearly related to P at low and moderate doses. The average value for P (given by [P]) for an exposed population can be used as an IH for the population when the maximum value for P does not exceed 1. Both P and [P] can be regarded as theoretical doses. The average risk for the population in terms of the average dose [P] is given by ln(2)*[P], and the expected cases of nonstochastic effects among N individuals by N*ln(2)*[P]. As an example of the application of the average dose [P], the expected cases of temporary sterility in males among the approximately 135,000 people evacuated within 30 km of the Chernobyl plant is calculated to be about 200