The BEIR Report Relative Risk and Absolute Risk Models For Estimating Effects of Low Level Radiation

Abstract

Two tests of the validity of the BEIR Report relative risk model as an ad hoc assumption are applied and the model fails badly whereas the absolute risk model is successful. The original basis for the relative risk model is explored and is questionable. Reasonable derivations of the relative risk model lead to an exponential rather than a linear dose-effect relationship, and if the former is used to analyze effects of low level radiation, these effects are reduced by a factor of .apprx. 3. A multiple-factor model is introduced and is mathematically equivalent to a relative risk model with the percent/rem varying inversely with age. It leads to values about 50% higher than the BEIR absolute risk model and rather unpalatable modifications are requred to explain the observed cancer incidence in those exposed as children. Even it these modifications are accepted, the multiple factor model gives much smaller results than the original BEIR relative risk model. The BEIR absolute risk model has no difficulty in explaining all data with the simple assumption that the ability to resist cancer decreases with increasing age. The results are insensitive to the form of the age dependence.