ABSTRACT
But the independent U.S. scientific body (Committee on the Biological Effects of Radiation) in its third report does accept that a simple linear extrapolation to low doses is unsupported by the facts; there is theory and evidence to suppose that a non-linear second-order fit (the linear-quadratic model), which starts less steeply and rises in effect more steeply with increasing dose, is better than a simple linear fit. The argument has been accepted by the ICRP and leads to the conclusion that low doses in routine emissions, or spread in accidents, will have less carcinogenic and genetic effect than previously assessed. The abandonment of the linear model does make the estimate of stochastic effects of radiation on a population considerably more difficult since the absolute level to the individual is now important and not just the integrated dose over the population. This is probably a sufficient practical reason why the linear and conservative model will still be favoured in protective and health physics calculations.
