Test of the Linear‐No Threshold Theory

I am writing in response to your Letter-to-the-Editor by A. V. Nero‘l) which states “ecological correlations should be taken with a large grain of salt, with the work of B. L. Cohen on indoor radon and lung cancer mortality representing a fine example of drawing conclusions that are recognizably unjustified because of confounding factors (in the case of lung cancer, smoking) with a substantial and uncontrolled for influence.” Our principal published papers on this ~tudy(~2~) treat the issue of smoking very thoroughly. In the main, they use the BEIR-IV approach which treats smokers and nonsmokers as completely separate species, with different and independent risks from radon (but they also show that our work is applicable to all other approaches). This procedure reduces the only possible problem to the determination of the fraction of the population of each county that are smokers. Six different and largely independent approaches to this determination are utilized, and all give essentially the same result. It is also shown(3) that our results are not nearly as sensitive to the smoking factor as Nero implies. Our papers clearly explain that our study is very different from all other ecological studies in that “the ecological fallacy” does not apply because it does not seek to determine the dose-response relationship, but simply is a test of the linear-no threshold dose-response relationship. They also show that all other limitations of ecological studies that have been discussed in the literature are not applicable. If Nero has an objection to our work, it should be presented as a paper or letter, rather than as a backhanded statement as the one quoted above.