RE: LNT models -Radon/smoking

[Bernard L Cohen <blc+@pitt.edu>]

On Thu, 9 Mar 2000, Bradshaw, Keith wrote:
> 
> 	If the major cause of lung cancer is smoking, it's inevitable that
> the uncertainty in lung cancer yield from this causes a great deal of
> statistical noise in the smaller effect that you are trying to measure, i.e.
> lung cancer yield from radon in homes.  So why not discount all households
> that have one or more resident smokers (including people who have given up
> in the last 20 or 30 years).  I realise  this will wipe out the majority of
> the data set but it might be better to do this than attempt to correct for
> it, especially since you appear to have a large study.

	My study starts with two basic pieces of information, lung cancer
rates and average radon level in each county. The important effect of
smoking is in affecting the lung cancer rates, and your suggestion would
do nothing for simplifying that problem; lung cancer rates are taken from
official statistics over which I have no control.
	Following your suggestion  would make a slight change in
radon levels if radon levels were different for smokers and non-smokers in
the same county, but this possibility has been investigated thoroughly in
my papers and it does very little to change the results. There is also the
problem that our radon data comes partly from EPA and State-sponsored
measurements which have no smoking info attached.
	Actually, the relative importance of radon levels, r, and smoking
prevalence, S, in determining our results is not that much different.
Mostly this is because variations in S among U.S. counties (Std Dev =13%
of mean) is much less than variations in r (Std Dev = 58% of mean). All
things considered, the importance of S is less than twice the
importance of r in determining variations in lung cancer rates in U.S.
counties if the BEIR-IV model is used. These matters are treated in my
paper in Health Physics 72:623-628;1997.

> 
Bernard L. Cohen
Physics Dept.
University of Pittsburgh
Pittsburgh, PA 15260
Tel: (412)624-9245
Fax: (412)624-9163
e-mail: blc+@pitt.edu


************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html