Re: LNT model question

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

On Wed, 8 Mar 2000 FIELDRW@aol.com wrote:

>  
> Field Response: As we suggested before (see references below), the risk model 
> you derive from the BEIR IV model is not equivalent to the BEIR IV risk 
> model.  You attempted to equate the model to the BEIR IV model by applying 
> rigid assumptions that include: 1) The time spent in the home is the same for 
> all individuals; 2) radon exposures occurring outside the home do not 
> contribute to lung cancer risk; 3) baseline lung cancer mortality is solely a 
> function of age and smoking; 4) all smokers have the same increased risk of 
> mortality; 5) smoking duration and intensity is the same for everyone; 6) age 
> is the only modifying effect of radon exposure; and 7) radon concentration is 
> directly proportional to the delivered dose from the radon progeny.

	--Anyone with experience in mathematical derivations will
recognize that it would be impossible to take every little possible
complicating detail into account in a mathematical derivation. All of
these issues have been addressed elsewhere in my publications. Items 1, 2,
and 7 were addressed in Table 1 and associated discussion in my paper in
Health Physics 75:23-28;1998. Item 4 is the assumption of BEIR-IV that I
was starting with. Item 5 is most fully analyzed in two forthcoming papers
accepted for publication in Health Physics and J. Radiol Prot. These also
cover Item 4. Items 3 and 6 are covered by my responses to the other
items. 
	Can you explain how these problems are handled in case-control
studies? 


> Your 
> relative risk model is only valid if these assumptions are not violated.  You 
> previously have  attempted in your published manuscripts to support your own 
> primary assumptions by use of secondary assumptions or qualifying statements 
> like: 1) "On average, people spend 75% of their time in the home…"; 2) "… 
> average radon levels are generally much higher in homes …"; 3) "….. 
> workplaces of various counties would strongly tend to be proportional ….." ; 
> 4) "….. it is generally agreed that radon gas level is a much better 
> indicator than WLM of radiation dose in homes …."; and 5) "This crudely 
> introduces the pack-years concept".  The very foundation of your assumptions 
> that you use to equate your risk model to the BEIR IV Risk Model is based 
> further on poorly documented assumptions.  The assumptions all have both an 
> error associated with them and a very significant NON-LINEAR component, which 
> as previously pointed out, can not be quantitatively described.  The BEIR IV 
> model, unlike your model, is not confined by either your primary or secondary 
> assumptions.     
>
	--This came out somewhat garbled, but as well as I can understand
them, these issues are coverd in my above citations

> -----------------------------------------------------------
> Dr. Cohen wrote - "This correlation is favorable for verifying LNT. If 
> smoking  rates were positively correlated, my correction for smoking would 
> have  increased the discrepancy with LNT. The negative correlation I adopted  
> reduced the discrepancy. The effects of various assumed correlations  between 
> smoking prevalence and mean radon levels are presented in my 1995  paper. If 
> the negative correlation were perfect, the discrepancy with LNT would have 
> been cut in half."  
> 
> Field Response:  My question regarded why you felt you found such a strong 
> inverse association between mean county radon concentrations and your 
> estimated county smoking rates.  I do not think this can be explained merely 
> by an urban/rural phenomenon. 

	--I think it can be explained by the urban-rural effect, but the
explanation is irrelevant to the issue under discussion, which is finding
an explanation for the large discrepancy between predictions of LNT and
our data.

> It would be interesting to see if this inverse 
> association held for living area measurements, rather than basement 
> measurements.

	--My analyses include both living area and basement measurements,
and the results are similar for both.

> 
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