Re: Radon and Smoking (other associations)

[Kai Kaletsch <info@eic.nu>]

It would be interesting to see graphical representations of other quantities

that are supposed to be related to county average radon levels. This would

include graphs of radon vs smoking, radon vs. SD in smoking, radon vs other

cancers, radon vs income, radon vs life expectancy, radon vs ethnic

background ....as well as radon vs lung cancer



Intuitively, I would guess that the smoothest curves with the steepest

slopes are the primary associations. Shallower, less well behaved curves

COULD be the result of residual confounding from the primary association.



> Dr. Gilbert pointed out, that that your results are confounded by smoking,

> because other smoking  related cancers are also negatively associated with

> your county radon data.



In that case, I would expect the radon vs smoking graph to be smoother and

steeper than the radon vs lc and radon vs other smoking  related cancers

graphs.



Are these kind of graphs available? (If it is an older paper I would

appreciate if someone could email me a pdf file.)



Kai



----- Original Message -----

From: "R. William Field" <bill-field@UIOWA.EDU>

To: "BERNARD L COHEN" <blc+@PITT.EDU>

Cc: <radsafe-digest@list.vanderbilt.edu>

Sent: Wednesday, February 20, 2002 11:49 AM

Subject: Re: Radon and Smoking (individual vs aggregate)





> Field wrote: I addressed a very specific problem below with your study.

You

> state that a

>  basic theory takes into account, "the most important things."   I

consider

> smoking intensity and duration two of the most important parameters to

> consider in validating your derived LNT formula.

>

> Dr. Cohen wrote: I am testing the BEIR formulas, and they do not include

> these

> things

>

> Field response:

>

> Dr. Cohen, the BEIR formulas DO indeed include "these things".  As we

> explained many years ago, HPJ 75(1), July 1998, page 13.    The BEIR model

> is quite general and allows for any degree of control for smoking.  You

can

> include factors such as pack-year rate and duration of smoking in addition

> to even non linear effects of smoking.  As was pointed out in BEIR IV,

"The

> choice of an appropriate age specific background rate for this calculation

> involves proper treatment of smoking, sex, and calendar time."

>

> Your previous response was that you "crudely" treat pack-year rate.  But

> your crude treatment is  after the cross-level bias already occurred.  My

> point which you have ignored for years is that you have not derived an

> equivalent BEIR model.  You can not assume smoking intensity and duration,

> are not important factors to include in your LNT derived formula.  Then

> latter try to treat the problem of smoking intensity and duration by using

> aggregate data.  You are not really testing the LNT using your formula.

> Your findings do not convince me the LNT fails, I am only convinced that

> your formula is not robust enough to test the LNT.  It is not surprising,

as

> Dr. Gilbert pointed out, that that your results are confounded by smoking,

> because other smoking  related cancers are also negatively associated with

> your county radon data.

>

> Bill Field

>

>

> ************************************************************************

> You are currently subscribed to the Radsafe mailing list. To unsubscribe,

> send an e-mail to Majordomo@list.vanderbilt.edu  Put the text "unsubscribe

> radsafe" (no quote marks) in the body of the e-mail, with no subject line.

> You can view the Radsafe archives at http://www.vanderbilt.edu/radsafe/

>



************************************************************************

You are currently subscribed to the Radsafe mailing list. To unsubscribe,

send an e-mail to Majordomo@list.vanderbilt.edu  Put the text "unsubscribe

radsafe" (no quote marks) in the body of the e-mail, with no subject line.

You can view the Radsafe archives at http://www.vanderbilt.edu/radsafe/