Re: Radon and Smoking (individual vs aggregate)

["R. William Field" <bill-field@UIOWA.EDU>]

Dr. Cohen,



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.  To exclude these factors a

priori, as you have done, invalidates your test.  As I stated  before, you

can not adjust for these "perturbations" after the fact unless you know the

joint distribution of these factors for each county.



Bill



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

From: BERNARD L COHEN <blc+@pitt.edu>

To: Field, R. William <bill-field@UIOWA.EDU>

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

Sent: Wednesday, February 20, 2002 8:17 AM

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





>

> The problem is in how scientists handle theories and their tests

> of them. No theory takes into account every possible variation in

> circumstances. A basic theory takes into account the most important

> things, and the other factors are treated as "perturbations" or

> corrections to the basic theory. That is how I have managed my study. If

> you have a specific criticism of any of my work, I would be happy to

> address it; the best way to start is if you criticize something in my

> paper on treatment of confounding factors in an ecological study. Much

> better still, give a specific example of something that could nullify my

> results -- your example need not be true, but it must not be highly

> implausible.

>

> On Mon, 18 Feb 2002, Field, R. William wrote:

>

> > Dr. Cohen wrote earlier this morning, "If you don't understand that the

> > BEIR Reports use a liinear relationship between lung cancer and

radiation

> > exposure, there is no point in discussing these matters with you. What

do

> > you think the LNT debate is all about?"

> >

> >

> >

> > Dr. Cohen,

> >

> > I agree, it makes little sense to continue when we are in disagreement

over

> > basic facts about BEIR VI.  However, rather than conveying my

> > interpretation of BEIR VI, let me quote:

> >

> >  "Miner data clearly indicate that the relationship of lung-cancer

> > occurrence in a person, to cumulative exposure, is not simply linear,

and

> > that the joint relationship of radon and smoking is not additive" (Page

376

> > BEIR VI).

> >

> > Relative risks for cumulative radon exposure and lung cancer have

> > dependencies on other factors other then smoking and radon.  There are

> > vectors of covariates that affect background lung cancer rates and other

> > vectors that modify the exposure response relationship.

> >

> > Let's take just one example, if you hope to test the LNT theory using an

> > ecologic study, you must know the joint distribution of radon exposure

and

> > smoking in EACH county (Guthrie's paper I faxed you goes into detail on

how

> > to possibly get at this information).  Nonetheless, you state by using

> > "rigorous mathematics" in your 1995 paper you prove you can test the LNT

> > theory using an ecologic study.  You may have used math, but you needed

an

> > assumption that all smokers have the same smoking duration and intensity

to

> > obtain the Cohen derived formula.  This faulty assumption invalidates

your

> > derived equation.  This faulty assumption can not be later verified or

> > treated by after the fact stratifications using more averaged smoking

data

> > especially without knowing the joint distribution of the covariates.  In

> > fact, you have even previously stated, your analysis only,  "crudely

> > introduces the pack-year concept."  Dr. Gilbert has already provided

> > evidence that your results are confounded by smoking, because other

smoking

> > related cancers are also negatively associated with your county radon

data

> > (Perhaps Jim has a mechanism that shows alpha irradiation to the lung

> > decreases the cancer rates in other organs - if so please share it).  So

> > when you say, the LNT fails, what you are really proving is that the

Cohen

> > derived LNT formula that used faulty assumptions failed.  While I am not

a

> > loyal LNT supporter, I find myself in agreement with Dr. Piantadosi who

> > said in 1994, your findings do, "more to discredit the analysis than the

> > theory."  I would not expect you to be able to explain your inverse

> > findings by using your "treatments" as you call, since the Cohen derived

> > formula used to test the LNT suffers from faulty assumptions.

> >

> > Regards, 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/