Re: It's the design, folks, it's the design

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

On Tue, 4 Mar 1997, Daniel J Strom wrote:

> Cohen's data are a good deal better than the study described above, but 
> the fundamental problem with any ecological study is the great leaps of 
> faith that are necessary in positing that one set of numbers is related 
> to another.  In Cohen's study, most of the cigarette sales data probably 
> don't apply to most of the lung cancer death data because different 
> people smoked the cigarettes sold in a county than died of lung cancer 
> in a county.

	---This does not matter in a test of LNT; it does matter in
determining a dose - response relationship, but that is not what my papers
are about. The only smoking data required for my test of LNT is
the fraction of the adult population that are smokers, S. I used several
different approaches to determining S-values, and cigarette sales was not
the primary one. I then assumed that my S-values were wrong and
investigated what S-values  would be required to resolve my
discrepancy, and found that the requirements are entirely implausible.
Unless you contest my judgement on their implausibility, you have no case
whatsoever on uncertainty of S-values as an explanation.

> Most of the radon measurements probably don't apply to 
> most of the lung cancer death data because different people breathed the 
> radon in the county than died of lung cancer in the county.  

	---This does not matter in my test of LNT; it does matter in
deriving a dose -response relationship, but that is not what my papers are
about. In LNT, the number of deaths depends only on the total dose
(person-rem), independently of how this dose is distributed.This is surely
familiar to all health physicists.

> Furthermore, the diet, lifestyle, genetic background, and other 
> confounders are not controlled, and cannot be controlled, because of the 
> lack of individual data.  

	---These correlate very strongly with many of the factors that
were considered, like socioeconomic factors, geography, climate, etc. What
I did was something much more elaborate than "control for" these
variables. Incidently, can you point out how the case-control studies
control for the factors you cite?

> This non-predictive relationship between various kinds of data isn't a 
> "result" worth $5000, it isn't news to those who have taken epidemiology 
> 101, and it is unlikely to find its way into a peer-reviewed journal, 
> since all of the reviewers would yawn and say, "No original content - 
> reject." 

	---My paper was not trying to determine a "predictive
relationship". LNT did the predicting; I was just testing its prediction.
In my papers, I cover all evidence presented in epidemiology 101 and in
all other literature published on ecological studies, and show that it is
irrelevant to my study. If this coverage is incorrect, someone should say
why. That would surely be a publishable letter-to-the-editor.
	If my papers are so obviously wrong, how come I have had invited
papers on it at national meetings of many different scientific societies,
and in many other places  in U.S. and in France, Austria, Netherlands,
Israel, Australia, Japan, Canada, and Czech Republic (in nearly all cases
with travel expenses paid), and how come I got a prestigious award
(including $1000) from a national scientific society for that work.

>  And contrary to Cohen's claim in the March Health Physics, 
> the miner study measurements were mostly made in the same mines in which 
> each miner worked while they were working (with some exceptions), not 
> after the miners died.

	---My statement in March Health Physics was referring to the
studies of radon in homes, not to the miner studies. It is difficult for
me to see how my statement could be misinterpreted, but if I wasn't clear
on this, I apologize.


>  For both case-control and miner studies, smoking 
> data (when available) were for the individuals involved, not for unknown 
> persons.

	---As pointed out above, this doesn't matter in my test of LNT

> As for the recurrent question of hypothesis testing, I agree that the 
> linear, nonthreshold model predicts the excess relative risk lung cancer 
> deaths in groups of individuals to be proportional to lifetime exposures 
> to radon progeny in those individuals.  If one knew what the 
> cradle-to-grave radon progeny exposures were for each individual in each 
> group, along with each individual's smoking, diet, lifestyle, genetic 
> predisposition, exposures to other air pollutants, and other risk 
> factors for lung cancer, one would surely be able to test the 
> hypothesis.  That's not what Cohen is doing.  When one measure Bob's 
> radon, Joe's smoking, and Sam's lung cancer and tries to correlate them, 
> one encounters inferential difficulties.  Some people who smoked 
> cigarettes purchased in County X also smoked cigarettes purchased in 
> Counties Y, Z, A, B, and C or in the Post Excange or the College 
> Bookstore or at the Casino or...  Some people who were exposed to radon 
> in County D were also exposed to radon in Counties E, F, G, and H.  Some 
> of those people died in County J.  Some people who died of lung cancer 
> in County X didn't smoke any of the cigarettes sold in County X and were 
> never exposed to radon in County X.  For my money, this isn't a very 
> cogent way to test a hypothesis.

	---All of these problems were addressed in my paper, and many of
them above in this message (especially smoking). Migration is discussed
extensively in my papers.   
	How about continuing this discourse by you writing a
letter-to-the-editor or a counter article for a journal?

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