Re: Cohen's ecologic study reanalyses

[BERNARD L COHEN <blc+@PITT.EDU>]

On Tue, 16 Sep 2003, Gary Howard wrote:

>

> What are your thoughts on this recent paper?

>

> Health Physics, Volume 85, Issue 4

>

> EPIDEMIOLOGICAL ASSOCIATIONS AMONG LUNG CANCER, RADON EXPOSURE AND

> ELEVATION ABOVE SEA LEVEL—A REASSESSMENT OF COHEN'S COUNTY LEVEL RADON

> STUDY

>

> Wesley R. Van Pelt*



	--I am preparing a letter to the Editor on this. Part of my first

draft follows:



The title of Wes Van Pelt's recent paper (Van Pelt 2003) calls it "a

reassessment of Cohen's county level radon study. The purpose of this

letter is to show that it is something less than that.

	My studies of the relationship between lung cancer rates and

average radon levels in homes for 1600 U.S. counties (Cohen 1995, 2000a)

with elaborate treatments of smoking behaviors (Cohen 1995, 1998, 2000b)

was designed and presented as a test of the linear-no threshold theory

(LNT) of radiation induced cancer. The key quantity in my analyses is the

slope, B, of the best fit to the data. In units of percent change in lung

cancer rate per pCi/L (37 Bq/m3) of radon, LNT predicts B = +7.3, whereas

fitting the data gives B = -7.3 (0.6) for males and B = -8.3 (0.8) for

females, discrepant with the LNT prediction by mopre than 20 standard

deviations.

	My studies include elaborate treatments of potential confounding

factors (CF) that might explain these discrepancies. I introduced a method

of stratification for obtaining a value of B free of confounding by a

given CF and showed that combinations of CF are not much more effective

than the single most important CF in affecting the results, including a

recent advanced treatment (Cohen 2004).

	Van Pelt has now discovered a new CF, altitude (elevation above

sea level) which can do more to make the value of B less negative than any

of the hundreds of CF previously considered. In particular, by use of

stratification on altitude, he finds B = -3.3 for males and B = -4.3 for

females. Moreover, he offers a mechanism involving changes in oxygen

concentration that can explain the behavior he hypothesizes.

	While Van Pelt's observation is certainly interesting, it does

little to affect the conclusions of my papers. In the first place, his

results can reduce the discrepancy in B-values with the LNT prediction, B

= +7.3, by only 27% for males (from [7.3 +7.3] to [7.3 + 3.3]) and by only

26% for females (from [7.3 + 8.3] to [7.3 + 4.3], which is still a long

way from resolving the discrepancy.

	In the second place, the stratification method provides only an

upper limit on how important a CF may be. For example, since radon levels

are strongly correlated with radon levels, two alternative views are

suggested:

A.	The negative correlation between lung cancer and radon might be

partially caused by a negative correlation between lung cancer and

altitude, as Van Pelt proposes; or

B.	 The negative correlation between lung cancer and altitude might

be explained by the negative correlation between lung cancer and radon, in

which case altitude is not an effective CF.

I see no reason to prefer alternative A over alternative B.



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

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/