Re: Cohen's Observation

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

On Sun, 15 Jun 2003 epirad@mchsi.com wrote:

> Dr. Cohen stated, he did not use the SEER data because it is not available for

> most counties, because I don't understand why it is better than mortality data.

>

> Response -

>

> The SEER Program is the only comprehensive source of population-based

> information in the United States that includes stage of cancer at the time of

> diagnosis and survival rates within each stage. SEE:

> http://www.seer.cancer.gov/about/

> Data is easily available using SEER stat which I previously sent Dr. Cohen

> information about.   http://www.seer.cancer.gov/seerstat/



	--It would take a lot of time and trouble for me to get the SEER

data into my data bases, and I don't think the results of that effort

would be publishable. If I can be assured that the results will be

publishable, I will undertake that effort. I still don't understand why

the SEER data are better for my purposes, and I don't see how I would use

info on stage of cancer at time of diagnosis and survival rates at each

stage. It is my understanding that lung cancer is nearly always fatal.



> Quantity of data should not be as important as quality of data.



	--It is important to have counties from all over the U.S. as that

reduces the potential for confounding. For example, in a single state, it

is easily possible that an ethnic group that is less susceptible to lung

cancer lives in a region with high radon levels. When all states are

included, with numerous areas of high and low radon and numerous

ethnic groups in populations, this would be extremely improbable.



 In many

> state cancer registries, lung cancer is reported even if it is a secondary

> cancer.



	--I did not know this. I sent an inquiry to CDC asking what

fraction of reported lung cancers are primary.



> Within your references, you also state that within your data –

>

> Low-income families are grossly underrepresented.

> High rise apartments are not represented.

> People concerned with environmental issues are overrepresented

> Urban areas are greatly underrepresented

> Cigarette smokers (and those socioeconomic factors associated with smoking)

> are

>           underrepresented

> Rented homes are grossly under represented



	--These problems are considered in several of our papers. The

principal reason for my confidence in our radon levels is the agreement

with EPA studies and with studies sponsored by various states



> You also indicated that 27% of bedrooms in your data set were located in the

> basements.



	--I am quite sure this is not correct. If you insist, I will try

to find my papers on this.



> When we replicate Cohen’s regression of Riggan’s lung cancer mortality data on

> Cohen’s radon measurements, we get the same large negative associations as

> reported by Cohen.  However, when we use the more temporally correct Iowa SEER

> lung cancer incidence rates and regress them on Cohen’s radon measurements and

> smoking percentages the large negative correlations go away.



	--See my above example of how info on an individual state can be

affected by confounding factors



>  when Cohen's adjusted smoking percentages for males and females were

> regressed on radon levels, significant (p < 0.00001) negative associations

> between smoking and radon were noted for both males and females.



	--Why is this a problem? I reduces the discrepancy with LNT. If it

were not so, the discrepancy would be greater.



> when we (Smith et al. 1998) repeated the regression of lung cancer mortality

> rates on Cohen's adjusted smoking percentages, the resulting R2 values

> indicated that Cohen's smoking summary data explained very little (23.7% for

> females; 34.5% for males) of the variation in lung cancer mortality rates.



	--R-squared reflects statistical fluctuations; see my paper in

Health Physics 72:622-627;1997. This subject is treated in more detail in

Item #15 on my web site -- see discussion around Eqn.(5)



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

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/