Re: Cohen's Fallacy

["Rad health" <healthrad@HOTMAIL.COM>]

Ruth's second points:



>

>1.  How much radon a given population is exposed to.



What you need to know is how much radon is an individual exposed to.



>2.  How many lung cancers are observed in that population.



What you need to know is how many primary cancers are observed per year in a 

population.





>3.  IN THAT PARTICULAR POPULATION, how many smokers (ever-smokers) there 

>are

>and what is the average amount  (and distribution) of the smoking they did

>(e.g., years and packs per year).





You get this information in a case-control study, but not an ecologic study 

unless Dr. Cohen would follows Field's advice and collect this information.



>4. How many of the observed lung cancers occurred in smokers.



You know this in a case-control study but not an ecologic study.



>5.  Subtract the observed lung cancers in smokers from the total observed

>lung cancers.



This can be done only in a case control study.



>6.  Look at whatever relationship exists between radon exposure and the

>remaining lung cancers (those that occurred in non-smokers).   It is also

>necessary to look at the secondary smoke exposure in non-smokers, since

>secondary smoke is also a pretty well extablished carcinogen.



This ignores the interaction between smoking and radon.  You get alpha 

exposure from both radon and tobacco.

This part of what Field did and this is likely what will be done in a 

pooling.

>

>Maybe this is what Field et al did -- I will have to get their papers and

>see.  But it seems to me that applying some kind of statistical correction

>for smoking that says, in effect, since the national risk of lung cancer 

>from

>smoking is x, and the national (or statewide) fraction of smokers is y, 

>then

>x*y *observed lung cancers are attributable to smoking and the rest to 

>radon,

>doesn't really cut it.   All that would tell you is the fraction of lung

>cancers that MIGHT be attributed to smoking.





Regarding a smoking correction - As posted before on this list by others, 

this is the most powerful way to adjust for smoking.



Quoting Hosmer and Lemeshow's authoritative book, "Applied Logistic 

Regression":



"One generally considers a multivariate analysis for a more comprehensive 

modeling of the data.  One goal of such an analysis is to statistically 

adjust the estimated effects of each variable in the model for differences 

in the DISTRIBUTION of and associations among the other independent 

variables.  Applying this concept to a multivariate logistic regression

model, we may surmise that each estimated coefficient provides an estimate 

of the log odds [of lung cancer] adjusting for all other variables [smoking]

included in the model."





_________________________________________________________________

Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp.



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

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/