Re: Radon Health Risks

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

Howard Long,



Your email below is really incomprehensible.



Comparing selection to sin??



The Iowa study selected lung cancer cases from the SEER registry with the 

requirement that they lived in their current home for at least 20 years.  

Nothing sinful about that.



The selection criteria to match cases to controls were by age (5 -year age 

groups), gender, and residency requirement.  They randomly selected 

controls, by 5 year age groups, from Drivers Lic. files and Medicare files. 

Again nothing sinful about this.  If they would have matched completely on 

smoking status and as Dr. Field has said, he does not think this is 

possible, it would no longer have been a population based study.  In the 

paper which I doubt you ever read, it says the results are most 

generalizable to women in the upper midwest.



By the way, age is an effect modifier.  Changes occur within the body with 

age, but age itself never killed anyone.  Correlation between age and death 

rate does not equal causality.  In fact, age adjusted lung cancers go down 

in those over 70 years of age.  Black women and men also have much higher 

lung cancer rates that white men and women.



It is impossible to have subjects equal in all aspects under study and if 

you did that would be over matching.

We have some space for an extra graduate student in epidemiology, are you 

interested?  Are you an M.D.?



Don Smith







>Selection is sin in epidemiology - and we all sin. Our rationalizations are 

>minimized by double-blind, placebo-controlled, prospective studies,

>repeated by disbelievers.

>

>However, such a definitive experiment is difficult with home radon. So, 

>Field matched controls for age (one cause of lung cancer) - but,

>statistically "adjusted" for smoking (another cause of lung cancer).  Only 

>32% of  Iowa "controls" smoked, vs 86% of cases. Such extrapolation, ie,

>extension beyond the data at hand, assumes applicability of data from other 

>studies, other populations. Extrapolation also derived LNT. One

>extrapolation in support of another stretches credibility.

>

>Also, selection of Iowa location (for whatever reasons) may disqualify 

>inference to other parts of the USA. Iowa women were a 1% outlier in 

>Cohen's

>mortality study. His variation within is small, with very short 1 SD bars 

>(18 at 1/4 pCi/l intervals),  and variation without is great (below 1 pCi/l

>vs above, or regression of same). The assumption of all epidemiology (and 

>case-control) is that despite universal uniqueness, residual unknowns can 

>be

>reasonably dismissed by having subjects virtually identical in all respects 

>except that under study. Infering to the whole from a part known to be

>different, is taboo - no matter how convenient.

>

>We need a definitive study.

>

>Howard Long (MD MPH epidemiology)

>

>Rad health wrote:

>





_________________________________________________________________

Chat with friends online, try MSN Messenger: http://messenger.msn.com



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

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/