Re: Radon Exposure Assessment for Cases

[Kai Kaletsch <info@eic.nu>]

Don,



> First, the average concentration does not equal the average exposure.



I never said that it did.



> The Iowa case control

> study has nothing what so ever to do with the LNT.



Let me quote from the study: "The associations between lung cancer risk and

observed radon exposures were studied by using LINEAR excess odds models..."

and the equation used, equation (2), does not show a threshold.



> Yes it does matter when the categories were chosen.  They could be

> chosen after the fact to either maximize or minimize the association by

> arbitrarily choosing endpoints that would support a specific finding.



It matters if you want to challenge the integrity of the authors, which is

something I am NOT doing. Just because the intervals were chosen a priory,

is not a guarantee that they do not produce non representative results. Let

me quote from the study: "Analyses based on continuous exposure variables

have the advantage of avoiding the discretionary nature of choosing

cutpoints."



> By using mean radon concentration data to predict risk you are

> falling into the ecologic fallacy by assuming grouped findings = risk to

an

> individual.



This is not a fallacy if you use LNT (your equation 2) as a statistical

model and if you match cases to controls according to likely confounders.

Anyway, I am not predicting risk. I am expressing surprise that the cases

had smaller radon source terms than the controls and the only reason they

had higher radon readings in the living areas is because of reduced

ventilation. If there is a distribution of radon source terms across Iowa, I

would have expected some of the cases to be the result of houses that had

elevated radon levels in the living areas because of a high radon source

term. Everything else being equal, that should translate into higher average

basement readings. (BTW, the Iowa study uses "grouped findings" (the groups

in the 5 exposure categories) to infer the risk to an individual. I think

you would have a pretty tough sell, if you tried to convince someone that

some agent follows a positive LNT, if your controls had higher average

exposure to the agent than your cases.)



> The cause you are concerned about is likely as simple as these cases

> developed lung cancer because the radon level was higher in their living

> area as compared to controls.



That seems to be the conclusion of the study, but it doesn't answer any of

the questions that I raised. (Why do houses of cases have such poor

ventilation as compared to controls? Is it reasonable to assume that the

ventilation in case houses is the same before as after lung cancer

diagnosis. Why do people who had high radon concentrations in the living

area AND high radon concentration in the basement not get lung cancer?)



It would be interesting to re-analyze the data and test the ratio of main

floor to basement radon readings (lack of ventilation) as the causative

factor.



Kai



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

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/