[ RadSafe ] BEIR VII

[Dale Boyce]

Hello again,

Actually the mean altitude of Hawaii is fairly high, although I will concede 
the population is mainly at low altitude. This is true of a few other states 
as well.  I tried to find a quick reference to the natural radioactivity in 
Hawaii. Being volcanic it is probably fairly high as well.

Regardless of whether the background is high there, my point is that any 
effect of radiation appears to be washed out by other factors. They don't 
have a lot of heavy industry for example. It is a correlation or rather an 
anti-correlation, and not a one to one correspondence. Any dose effect due 
to background changes is washed out by other factors. I am not even claiming 
it is heavy industry, just using it as a possible example.

For the moment let's say BEIR VII is correct and we use the canonical 3.6 
mSv/yr background, then over an average lifetime if you don't apply DRR 
factors almost 10% of cancers are due to background radiation. If someone 
lives in a high background area say 10 mSv/year, then even applying a DRR of 
1.5 puts radiation as the cause of 20% of cancers in those areas.

What we see tends to be a 20% decrease instead of increase. Again I am not 
calling what I am writing anything more than a casual obsevation that BEIR 
VII predicts that a significant fraction of cancer and cancer deaths are due 
to background radiation. If it is significant it should be measurable.

There are three (maybe more, but these are the ones I tend to think of ;) 
uses for dose response modeling:

One.  Establishing acceptable exposure levels for the purposes of 
regulation. I'm okay with LNT on that.

Two. Demonstrating probabilty of causation in litigation. Since LNT is used 
in establishing regulations, I'm okay with its use in this. Companies and 
institutions set their ALARA policy based on the risk they are willing to 
accept.

Three.  Estimating excess cancer deaths in large populations exposed to low 
dose radiation. This is just plain wrong. Extrapolating a model outside the 
domain of statistically significant observations for this purpose should 
only be used to design experiments to actually measure and extend the 
domain. If that can't be done so be it.

I don't really have a problem with the risk factors presented. It is 
presenting them in a way that invites the use in the third way without a 
disclaimer that the numbers really are not valid when used in this way.

Dale
daleboyce at charter.net




----- Original Message ----- 
From: "Otto G. Raabe" <ograabe at ucdavis.edu>
To: "Dale Boyce" <daleboyce at charter.net>; <radsafe at radlab.nl>
Sent: Thursday, June 30, 2005 5:54 PM
Subject: Re: [ RadSafe ] BEIR VII


> At 02:41 PM 6/30/2005, Dale Boyce wrote:
>>I periodically like to point out when discussions like these arise that if 
>>you take American Cancer Society data on cancer death rates by state and 
>>plot them versus the mean altitude of the state there is a strong 
>>anti-correlation.  That is the higher you live (and therefore the higher 
>>your probable background exposure) the lower your risk of dying of cancer.
> ***********************************************
> Hawaii doesn't fit very well since it is low in altitude and low in 
> cancer.
>
> Otto
>
> **********************************************
> Prof. Otto G. Raabe, Ph.D., CHP
> Center for Health & the Environment
> (Street Address: Bldg. 3792, Old Davis Road)
> University of California, Davis, CA 95616
> E-Mail: ograabe at ucdavis.edu
> Phone: (530) 752-7754   FAX: (530) 758-6140
> ***********************************************