Re: absence of proof -Reply -Reply

[Mike McNaughton <mcnaught@lanl.gov>]

>I guess I'm a bit confused ('happens often)....  I thought radium-dial
>painter data showed a clear threshold of bone sarcoma incidence at
>10 Gy average skeletal Dose (Rowland 1997).  How does that agree
>with a linear model (logarithmic or normal plotting)?

First, let me state I am not a supporter of the linear-no-threshold model,
LNT; I believe Bernie Cohen's data demonstrates a definite departure from
LNT. The radium dial painter data, however, is often misinterpreted.

The data are graphed on page 198 of BEIR-V with a "curve to guide the eye".
Take a pencil and assume the following linear model for % tumor incidence
vs dose.
100 rad: 0.5%
1000 rad: 5%
10000 rad: 50%
This curve (which is an exponential on a linear-log graph) passes within or
close to the error bars of all data except the 20,000 rad point. Therefore,
the data are consistent with LNT up to 10,000 rad, but the curve flattens
between 10,000 and 20,000 rad.

In summary, the "curve to guide the eye" is misleading; an equally
acceptable fit is: LNT up to 10,000 rad with a departure from LNT near
20,000 rad.

mike
"Shlala gashle" (Zulu greeting, meaning "Stay safe")
mike (mcnaught@LANL.GOV)
************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html