Rants, again

[JMUCKERHEIDE@delphi.com]

Paul, et al

Paul Frame wrote, 3/15.  (Sorry about the delay.)
> 
> Might it be possible to narrow the focus and length of these
> discussions by considering individual studies or specific pieces of
> evidence? This would permit a more informative "scientific" exchange
> of information. For example, shoot down the IARC study of 96,000
> radiation workers (in Canada, the US and Great Britain) that
> indicated excess risks of leukemia. It was in Vol 344 of the Lancet,
> Oct 15, 1994 and was referred to by Nuclear News as the "Best
> estimates of low-level radiation risks". Radsafers could then learn
> the details of an important new (and easy to obtain) paper. 

Finally caught up with this "important new paper", and found:  nothing.

There is no data by dose (with 2 instances where it says data will be
published "elsewhere"). There are no SMRs of other measures of health effects. 
 

The closest to a data point is the statement that the RR for leukemia (ex CLL) 
is 1.22 and for all-cancer ex leukemia is 0.99 for 10 Rem vs the 0 Rem; and
that the ERR is 2.2/Sv ( 0.1 to 5.7 @ _90% CI_ [implying not significant at
95%! ] ) for leukemia, and -0.07/Sv ( -0.39 to 0.30 @ 90% CI ).  

It then adds, with no foundation, that they multiplied doses by 1.2 for
leukemia on the (new, unsupported) judgement of the dosimetry that bone marrow 
dose is underestimated, making leukemia 2.6 ( 0.1 to 7.0 @ 90% -- still not
95%! ) 

It then states the ERR for leukemia is between the 3.7/Sv for atomic bomb
survivors with a linear model, and 1.4/Sv from the linear-quadratic model,
_extrapolations_. Fascinating rationalizations, since the actual bomb survivor 
data is positive for leukemia and null to slightly positive for all cancers
except leukemia, in the low dose range of interest here (yet the
"extrapolations" from >200 Rem and >400 Rem are applied, that themselves
ignore the data at <40 Rem). 

The article says the data for leukemia compared to RERF/BEIR V model finds
that a DDREF would be ~1.7 (lower limit of 0.6 and upper limit of 37!!) 

This, of course, as usual, presumes the answer (linear/linear quadratic
"model") and then just "fits" the data to it, without any consideration of the 
data, or the validity of the "model" to the data. 

Of course, only 16.5% of the population is deceased (15,825/95,673) with 3830
from all cancer ex leukemia, and 119 from leukemia (ex CLL) (ignoring 6 in
cases exposed to > 40 Rem, but not stating how many total >40 Rem cases
exist).  Without any age or SMR data these numbers can't be compared, though
at <25% cancer deaths and <0.8% leukemia deaths the implication is .... !! 
(unless the deaths are very skewed to very young ages!) 

Typical of such reports, this kind of article enables references to
"conclusions" from obfuscated data with no opportunity to review the basis.
Then when the data is released (IF the form is accessible, which is often NOT
the case) and a critical review is possible, it is relatively easy to ignore
-- the answer has already been announced! and can be a continuation of
self-referential results to meet the mandate. 

The author of the NN article confirms that there was no independent technical
review in his report. The headline is just a self-congratulatory statement
from within the IARC article itself. (This was a news report, not a technical
review.) 

Of course, since each of the significant data sources compiled here have been
previously reviewed and found to have nul or negative correlations with dose
when actual dose-response data is applied, the result presented here is
impossible, except by the standard government mandate to presume the linear
solution and ignore the actual data, as has been done previously with the
government reports of the individual cohorts, similarly obfuscating this data. 

Since I don't have library access to the Lancet, and can just request the
original article without more precise references to letters, I would
appreciate anyone else's input on whether there were any follow-on letters to
this article in subsequent issues of The Lancet. 

> Then we could turn our attention to Cohen's latest report.

I think that's a great idea!  He's pretty well destroyed the nonsense that
"ecologic studies" don't/can't have scientific merit! (This has always been a
funny/disingenuous argument since the same people who make it then argue they
can find environmental carcinogens from local releases/sources that have 1000s 
of times less statistical validity from population release/dose
considerations!) 

> Best wishes
> 
> Paul Frame

I'll provide some other scientific data to better demonstrate the actual
evidence of the LACK of health effects associated with low to moderate
radiation doses to avoid "generalizations". 

It is encouraging that HPS leadership is taking a formal and serious review of 
this issue! As Dr. Goldman stated in the Feb 95 HPS Newsletter message, are we 
really going to spend/waste $ Trillion to clean up sites with negligible
public health benefit? Is this really justified in the light of national
public health/environmental needs that would benefit the public instead of the 
government bureaucracy? 

Of course we can reach this conclusion even assuming the linear no-threshold
"model" if the evaluation cares more about the potential significance to
public health rather than to government programs. But BEIR, NCRP, etc make
those statements out of the spotlight, leaving the agencies to abuse even the
linear model into causing massive public costs for no public benefits. 

Along with HPS leadership, we all should be concerned, and technically
challenged to contribute to public policy and understanding about such an
extreme waste of the public wealth (forced from private expenditures as well
as government funds), and make that case to the appropriate policy interests
for corrective action. (I would like to think that the Clinton Admin
willingness to defer DOEs massive cleanup costs in the priorities for
government funding has been the result of growing understanding of the
fallacies that have been generated by the abuse of the linear no-threshold
model to create government programs and regulations.) 

Perhaps part of the answer is to focus a little less on what we don't know
about low level radiation health effects, and reexamine and report on the data 
we do have about the levels of exposure that produce known health effects! 

Thanks.

Regards, Jim Muckerheide