Healthy survivor effect

[William McCabe <WMCCABE@tnrcc.state.tx.us>]

I often hear reference to the "healthy survivor effect" in response to
observations like the one Charles Willis made earlier about "the A-bomb
survivors have a lower cancer mortality rate than that of the rest of the
Japanese people (age-adjusted)."

My gut feeling is that this (the "healthy survivor effect") can't be a real (or
at least statistically significant) effect given the input
parameters/conditions and data to the analyses. 

Can anyone point me to articles/studies discussing the "healthy survivor
effect" and whether it is real or not?

Thanks in advance.

William J. McCabe
Health Physicist, MC-131
Texas Natural Resource Conservation Commission
P.O. Box 13087
Austin, Texas  78711-3087
wmccabe@tnrcc.state.tx.us
(512) 239-2252     fax: (512)239-6362

>>> Charles Willis <CAW@nrc.gov> 19 May 98  09:20 >>>
Ken,

Well done.  Writing an editorial of this quality is an accomplishment, but
getting published, that is an achievement.

Of course, it is almost a duty to do a little nit-picking.  Your closing
statement seems to under-state the regulatory importance of Federal
Guidance Reports.  Historically, these reports have gotten Presidential
endorsement that forces the other agencies to adopt the EPA guidance.

You say, quite correctly of course, that people over-estimate the risk
from low doses.  It might be added that people also overestimate the
risks from high doses.  People I have talked with find it difficult to believe
that the A-bomb survivors have a lower cancer mortality rate than that
of the rest of the Japanese people (age-adjusted).  The response is
even more skeptical when people are told that as of 1990, of the
A-bomb survivors with the highest doses (>200,000 mrem), only 17
percent had died of cancer.  

A question: Shouldn't more attention be paid to dose rate?  When the
NCRP tried to find data for or against a dose rate reduction factor (for
risk) they could find no effects for comparison where the dose rate
was below 8.3 rad per day.  Considering the biological repair process, it
seems reasonable that dose rate could be important, perhaps even
having a threshold.

Charlie Willis
caw@nrc.gov