[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Ecological Study- I Need a Primer!
I highly recommend Laurier, et al, "Radon Exposure and the Risk of
Leukemia: A Review of Epidemiological Studies", Health Physics,V81,#3,
Sept. 2001. This is one of the better papers I've seen in terms of
describing different kinds of epi studies, including ecological, cohort
(prospective) and case/control studies, with some discussion of the
utility and pitfalls of each type. It also shows pretty clearly how
ecologic studies can come to conclusions that are wrong.
The way I learned it in Public Health school, ecological studies are
useful for hypothesis generation, and next to useless for proving
anything. Case control studies are better, and cohort studies are
generally regarded as the best evidence (not proof) of a cause/effect
relationship. "Proof" comes from may sources, including reproducible
studies, and must include a mechanism.
Few epi studies can "prove" a mechanism, and fewer try to. This
doesn't make such studies useless. The textbook example is John Snow's
work with cholera in London. By spotting cases of disease on a map he
showed pretty clearly that there was a relationship with disease and
drinking water. This was before the germ theory, so he had no idea what
the mechanism was. Nevertheless, by stealing the handle to a water pump
(thus cutting off that particular water supply) he saved a lot of lives.
Mechanisms are nice for proving a cause/effect relationship, but most
MD's are willing to take actions that will prevent disease without
waiting for that relationship to be proven.
Tony Harrison, MSPH
Colorado Dept. of Public Health & Environment
Laboratory and Radiation Services Division
(303)692-3046
tony.harrison@state.co.us
>>> "Estabrooks, Bates (IHK) " <IHK@Y12.doe.gov> 12/14/01 06:51AM >>>
Radsafers,
Help me out here in regards to this issue of "ecological" studies. Let
me
say right up front, that in this arena I am generally ignorant. As for
me,
the answer to the now-famous question: "Are you a statistician?" is
emphatically: "No!"
As I understand it, an "ecological" study simply looks at a broad
cohort of
people and looks to see whether certain conditions (health indicators)
exist
in relationship to some varying influence, without defining a mechanism
for
how the influence produces an effect on health. Is this correct?
By contrast, an "epidemiological" study seeks to prove something by
identifying the mechanism. Is this correct?
If my assumptions are correct, I am puzzled as to why the use of an
"ecological" study for disproving LNT is so roundly impugned, when
this
seems to be a credible approach in other technical fields. For
instance,
yesterday NPR had a story about a new study showing that women with
breast
cancer do not live longer if they participate in a support group. A
previous study had asserted the opposite. Both of these studies (as I
understand) simply looked at a large group of women and compared group
involvement vs. mortality, with no significant attempt to identify a
cause-effect mechanism. Yet, within the medical community these
studies
carry notable weight.
So, educate me. Please feel free to respond directly to my email, or
post
if you think the group would be interested.
Thanks.
Bates Estabrooks
Facility Safety-EUO Restart
BWXT Y-12
9983-FS
P.O. Box 2009
Oak Ridge, TN 37831
865-574-7376
865-241-5780 (Facsimile)
ihk@y12.doe.gov <mailto:ihk@y12.doe.gov>
************************************************************************
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 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.