Subject: Re: Licensed Naturopathic Physician Quacks?
From: email@example.com (Jay Mann)
Date: May 25 1996
Matti Narkia (firstname.lastname@example.org) wrote:
: There are also ethical considerations to be taken into account. If a
: medication is expected to have some effect (it usually is, because
: it is subjected to testing), is it ethical to test it against placebo with
: seriously or even terminally ill patients? A comparison test with some
: other medication or with a historical group (like Cameron and Pauling did)
: seems more appropriate.
Some of us graybeards will have read Sinclair Lewis' novel "Arrowsmith",
which deals specificially with the ethical problems of testing medication
vs placebo. The logical trap is the "expected to have some effect"
argument. Until the test is performed, that expectation is only a
hypothesis. Once the results confirm the efficacy of a treatment, then
it could well be terminated early. This is exactly what happened in
China with the test of whether selenium supplements could prevent the
childhood heart disease in a selenium-poor region of that nation.
Although the plans were for a three-year trial, by the end of the second
year the data was so statistically convincing that there was no longer
any moral justification for withholding selenium from the kids who had
been getting placebo pills. BUT until that moment, the double-blind
testing was the only moral way to go, because there was no proven treatment.
As this posting correctly states, even double-blind tests can be
deceptive. But that doesn't mean that double-blind isn't the way to go,
because non-blinded or even single-blind tests offer unlimited
opportunities for self-deception by the experimenters and/or patients. A
chap I worked with had been a test subject for what was purportedly an
extract of schizophrenic blood that would induce psychotic symptoms in
others. He said that the experimenter, a well known psycho-physician in
a southern US state, first administered a placebo injection; this was
done casually. Then the physician with great care, not spilling a drop,
filled a syringe with another extract, injected it, and leant forward
tensely waiting for reports of psychological symptoms. Since the extract
was known to have ammonium sulfate in it, which would cause flushing and
other physical symptoms, it wouldn't have been difficult for a gullible
subject to start expressing psychological unrest, particularly when
unconsciously encouraged by the (non-blinded) experimenter.
For another example of this, read Martyn Gardner's chapter about Hans, the
horse that was supposed to be a mathematical genius -- purely prompted by
the owner. One of the late-night interview people in L.A. had a woman
with a similarly mathematical dog who barked out the answers. --- or
more obviously, barked until his owner relaxed. When the woman made a
mistake in a multiplication problem, so did the dog. Funny that.
Jay D Mann <email@example.com>
Christchurch, New Zealand
From: firstname.lastname@example.org(Steven B. Harris)
Subject: Re: A Question for the ALT crowd, if they can drop into the real world
for a minute
Date: 27 Apr 1999 01:18:54 GMT
In <email@example.com> "mjdgdc"
>One more point: no procedure, manual, surgical or otherwise, can be
>tested through a "double blind" study, so stop throwing that term around
>(or at least go review what it means). In a double-blind study, both
>patient and provider are unaware of whether the tested method or the
>placebo is being delivered. How would you propose to design a
>double-blind study for any procedure? A surgeon who doesn't know if they
>are removing a disc or not? Have the chiropractor close their eyes and
>flail their arms in front of them? A massage therapist who doesn't know
>if there is a patient on the table or not? The best any procedure can
>hope for is single-blind in which the patient does not know if the
>treatment was performed (which in some procedures isn't even possible),
>but the provider would obviously have to know. So if we reject all
>procedures that were never tested in double-blind studies, you would
>eliminate ALL surgical procedures, and just about every aspect of
>physical medicine. Since that is neither reasonable nor fair, you must
>accept that such procedures are tested in other ways, but not
>double-blind. If you're having trouble with the concept, take a second
>and try to imagine a double-blind study on bypass surgery.
Sigh. They should have explained this to you in school. Double
blind doesn't mean the therapist doesn't know what he did. It means
the patient doesn't (has no way of getting it by non-verbal cues) and
the evaluator doesn't. In practice this is difficult to insure unless
the patient is unconscious, but it's quite easily done then.
And FYI there have been a number of famous double blind surgeries.
Two that come to mind were for angina, a notoriously difficult problem
to evaluate unless done in a blinded manner. One study tested the
effectiveness of "poudrage", in which asbestos powder was dumped on
the naked heart under the pericardium, in the belief that the
inflammation generated thereby would produce new circulation. In
unblinded studies it seemed to work clinically. But patients who
didn't know if they'd gotten no powder vs ones who had their chests
opened and did, and who never saw their surgeon after the procedure and
before evaluation, and were evaluated by doctors who didn't know what
they'd had done, showed no improvement over controls. A similar study
looked at internal mammary artery ligation, which was supposed to shunt
blood to the heart, where one group had a sham opperation (skin cut
then sewed back up), and the other actually had the artery ligated
(tied off). When evaluated as above, no difference. That was the end
of both procedures. Never say surgery is never scientific.
Steve Harris, M.D.
From: "Steve Harris" <firstname.lastname@example.org>
Subject: Re: Double Blind Test
Date: Wed, 16 May 2001 10:30:50 -0700
Jay wrote in message <email@example.com>...
>What is a double blind test?
>What is the purpose of it?
It's a test of a medical treatment in which neither the
patients nor the people who administer, coadminister,
or evaluate the effect of the treatment by having
contact with the patients and their caretakers,
know whether or not the treatment
or the placebo control is being given to any given
patient. This avoids (or controls for) the placebo
effect in the patients,
and also the unconscious biased observation
effect in the evaluators.
The randomized double blind clinical trial is the
gold standard for gauging the effect of treatments
in which the treatment endpoint is at all susceptible
to unconscious bias by either treators or treatees.
Obviously, not all treatments NEED to be looked at
in this manner. For example, since nobody really
seriously thinks that people mentally influence the
size of their tumors, a drug which shrinks tumors
can be tested in a single blind fashion (ie, only
the radiologists aren't told who got the drug).