Ridecamp: Re: [RC] [RC] [RC] shoes reopened...NOT; now magnetic boots

" The twenty nine patients with active magnets reported, on average, a
significant reduction of pain (from 9.6 to 4.4), while the twenty-one
patients with shams reported a much smaller average reduction (from 9.5 to
8.4). This is a substantial difference, and if the double-blind study was
successfully conducted, cannot be explained by a placebo effect. "

Is this based on measurements of chemicals in the blood related to pain, or
simply perception on the part of the patients?  Did they have any subjective
way to classify level of pain between patients? Different ppl have different
pain thresholds--my grandfather had to really be in pain before he'd even
take anything for it, much less admit that he was hurting. Unless the
results are consistently repeatable, over multiple studies, the researchers
can't draw any real conclusions. Just because the study is double-blind
isn't enough to go on at this point.

I can't help but being skeptical--pain, both the causes and perception of
it, is a very complex issue.  Also, I have several issues with the
experimental design of this particular study:

In this case, we are dealing with a company-funded study where only 1 brand
and type of magnets were used, supplied by that company. The results would
be more impressive if there were multiple brands represented.  Also, from
your description, the statistical design of this study was randomized
complete block, the simplest form, and the one most likely to result in a
Type I or Type II error because it does nothing to eliminate random
variables. So, what, you may ask, are Type I and Type II errors, and why
should I care?  Quite simply, a Type I error in statistics is accepting a
true hypothesis as false on the basis of a test of the statistics, and a
Type II error is accepting a false hypothesis as true.  There are a lot of
random variables in this study that could be causing the authors to commit a
Type II error in this study:  Effect of day (pain levels typically aren't
the same for ppl on a day to day basis--it depends not only on the pain
itself but what else is happening in their lives), pain threshold of each
patient, relying on pain perception only as the criterion for evaluation,
being unable to adequately compare pain levels between patients.

The design is unbalanced, with more ppl being exposed to magnets than not.
Being the skeptic that I am, I'd have to look at that and question just how
significant the results really are and whether that was a deliberate attempt
on the part of the company supplying the magnets to skew the results.  Eight
missing samples out of 29 is fairly significant.

Everyone is sitting there and saying, but she believes the arthritis surgery
study, and not only is it based on perception of pain but they couldn't
switch treatments on the patients at all.  Both arguments are true, but the
surgery study has several things going for it that this study doesn't--in
the arthritis study, there were two different types of surgery compared
along with the sham surgery.  Also, more patients per study, which helps
decrease variation within treatment.  The study also evaluated the patients
over a period of 2 years, so there are multiple evaluations over different
pain levels for each patient, which helps to eliminate effect of day and
difference between patients.

I'm not trying to start a war, I'm not deliberately being stubborn, and I'm
not trying to make up your minds for you.  I am, however, showing you what I
look for when deciding whether or not to believe something, and I take the
same approach regardless of whether the product is alternative or
conventional in nature.

What I want everyone who looks at these alternative therapies (or for that
matter, anything) to do is be able to sit back and evaluate reports of their
effectiveness with a critical eye.  Statistics are held up to the public by
the media as being something magical and unquestionable, but they aren't.
As Queen Victoria's PM, Benjamin Disraeli once said, "There are three types
of lies:  lies, damned lies, and statistics."  Statistics can be very
useful, but for them to be used effectively and to be believable the
experiments which generate them have to stand up to scrutiny.  This one
doesn't, at least in my opinion.

Lisa


=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Ridecamp is a service of Endurance Net, http://www.endurance.net.
 Information, Policy, Disclaimer: http://www.endurance.net/Ridecamp
 Subscribe/Unsubscribe http://www.endurance.net/ridecamp/logon.asp
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=