Statistics

[Guest guest]

>>My post on Paul's comments about statistics

was not meant to be an assault on research.

Those who read it that way read it wrong and

if you care about that, just go back and read it

again. Ken>>

 

My take on this issue is that there is not anything necessarily wrong

with statistics, and I don't think that this is what Unschuld meant

to imply, although I was not there.

 

Statistical methodology can be profound, but ultimately it may have

little or no relevance to the individual case. If a medicine is shown

to be slightly more efficacious than a placebo, this may show that

has some efficacy, or perhaps better phrased, potential for efficacy

in certain circumstances, but that doesn't prove that it's a good

treatment, or even suitable for an individual. In fact, it may be

revealed to be a bad treatment. A number of years ago, the Journal of

the American Psychological Association published a review of placebo

controlled studies of Prozac, in which it was found that Prozac was

only very slightly more efficacious at treating depression than an

active placebo. Now, in clinical practice, I've seen individuals who

benefitted greatly from Prozac, and I've seen people for whom it was

very doubtful whether they were benefitting at all, in their or my

estimation.

 

That's the problem with statistics - when enough information is

available, and research has been rigorous, they can tell us

something, but perhaps very little or nothing that is relevant to

the individual case. It may certainly be inappropriate to come to

firm conclusions about the efficacy of a treatment on the basis of

statistics.

 

This is one issue that I imagine Unschuld is commenting on - the

drawing of false conclusions, or inappropriate recommendation of

medical treatment, based on inappropriate use of statistics.

 

Another issue that I gleaned from the report of Unschuld's comments

was that people who fit outside the pigeon holes of biomedical

diagnosis become marginalized. In the UK, I've even noticed, first

hand, a tendency to accuse people of Factitious Disorder or

Munchhausen by Proxy when a firm diagnosis couldn't be determined,

when to me it was totally evident that the individual had a real

medical condition. Freud warned psychoanalysts about reification, and

I think that reification is an issue in biomedicine (and possibly in

any type of medicine).

 

Two additional points - it may be possible to further refine

biomedical prescribing by analysing pharmaceutical drugs according to

Chinese methods of differentiation, and tailoring drug selection to

the individual. This would be a form of integrated medicine in

reverse to the way it's often conceived.

 

Another point - it is precisely when patients' conditions fail to

correspond to well defined patterns that Yi becomes important in CM

therapeutics (not the only time Yi is important, as it's always

important, but a situation in which Yi is very signficant).

 

Wainwright