On 13 December 2006 Paul Okami wrote [snip]:
As a last word, critiques of Kirsch & Saperstein 1998 are old news,
and addressed in the study using the suppressed FDA data, published
in 2002.
The idea that a meta-analysis published in 1998 of studies of the efficacy
of antidepressants is "old news" is an interesting notion, but hardly one
to be taken seriously. Kirsch et al (2002) was a meta-analysis of efficacy
studies submitted to the FDA by the pharmacological companies, a very
different animal from an analysis of studies undertaken by researchers in
the field (Kirsch et al [1998]). The notion that the second somehow
supersedes the first doesn't bear serious consideration - they are
examining a very differently based series of studies.
Once again, the commentaries following that article all affirm
that the basic statistical findings are accurate -- it is the
interpretation
of the meaning of these findings that differs between those who
support the "Emperor's New Drugs" view of antidepressants
heralded by the title of the article, and those who support the
use of anti-depressants and believe they are important weapons
in the treatment of depression.
Once again? If that is supposed to mean that the commentaries following
the *1998 article* all affirm that the basic statistical findings are
accurate, it is erroneous. Of the four commentaries, as I've already
noted, two (Dawes and Klein) disputed the validity of methodological
features of the meta-analysis, Klein's being titled "Listening to
meta-analysis but hearing bias".
http://journals.apa.org/prevention/volume1/toc-jun26-98.html
Is it true that the nine commentaries to Kirsch et al (2002)endorse their
findings?
http://journals.apa.org/prevention/volume5/toc-jul15-02.htm
One of them (Salamone) writes of the Kirsch et al contentions
"Nevertheless, it is important not to exaggerate the claims of minimal
effect size...", hardly a total endorsement. Another (Brown) writes
"However, these drugs may have more inherent effectiveness than they
appear to in clinical trials." A third (Hollon et al) writes "We suggest
that not all patients necessarily respond to a given medication and that
effect sizes based on the 'average' patient may underestimate drug-placebo
differences for those who do. Data submitted to the FDA can also
underestimate how a drug will perform in clinic practice..."
These hardly constitute affirmations that Kirsch et al's basic position on
the efficacy of antidepressants in clinical practice, which is where it
really matters. But in any case, that commentators responding directly to
an online article do not generally criticize the methodological procedures
of the article is hardly surprising. The 2002 Kirsch et al article (like
the 1998 article) was published in "Prevention and Treatment", a rapid
publication electronic journal which ceased publication at the end of
2003. This is hardly the kind of journal to rely upon for comprehensive
critical discussion of the methodological limitations of published
articles. I have no idea why people like Quitkin and Klein did not supply
commentaries, but it could well be that, having published detailed
critical analyses of the 1998 meta-analysis, they didn't feel inclined to
do the same again for another like article by the same lead author in an
online journal of a calibre that evidently does not have the status of
professional journals like the American or British Journals of Psychiatry,
surviving a mere five years.
Paul's implied contention that the commentaries to such a journal
constitute evidence that the Kirsch et al claim that their analyses show
that antidepressants are no better than placebo are no longer in serious
dispute does not bear serious consideration. How could Paul possibly make
such a claim on the basis of contributions to a single journal whose
professional status is such that it lasted a mere five years? The Parker
et al (2003) article I've already cited several times includes Kirsch et
al (2002) among articles it criticises, indeed saying specifically that
their conclusion that "the pharmacological effects of antidepressants are
clinically negligible" may well be viewed by clinicians as "specious" -
hardly an indication that Kirsch et al's analyses are generally taken as
proven. Paul's reiteration of his theme that antidepressants are no better
than placebo also ignores the fundamental point I've made more than once,
that Parker has provided cogent reasons for arguing that the current (DSM)
classification model on the basis of which studies are predicated is so
flawed in its failure to discriminate in terms of targeting of
pharmacological products that it has the result that most randomised
controlled studies are not representative of the clinical population, and
cites a study that suggests this is in fact the case.
Summing up, Paul's contention that "it is the interpretation of the
meaning of these findings that differs between those who support the
'Emperor's New Drug' view of antidepressants heralded by the title of the
article, and those who support the use of anti-depressants and believe
they are important weapons" is, not to put too fine a point on it,
uninformed nonsense. To take just one instance, Quitkin et al (2000)
argue, on the basis of a close analysis of articles such as Kirsch et al
(1998), that "Our examination of the original source material cited by
antidepressant sceptics suggests that these critiques of the
antidepressant literature are largely unsubstantiated."
What I find interesting is that early in this thread (9 December) Paul
chided Annette for allegedly failing to appreciate precisely what is meant
by a sceptical approach: "Skepticism is the philosophy of the scientific
method." Yet in the current discussion Paul takes as given that the claims
of Kirsch (lead author), whose basic philosophical position indicates, as
I pointed out, that he hardly comes to the debate from a disinterested
position, are proven, period. I already wrote that the debate about
disputed methodological procedures involves great complexities for the
non-expert, and I don't pretend to be able to get near to grasping the
subtleties (or even less subtle points) of what is in contention. Should I
congratulate Paul for having such mastery of the subject matter, and of
the methodological complexities, that he is able to confidently declare,
in relation to contentions that are vigorously disputed, that Kirsch et
al's conclusions are definitive?
Allen Esterson
Former lecturer, Science Department
Southwark College, London
http://www.esterson.org/
---------------------------------------------
Wed, 13 Dec 2006 08:58:02 -0500
Author: "Paul Okami" <[EMAIL PROTECTED]>
Subject: Re: SSRIs and depression and anxiety
As a last word, critiques of Kirsch & Saperstein 1998 are old news, and
addressed in the study using the suppressed FDA data, published in 2002.
Once again, the commentaries following that article all affirm that the
basic statistical findings are accurate--it is the interpretation of the
meaning of these findings that differs between those who support the
"Emperor's New Drugs" view of antidepressants heralded by the title of
the
article, and those who support the use of anti-depressants and believe
they
are important weapons in the treatment of depression.
Paul Okami
---
To make changes to your subscription go to:
http://acsun.frostburg.edu/cgi-bin/lyris.pl?enter=tips&text_mode=0&lang=english
__________ NOD32 1919 (20061213) Information __________
This message was checked by NOD32 antivirus system.
http://www.eset.com
