You are on page 1of 4

Thinking about Life Sciences: FDA panel votes to retain Avandia on the m...

Thinking about Life Sciences

Wednesday, August 01, 2007

FDA panel votes to retain Avandia on the market: DrugWonks celebrate;

Prince Hal rails against Falstaff ...

An FDA panel voted 22 to 1 on Monday (7/30) to keep Avandia on the market; the vote on the heart attack
risk was 20 to 3.
This “result” has been touted as a “victory” by certain pro-pharma groups while my position has not so much
been to take sides but rather to highlight these Avandia (and Vioxx) debates as springboards for the greater
issues of the growing importance of safety, the evolving risk-benefit balance in “post-modern” medicine, the
impacts on FDA reform and the “deconstruction” of the pharmaceutical business model.
Unfortunately others have made this into a highly personal issue which is fine for sandbox politics but not for
a topic that demands the highest rigors of scientific discourse and the deepest levels of human compassion.
We are dealing with lives here and fragile egos and statistical minutiae are ultimately only minimally
I am not here to defend Dr. Steven Nissen (whom I have never met though I have expressed admiration in
the abstract for at least taking a position, however precarious, on safety) nor am I here to defend particular
pharmaceutical company interests. This is an independent blog and as such harbors no agendas, no stock and
values logical thought and open, transparent discourse. Having stepped into the debate, however, I have seen
how scientific rigor and medical professionalism can unfortunately play less of a role in shaping public
opinion than would be hoped.
The painful purpose of this blog entry, then, will be to expose the dangers of the polemics polluting the
debate and hopefully in at least a small way facilitate a more “adult” and rational discussion.
Argumentum ad hominem
One doesn’t need to be a scientist or a lawyer to know that arguments on an ad hominem (or “argument
against the person”) basis are not only objectionable from the perspective of civilized discourse but are
also logically false in a formal sense. Even an elementary school student (the show “Are you smarter than
a 5th grader” comes to mind) will recognize that the following thought process is false:
1. Dr. David Graham says Avandia is unsafe
2. Dr. Graham is “Dr. Evil”
3. Therefore Avandia is safe
In this regard, I thought it important to point out how the Center for Medicine in the Public Interest
(CMPI) – a policy think-tank describing itself as "a forum offering rigorous and compelling research on
the most critical issues affecting current drug policy" – has, in fact, been a disservice to the pharmaceutical
industry by using unabashed, non-rigorous ad hominem attacks as the cornerstone of their efforts to advance
the interests of their drug company benefactors. For an industry that appropriately prides itself on being
based on science and ultimately in the service of people, it would be my recommendation that they
re-evaluate the value provided to them by - the CMPI associated blog - and related sites.

1 of 4 11/17/2008 12:38 AM
Thinking about Life Sciences: FDA panel votes to retain Avandia on the m...

DrugWonks is gleefully celebrating yesterday’s Avandia “victory” but it should be clear that the ad hominem
methods and techniques used liberally in are not sustainable in the long run. Having
the right forum for presenting a positive, logically consistent, accurate and open message is particularly
important for an industry that is undergoing significant structural change (c.f. the layoffs announced by
Johnson & Johnson this morning and my discussion of pharmaceutical business models in “FDA Reform
Redux: On Business Models, Regulatory Reform and Safety”).
A Statistical Analysis of
As pointed out in the article “Avandia, the Drug Safety Debate and How to Get a Nobel Prize,” building a
scientific career around drug safety does not put one on the fast track to scientific fame (or fortune). For
various professional, cultural and aesthetic reasons, modern science reasons doesn’t necessarily value such
safety-oriented studies despite the great societal need for such research. What is clear, however, is that
doing such work is certainly bound to incite more wrath than admiration as evidenced by the vicious
attacks sprinkled liberally throughout the blog.
Being that statistical analysis is the order of the day, here’s my analysis:
Hypothesis: – a pharmaceutical industry-sponsored think-tank/blog – engages in
ad hominem based arguments (also known in political circles as “negative attacks”)
Methods: Review any and all references to “Dr. David Graham” (or Graham and other nominative
equivalents) on today’s main blog page and assign an “ad hominem index value
(AHIV)” to that reference depending on whether any proximally situated adverbs or other qualifiers
were of an inflammatory or pejorative nature. An ad hominem index value (AHIV) of “0” signifies a
completely neutral stance while an AHIV of “1” would represent a position that was either
gratuitously or in fact denigrating of the target (in this case Dr. Graham). Tabulation of results and
statistical results thereof should provide some insight into the validity of the hypothesis as proposed.
Results: There were twelve references to Dr. Graham and the results are presented in tabular form

Conclusion: A total of 12 references to Dr. David Graham (or Graham) were found. Of these 11
had at least some denigrating or pejorative description, tone and/or implication. All of these 11
denigrating comments were assigned an AHIV of “1” with the exception of one which was deemed
to be only moderately inflammatory and hence given an AHIV of “½.” Only one out of the 12 had
a completely neutral stance towards Dr. Graham. By way of example, one of the negative entries

2 of 4 11/17/2008 12:38 AM
Thinking about Life Sciences: FDA panel votes to retain Avandia on the m...

entries (#10) refers to Dr. Graham as “Dr. Evil.” The mean AHIV was 0.88, the median AHIV was
1.00, the mode was 1.00 and the standard deviation 0.31. The data appear to point strongly towards
the frequent use of ad hominem argumentation on the site. More data is required to
ascertain statistical significance. Neverthless, the reader is encouraged to visit that site to form their
own opinions in this regard.
As The Statisticians Play On
Of course, the discussion above is somewhat absurd. You don’t need this statistical analysis to tell you
that is almost infantile (sorry for the ad hominem here) in their personal vitriol and
polemic. Like Justice Potter Stewart famously remarked in evaluating whether something was
pornography or not: "I know it when I see it." Feel free to read the DrugWonks blog for yourself. I hope
that this exposé may impel the pharmaceutical benefactors for DrugWonks to exert some adult
supervision over what must be an embarrassment to an otherwise fine segment of our economy.
But here’s the point. Statistics alone will not solve the problem of drug safety. To some extent, it doesn’t
really matter who is right, Dr. Bob Goldberg, Dr. David Graham, Dr. Steve Nissen or anyone else for that
matter. What is important is that there is a (hopefully) serious debate about drug (and device) safety and
to that end that improvements to the system – for the sake of patients and for the sake of the continued
vitality of our pharmaceutical industry – are being considered. To the extent that claims of “victory”
provide complacency along these lines, then such victories are pyrrhic at best, tragically misguided at
The essence of the problem is that the main burden of safety evaluation is assigned to post-market
surveillance studies which because of their retrospective, individual case report basis are intrinsically (and
notoriously) problematic in providing statistically robust results. As I had written in a comment on the blog (but which was not published):
The primary means for evaluating drug safety - apart from the rather routine and often perfunctory preclinical and
phase I studies - is the system of post-market voluntary surveillance data. This means, by definition, that such studies
are non-randomized, non-prospective, post-hoc and essentially meta-analytic. Even worse they are based on the
agglomeration of individual single case reports. That is about the most unsatisfactory dataset one could ask for in
supporting the most important mission of the FDA – namely protecting the safety of our medicines.
So: while the statisticians play on as the Titanic sinks, I ask this: “how can we truly expect safe medicines, if this is
the system we have to work with?”
The FDA reform proposals discussed in several postings on this Life Sciences Daily blog (for example:
“FDA: Tortoise, Hare or Something Else”) point to the need for more innovation, more rigor and
more priority being given to pre-market safety studies. Often, the immediate response to the
proposal to shift more safety evaluation to pre-market phases of development is that this would not be
feasible and would require large numbers of patients to conduct such studies. My reply to such critics is
that such a stance is (as the 9/11 Commission once famously said in regards to another national tragedy) a
“failure of imagination.” To this end, we need to apply new and innovative approaches to pre-market
safety evaluation, including, for example:
· Considering innovative surrogate safety endpoints,
· Utilizing prior data (especially the results of “negative” clinical trials),
· Developing more sophisticated and comprehenisve animal models,
· Using computer simulations to prospectively anticipate complex drug interactions and so on.
Falstaff and Prince Hal
Ad hominem arguments do have their place - we are all human after all and attacking one another is just a
part of the stage we all play on. And, whether it be Jay Leno or Shakespeare, roasting others is clearly a
much-enjoyed staple of late-night and theatric comedy.

3 of 4 11/17/2008 12:38 AM
Thinking about Life Sciences: FDA panel votes to retain Avandia on the m...

Shakespeare also used ad hominem attack not only for comic effect but also – in its own beautiful and
profound way – to create an enduring message of friendship and redemption. Take a read of the
wonderfully hilarious dialogue (Henry IV Part I, I.ii) below between Falstaff and his liege Prince Hal (the
later Henry V of Agincourt). Here, Prince Hal completely demeans Falstaff with a tirade that is both
entertaining and illuminating. From this, Shakespeare has us realize that it is not Falstaff who is made the
fool but rather it is Prince Hal who is the fool which magnifies the contrast between this princely
foolishness and his later redemption as the prodigal son, and heroic king, as Henry V.
Now, Hal, what time of day is it, lad?
Prince Henry
Thou art so fat-witted, with drinking of old sack and unbuttoning thee after supper and sleeping upon benches
after noon, that thou hast forgotten to demand that truly which thou wouldst truly know. What a devil hast
thou to do with the time of the day? Unless hours were cups of sack and minutes capons and clocks the tongues
of bawds and dials the signs of leaping-houses and the blessed sun himself a fair hot wench in flame-coloured
taffeta, I see no reason why thou shouldst be so superfluous to demand the time of the day.
Which leads to my last point. A posting from today - “Who ‘Lost’ Safety”
concludes by asking that:
“It's time for the grown-ups to step forward and take charge of the debate on drug safety.”
I agree 1000%. The pharmaceutical industry, the FDA, patient advocate groups, healthcare payers,
clinicians in practice and the biomedical academic community should continue to make progress in
reforming our approach to drug (and device) safety. We should let individuals and groups who are
inclined to rational, scientific debate take charge and jettison all other “noise” from polluting the debate.
And, like the fool Prince Hal who was later transformed into a great king, there is likewise redemption and
opportunity for DrugWonks … so long as they behave like adults.

Ogan Gurel, MD MPhil

4 of 4 11/17/2008 12:38 AM