Examination of Witnesses (Questions 680
- 699)
THURSDAY 13 JANUARY 2005
MR EDDIE
GRAY, DR
STUART DOLLOW,
MR CHRIS
BRINSMEAD AND
DR JOHN
PATTERSON
Q680 John Austin: Can I go back to
something you said earlier about when there was a discussion about
the media and the way in which the pharmaceutical industry may
be now viewed because of high-profile and perhaps exaggerated
claims in the media, for example, "New blockbuster drugmiracle
cure", et cetera. You seem to be suggesting that the
media were at fault here. Is not the industry sometimes at fault
by having sort of over-inflated claims for some of its new products?
Perhaps I could put the question back to AstraZeneca and let's
take Iressa, an innovative cancer drug, which you say in your
evidence to us has been described as "the biggest advance
ever in the treatment of lung cancer", yet I understand that
last month the Food and Drug Administration indicated it was considering
withdrawing Iressa from the market after the drug had failed to
prolong the lives of people with advanced lung cancer. Are there
any lessons which we should draw from this?
Dr Patterson: Well, there are
lots of lessons and I could spend a long time talking to you about
them. Let me come back to your first point though, that somehow
I had implied it was all the media's fault which I think the Chairman
said in response to my first answer. I am not saying it is the
media's fault. I am saying it is an outside world that has changed
and that that activity is going on and, therefore, that is why
there is so much visibility of what is happening with our products.
Let me now switch to Iressa. It is a classic example of where
sometimes we are damned if we do and we are damned if we do not.
Perhaps later on you will be talking to us about only doing me-too
products. Well, here is a very good example of something that
is an absolute breakthrough product, brand new type of agent,
an EGF (Epidermal Growth Factor Receptor) tyrosine kinase inhibitor,
based on very good pre-clinical science and brought to the marketplace
in Japan and the USA and a European licence applied for. The early
data showed some dramatic responses in patients with lung cancer,
many of whom had failed all previous therapies, actually having
a complete regression of their disease and some of them going
on to live for several years on this product and with very few
side-effects, which is something that just does not happen spontaneously.
In the open studies in third-line usage in previously failed patients
we saw and we treated something like 70,000 patients, stabilising
disease in about 40% and about 10% of them giving a very significant
response. Now, interestingly enough, there were certain sub-groups
who showed a very good response, in particular females and oriental
people, Asian people particularly from Japan. We then achieved
a licence in the United States under a system that does not exist
here, called "Sub-part H" in their procedure which allows
you to enter the marketplace while undertaking what are called
"the definitive trials", and in this case the definitive
trial was a placebo-controlled mortality or survival study in
a very significant number of patients followed all the way through
to death. That study was undertaken and unfortunately, (in spite
of some of the evidence you have had, to the contrary) our studies
are designed to give both positive and negative outcomes, they
are balanced, this one showed that although there was the same
response rate as we had seen previously, the same time to treatment
failure, the survival benefit did not reach statistical significance.
As a result of that, our European licence, which was under review,
was withdrawn and we had further discussions with the Food and
Drug Administration and we will go through some advisory committee
processes as to whether it is appropriate to continue to sell
that product. We have withdrawn all of our marketing support of
the product while those discussions are ongoing. Interestingly
enough, within that there is a sub-group of patients from South-East
Asia who had a very good survival benefit, some 22%, so we are
learning a lot about it. During the course of last year some doctors
in the States described some mutation changes that seemed to associate
with those people who were getting these very good proven responses,
so it may well be we will be back later in the year to say that
there is now a sub-group of patients whom we can identify who
actually will be the correct sub-group of the population to give
this agent to and who will much more likely get survival benefits.
It is a rather long answer, but it is a classic example of how
we, as an industry, develop modern, very potentially effective,
but completely unknown medicines, the kind of issues we can get
into and the kind of science we have to keep putting into these
products over many years to get all the right answers.
Q681 Dr Naysmith: Maybe the question
I am about to ask is perhaps related to what you have just been
saying in terms of the information you get from clinical trials,
how they are designed and so on and how you get the right answer
because we have heard quite a lot of evidence here to suggest
that the design of clinical trials and the manipulation of the
results is sometimes done in a way which favours the product which
is about to be launched, things like what you actually test the
product against, how much you put into negative results, how many
negative results are published and that sort of thing. Perhaps
Dr Dollow, who must have had a lot to do with a number of clinical
trials, would like to suggest, if there is this criticism of the
way clinical trials have been carried out in the past and you
agree with it, why people, patients and clinicians, should trust
the information they get from you in other ways?
Dr Dollow: Well, I think it might
be helpful if I explained to the Committee how we design our overall
clinical development programmes. When we are looking at a potential
medicine before it actually gets to even getting to a pill that
a patient might take, the programme is designed, first, to see
if it has acceptable efficacy and safety in the target that we
want and then we design the programme, saying that assuming that
these are going to be correct, we will look at our phase two,
for example, dose-range finding studies and then our phase three
studies. All of these will be pulled together and seen as a global
programme. We discuss this programme with regulatory authorities
across the world and we discuss it with experts in the field to
say that in some cases with some of the studies there is a need
to prove that the product has some efficacy, and those will be,
for example, placebo-controlled studies. In other cases, it is
important to see the incremental benefit that a medicine may bring
and those will be comparator studies, and the way we choose our
comparators would be looking at which are important medicines
which are being used across the world, so because all of our studies
or the bulk of our studies are multi-national studies, something
that is important in one market may not be important in another
market, so we will look at that in the round. The other thing
you talked about was around disclosure and, just to reinforce
the point that Eddie made earlier regarding the clinical trials
register, every single one of our studies for our medicines from
the earliest clinical pharmacology studies in healthy volunteers
through to every single clinical study that is done subsequently
will be made available on the GSK clinical trial register at the
time that medicine is first launched, and then subsequent studies
will be publicised within 10 months of each study finishing.
Q682 Dr Naysmith: Are you saying
then that it is impossible to carry out trials at any stage in
the procedure you have just outlined which produces negative results
and for that not to be known?
Dr Dollow: I think it is certainly
possible that studies produce negative results, and John Patterson
just gave you a very good example of when a study produces a negative
result. We always try to publicise one way or another the results
of our studies and we know and, as evidence you have heard previously
has shown, the journals are actually sometimes reticent to publish
negative studies. If I look at The Lancet or the BMJ
websites, they say that they actually only publish around 10%
of all the studies they receive, so as a result some of our studies
are put into journals, some of our studies are published as abstracts
and some of them are presented at conferences.
Q683 Dr Naysmith: I am not saying
it is necessarily your fault, but that in this whole process what
sometimes obviously happens, and some of our witnesses have told
us this, is that results are cherry-picked out of a pool of possible
results to establish the effectiveness and sometimes the safety
of products and others which might be used for other purposes
are kind of ignored and sidelined.
Dr Dollow: All I can tell you
is our position, that, firstly, we endeavour to publicise the
results of all of our studies, be they negative or positive, and,
secondly, the clinical trial register will put an equal footing
on that for all of GSK's studies from the year of the start of
GSK in 2001.
Dr Patterson: Perhaps I could
just add a couple of things to that. First of all, when we apply
for a licence to sell a product, we submit data on every single
patient who has been treated with that product in positive or
negative studies, the whole lot, so there is absolutely no selection,
whether it is on efficacy or safety, and all licences, therefore,
are granted on the totality of the evidence.
Q684 Dr Naysmith: So where do you
think this comes from, that witnesses have been telling us that
sometimes there are negative results which are not published?
Dr Patterson: Once the product
is in the marketplace, then we continue, as I said earlier, to
undertake clinical trials and we have had a policy within AstraZeneca
now for a number of years where we would seek actually to publish
all of our clinical trial data and we have lots of examples where
we have done that. There is a product called Viozan that failed
in phase three trials and we published all of that data in a supplement
to one of the journals, but, as Dr Dollow says, it is not always
easy to get negative results published and that is why I think,
in response to the questioning that is going on, we are saying
that perhaps the best way is we are going to put them up on a
website ourselves and then everybody can see. The other criticism
that some of your witnesses have given in the past is this one
that we pick the easy comparator, we just go against the old medicine.
I think the example they used was the COX-2 inhibitors and why
did not Vioxx compare with Celebrex. Of course that is an interesting
one because at the time when they were both doing their phase
three trials and they were within one to two years of each other,
remembering you do your phase three trials and they take about
two years to run, you then go through a registration process which
takes maybe another two years, so there is a four-year window,
therefore, by the time they got to the market, neither product
could possibly have done a comparison with the other one because
it was not a registered, marketed product, so they did comparisons
with Naproxen which was at that time the gold standard in non-steroidal
anti-inflammatory drugs and they could not have done anything
else. Post-marketing, then you can do head-to-head comparisons
and most of our products that are coming to the marketplace now,
unless we have done a head-to-head comparison with the market
leader, and that can be, by the way, by volume use as well as
by value and I suspect Naproxen is still probably, by volume,
the market leader, then we are not going to get on to formularies
and get acceptance.
Q685 Dr Naysmith: Well, that brings
us on to another topic which, I must say, has really surprised
me since we started this inquiry, which is the question of ghost-writing
and looking around for clinical scientists and medical scientists
who have the reputation to put their names on papers which carry
data which they themselves have not had very much to do with and
may even know nothing about. Do you think that this practice is
widespread?
Dr Patterson: It is absolutely
not practised in my company. We have a set of publication policies
that we have had in place for a number of years and we work with
doctors in the outside world in our clinical trials. Perhaps I
can take a step back and just give you an example which I brought
with me from a phase three clinical trial that we ran on an anti-coagulant.
It was undertaken in 260 hospitals with some 3,400 patients. The
amount of data that generates is staggering, with 16,000 events
monitored, 30,000 blood samples, 50,000 patients visits, a total
of 12.5 million data points. We then, as the company, analyse,
collect that, quality-assure it by AstraZeneca personnel to make
sure the data are correct at the hospital sites, and we then turn
it into a report that is some 200 pages long. Those kinds of studies
we start by having a group of doctors working with us or an individual
who will help us design the study and they will stay with the
study throughout. We will often have a separate data-monitoring
or safety committee which will look at the data to make sure we
are doing no harm during the course of it, and at the end of it
we do employ professional writers within the company to turn that
200 pages into a draft publication, but it is done with the people
who are going to become the authors and they have the opportunity
to question and challenge every single conclusion and every single
table that is put in there. At the end of the day they have a
responsibility, just as we have the responsibility for getting
the analysis right, in putting their names to it to say that they
have taken adequate care and worked with us to deal with it, so
that is, if you like, the one extreme.
Q686 Dr Naysmith: I am not going
to stop you saying any more, but if that is an example of what
happens and you are talking about a really big trial, sometimes
there will be, and it is hard to say how many, but dozens of names
of people who have been involved in the trial.
Dr Patterson: Yes.
Q687 Dr Naysmith: So how would you
decide which names go on the scientific paper?
Dr Patterson: Well, there was
actually in the case of this trial, and I have got a flow diagram
which I can show you, the committee which we had set up to begin
with which met and had a meeting to decide whose names would go
on it. It is often quite a fight amongst them as to who goes on
it, but then whoever does will get the responsibility on behalf
of the others to make sure it is right. That is no different from,
say, something that the Medical Research Council runs as one of
its major trials or other co-operative groups outside of the industry
itself, so yes, you will get a small number of names and if there
is a significant AstraZeneca author as part of that, we would
hope and expect that their name would be there as part of that
report.
Q688 Dr Naysmith: But you are saying
that you would never seek someone who may be someone who is also
a scientist at an institute where this trial has been carried
out who is a big name and well known to put their name to the
paper, not having been involved throughout?
Dr Patterson: And not gone and
worked through all the data and not had the responsibility? That
is an absolutely unacceptable activity.
Q689 Dr Naysmith: From what you have
said, Dr Patterson, do you think that the stuff that we have had
describing this, that it would not be an acceptable practice and
do you think it would be open to abuse, what some of the witnesses
have said to us?
Dr Patterson: If what the witnesses
are saying is correct and they have evidence to support it, then
I think it is an unacceptable activity. The idea of hawking around
a set of results and a cheque book and saying, "Put your
name here and we will give you some money" is against every
principle that we and the medical journals live by.
Dr Dollow: I was just going to
add to John Patterson's point that the International Committee
of Medical Journal Editors have set out some guidelines for authorship
and they read that, "Authorship should be based upon substantial
contributions to conception, design, acquisition of data, analysis
and interpretation, drafting the article, revising it critically
for important intellectual content and final approval of the version
to be published". That is something that we strictly abide
by. We may well have people who put the first draft together,
but absolutely those people who put their name to the authorship
must have been involved at every step of the way. It says, "Authors
should meet conditions 1, 2 and 3". The other thing that
Dr Horton mentioned in his evidence was that people who contributed
can also be listed as contributors, if not formal authors, and
again that is something we support as well and is something we
have been doing for many years. The issue of ghost-writing, as
alleged, is not something I recognise at all.
Q690 Dr Naysmith: So you would say
that it is an unacceptable practice if it does happen, trying
to get a prominent scientist or a prominent medical person to
put their name on it without having any kind of involvement in
the process at all?
Dr Dollow: We would not support
that in any way, shape or form, absolutely not.
Q691 Chairman: Are you saying in
your experience that it does not happen within the industry? You
have explained your own company's position on this, but it does
not happen in the industry, as far as you are aware?
Dr Dollow: I can only speak for
GSK. My personal experience is that I was working for one other
company before I worked with GSK and it certainly did not happen
in that company and I have been in the industry for 14 years now.
Q692 Chairman: But you are not making
the point that it does not happen at all, but that in your experience
of the company you worked for and your current company it does
not happen?
Dr Dollow: Absolutely.
Q693 Dr Naysmith: Is that the same
for Dr Patterson?
Dr Patterson: It is completely
unacceptable in my company, but I cannot say it has not happened
in other parts of the industry.
Q694 Mr Jones: I just want to clarify
a slight ambiguity that I thought arose in those two answers.
When Dr Dollow read from a code of practice, one of the requirements
was that the author drafted
Dr Dollow: ". . . a substantial
contribution to drafting", so they have to have been involved
in reviewing and writing, so absolutely, yes.
Q695 Dr Naysmith: But in Dr Patterson's
answer, he referred to professional writers, so perhaps you forgot
about the drafting.
Dr Patterson: Not at all, no.
We do employ professional writers and they are involved. Most
of them are not experts in that particular field of medicine,
but they are very good at writing good English and they also understand
what it takes to write an article that can be published.
Q696 Dr Naysmith: Perhaps we have
difficulty in understanding what "drafting" means then.
If the person whose name is at the bottom did not either write
part of the article or how the article would come out, then they
did not have any involvement in drafting, as I understand drafting,
and, therefore, the code of practice is broken.
Dr Patterson: The way that it
works with professional writers is that they will possibly produce
a first draft which will then be reviewed by the author or authors
or a committee and who will have significant input into that drafting
and how it should then subsequently go forward, so we start with
a 200-page report, as I described, and the medical writer will
then pick that up and try and turn it into something, if it is
going into The Lancet, that is in the format that these
people recognise, remembering that many of the senior doctors
who have worked with this may not have English as their first
language or are not going to have the time, the effort and the
knowledge as to what it takes to have the right format for these
journals. That is drafting, but that is not the whole of the drafting.
It then goes through a series of iterations.
Q697 Dr Naysmith: I would guess there
would be a problem if you are talking about translators. I think
I would not be asking these questions if you said you employed
translators.
Dr Patterson: No, this is not
a translator who takes somebody's foreign language into English.
This is somebody whose native language is English who is a professional
writer who takes a report and turns it into a draft which would
be acceptable to a medical journal.
Q698 Mr Burns: Just to clear this
up, maybe you could help us with this analogy. Between 1994 and
1997 in particular, but also thereafter, there were dozens and
dozens of articles that appeared in papers like The Sun,
The Daily Express and The Daily Mail, written by Alistair
Campbell or Peter Mandelson, but having Tony Blair's name on them,
although he had clearly never seen the articles! What you are
saying is that that does not happen, but there are some people
who may not be as fluent in the English language who need a professional
draftsman to put their thoughts and their research into English
to make a compatible article for reading. Would that be helpful?
Dr Patterson: Yes, or who even
collects together the information for a first draft from which
you can then present your thoughts.
Q699 Mr Burns: But it is more accurate
that way than just having some skivvy writing an article for a
newspaper that you have never seen until you read the newspaper?
Dr Patterson: And then putting
somebody else's name to it, correct.
Chairman: Are you making a point there,
Simon!
|