interview with Harold Varmus on Open Access and PLoS
- FREEDOM FIGHTER
It must have been amazing news for Harold Varmus. More than 500,000 hits on
13 October had crashed the servers as people rushed to log on to the debut
issue of the pioneering science journal he had helped create [www.plos.org].
The furore caused by the journal shows no signs of going away. Luckily,
Varmus is used to the heat: as head of the [US] National Institutes of
Health, it went with the territory. But what made him, now in another
high-profile job, sign up for this particular controversy? Kurt Kleiner was
1 November 2003
NS: You are in the middle of a distinguished career, and the system has
treated you well, so you don't seem a natural for an idea like this. Why
does it strike such a chord?
Varmus: Because publication is the heart of the scientific effort. Nothing
glues us together as a community more than publishing. That's what people
work for. It's the moment of revelation and potential embarrassment. You are
showing your data and your conclusions and your way of thinking and the
heart of your life's work to your critical and competitive colleagues. So
it's a big moment in everyone's life.
NS: But why open-access publishing? And when did the idea hit you?
Varmus: I was converted by Pat Brown. Pat is a biochemist at Stanford and
had done a presentation in his own lab on the work of a guy named Paul
Ginsparg, one of the founders of the open-access movement. He set up a
website [now developed into the arxiv.org site] that published physics
preprints, and many of these articles eventually got published in
conventional journals. But the preprints were all online and free access.
That movement began a revolution. I was then the director of the National
Institutes of Health and I realised there was incredible potential to do
something in the biological sciences that would be really, deeply important,
not just for the advancement of science but for providing information that
the public really wanted to know. I have a special interest in advancing
science in poor countries, and this was obviously an important way to do
that. Open-access publishing requires no subscriptions to use the digital
version, allows any use of the material as long as attribution is
maintained, and involves placing the material in a public digital database
that can be rigorously searched. Many of us are doing so-called
high-throughput analysis that generates much more data than we ever
interpret. Allowing others to mine that data for new observations is
incredibly important and exciting. So I thought about this for a while and
then wrote a manifesto called E-Biomed.
NS: What happened next?
Varmus: That was an interesting moment. I must have known that I was not
going to be at NIH for much longer, because this caused a tremendous
political argument: what the hell was I doing trying to destroy the
publication industry? And actually I went too far. Politically I was naïve
to describe the full vision rather than proceed one step at a time. People
in Congress who had been, and who still are, my friends wanted me to defend
what I was doing because they were being told by the lobbyists of the
publishing houses that I was out to destroy the capitalist economy, so I
think I would have done that differently. But it definitely got people's
attention. I revised the manifesto and tried to make it more targeted, and
by the end of the year we had PubMedCentral [www.pubmedcentral.nih.gov] up
NS: So what is PubMedCentral?
Varmus: It was a first step in what we hoped would become a digital
repository of all the works in biomedical science. Though ideally those full
texts would be deposited at the time of publication, or even before
publication, we realised that for most journals this was not an option. They
would only provide content after a delay of anywhere from two months to a
year, if at all. But even faced with good evidence that you are not going to
harm the personal subscriber base if you delay deposition for six months to
a year, many journals were unwilling to take the step. Frustrated by the
slow progress, we decided we would try to generate interest in the
scientific community, saying: "We're not going to submit our papers, provide
reviewing, or do editing for journals that don't provide their content to an
open-access repository like PubMedCentral."
NS: And when the due date came around, what happened?
Varmus: It was clear that not as many journals as we would have liked had
gone over to the PubMedCentral model. That's when we realised we needed to
do something a little more ambitious. One of the things we were interested
in doing was creating journals that did it right. So three of us got
together: Pat Brown, Mike Eisen [an evolutionary biologist at Berkeley] and
myself. We were the founders, more or less. We decided we would write a
prospectus and shop it around. And we finally got lucky with the Gordon and
Betty Moore Foundation, who were very responsive and gave us around $9
million, which we used to hire extraordinary people.
NS: But why bother? What's wrong with the established scientific publishing
Varmus: Well, it doesn't live up to the opportunities that are created by
the internet. The system as it exists has produced many good journals, but
journals are expensive and increasingly people are reading and searching
online. There's an opportunity here to eliminate boundaries between the
individual and the information, and between pieces of information. I think
all of us were startled by the incredible power that the internet provided
for looking at and working with the genome. If we had published pieces of
genomes paper by paper we would be much less far along than we are. That
model has been a powerful force in helping people to think about how the
scientific literature can be worked with. An important issue is having
widespread searching through a public library. That's why we use Public
Library of Science, PloS, as our name. We strongly believe in this concept
to go to one place and look at everything. But the issues are many. Most of
us who are of a certain age grew up at a time when there was essentially no
science in the developing world because there was very little access to
information. One of my prime motivations is simply getting the information
that governments and other philanthropic organisations have paid for into
the hands of the people who have a vital interest in seeing it.
NS: I can already buy scientific articles online. The access is there...
Varmus: But is the access really there? You may say you'll pay for them, but
it adds up pretty quickly. Imagine you are a doctor and have encountered a
patient with an unusual disease, or a high-school student trying to write a
paper. You could identify quite a number of papers you'd be interested in
looking at. But until you look at them you don't really know whether you
want to read them. The prospect of paying several hundred dollars to
identify what you want to read is usually not very appealing . Don't you ask
yourself when you do that, why should I be paying for this when most of this
research was done with money I provided to the US government as a taxpayer?
NS: What about those who say you are just shifting costs from library
budgets to research budgets? You are now asking researchers to pay to
publish their articles...
Varmus: But that's fine. People criticise the business plan of open-access
publishers because it involves authors - and that means usually the funding
agencies the authors use - paying a fee upfront of $1000 or $1500. They say
that is a new expense. But it's not. The organisations--the funding agencies
and research organisations - paying those costs will not change. It's the
way the payment is done that's changing, and the total costs will certainly
go down. The research environment has been sustaining the publishing
industry for years. That's not a bad thing. Publishing is crucial to
research. In fact, one of the things we are trying to get across loud and
clear is that publishing has to be considered part of the cost of doing
NS: So have you got the cost basis right?
Varmus: People legitimately argue about whether we have figured the best way
to do this. We probably haven't gotten all the details quite right yet. We
wanted to show how we thought it ought to be done by producing a journal
that tested some of the principles of open-access publishing, including the
business model, and attempted to make an open-access journal that was
considered to be prestigious. This is an important point, because when you
begin to study the culture of the scientific community you find that people
are very, very sensitive to the place in the hierarchy accorded to the
journals. So if you want a job at Harvard or Sloan-Kettering or other
places, all too often it's the perception, and probably the reality, that
publishing in only three or four of the several thousand journals of
biomedical sciences - The New England Journal of Medicine, Nature, Science,
Cell, a couple of others - is virtually a requirement for hiring. Many of
the first open-access journals published didn't have that kind of cultural
credibility. We think we can achieve that for PLoS Biology by virtue of the
rigorous review we are doing, the high quality of people associated with the
journal, and our efforts to make it a journal of distinction.
NS: Is there an inherent difference between science publishing and
publishing per se?
Varmus: Sure. All of us who do science for a living make an incredible
effort to ensure that the scientific publishing industry works well. We
provide our papers for free, and our tradition is to assign copyright to the
journals, which was a huge mistake, in my view. We do editing, reviewing,
and we provide these services which in journalism or book publishing would
be work for which you would be compensated quite nicely. We all think of
this as our civic duty. What most people don't realise is that when they do
this for a for-profit publisher they are actually filling the stockholders'
NS: I talked to someone at Science, and he said: "We wish them luck, we're
also a scientific society, but we can't afford to experiment. Our journal
finances all our good works, and I'm not sure that what they [PLoS] are
doing is actually going to be sustainable"...
Varmus: Sounds like Alan Leshner [Science's executive publisher]
NS: It was...
Varmus: Alan's a good friend of mine. We had dinner a few weeks ago and
talked about all of this. I think that position is highly flawed for the
following reasons. First, Science is actually a very special case. I buy it
and I'm glad to get it every week because it's my weekly New York Times for
science. It's the product of hard-working journalists who are doing science
journalism, writing the book reviews and obituaries and editorials and many
reviews and, most importantly for me, political news about science. So it is
NS: OK, so what about other journals?
Varmus: Almost all scientific societies publish journals, which are usually
very good. My own very strongly held opinion is that scientific societies
are like guilds, they're like unions. They should serve the members, and
when they don't there's no reason to keep them going. Most societies provide
meetings, workshops, educational programmes, and these activities should be
encouraged. On the other hand, I don't believe that traditional business
plans that depend upon the sale - the inappropriate sale from my point of
view - of subscriptions to these journals should be how these societies
finance their activities. To best serve their members they are simply going
to have to adapt to the opportunities for much more efficient and useful
publications of science by the internet.
NS: Do you want to show Nature and Science and everybody else that this is a
better way? Do you want to drive them out of business?
Varmus: No, no, we're not trying to put people out of business. The ideal
solution from our point of view is to see journals convert to an open-access
NS: But you tried your petition and you didn't really convince these guys...
Varmus: The petition [www.plos.org/support/openletter.shtml] was not
useless. First of all it brought the open-access movement further along.
Over 30,000 people signed. That's not to be ignored. And a lot of journals
did sign up. We were hoping for a thousand journals and we ended up with
about a hundred, and more are coming on.
NS: But you've drawn attention to the issue, and they haven't come around
Varmus: We're really working on several different issues at once. One big
issue is the journals themselves. The private publishers are unlikely ever
to like this, let's face it. I don't know yet what they are going to do. I
think they could become open access. It's conceivable that a private
publisher could be very successful by charging, say, $4000 an article
instead of $1000. It's possible that journals will find other ways to raise
money in open-access format. But in general I think you are going to find
that the private, for-profit publishers are not going to like this because
the kind of profit margins they've had, exceeding 40 per cent - much higher
than that in some cases - simply aren't going to be tenable in the long run.
NS: Is it eventually going to be pressure that brings them around?
Varmus: Yes, I think it would be. But our real target is the society
journals. Those are the journals we think should move in the direction of
open access. And the way I'd like to see that happen is by having the other
journals say: "Look, PLoS journals are getting the best articles because
people see that not only do they do an incredible job in the editing and the
promulgation, but it's better to be published in a form that allows
everybody to see it instantaneously." And if we, BioMedCentral and other
open access journals attract the best articles and most of the articles, the
other journals will have to come around because otherwise they won't get