Just in time for the new century, the National Institutes of Health (NIH) quietly
launched an innovation that promises to do for biomedical research what the Gutenberg
Press did for literacy. The project is called PubMed Central, and its premise is
disarmingly simple: to bring the widely scattered findings of the world's life science
research together under a single electronic roof, searchable by anyone.
To be integrated with the existing PubMed biomedical database, PubMed Central is now
online at http://www.pubmedcentral.nih.gov.
The historic undertaking is the brainchild of Nobel laureate Harold Varmus. As NIH
director from 1993 until 1999, Varmus's idea was to create a virtual library where anyone
with a computer - doctor, farmer, biochemist, flu sufferer - could search for
comprehensive and current information on the latest medical research, even if some of it
hadn't yet been through the laborious peer review and publication process.
By making connections across disciplines and cultures, this huge new public database
promises to be a pivotal event shaping the life sciences, where explosive growth and
dramatic breakthroughs are on the near horizon. The pace and magnitude of today's research
is illustrated by the Human Genome Project, the global effort to document all human genes,
which last year released roughly five million new DNA base pairs every day. In scientific
discovery, the traditional image of the lone genius at the lab bench has given way to that
of a networked team of researchers collaboratively pushing the scientific edge.
Varmus's concept is aimed less at consumers than at professionals. He hopes that PubMed
Central will set off a cascade of new discovery in life science. And not just in the lab.
Through PubMed Central, data from field and clinic will interact more easily with
laboratory research, continually improving both the research and its applications.
"You might suddenly recognize, by looking at a gene sequence, that someone's study
of an arcane process in fruit flies is the richest source of insight into a problem of how
blood vessels develop," says Patrick O. Brown, a genomics scientist at Stanford
University who consulted with Varmus on PubMed Central. "More and more scientific
research involves making connections across totally unpredictable distances in the body of
knowledge. Suddenly all the Balkanization of the scientific literature becomes a major
obstacle. You don't have, in the small set of journals in your office, all the information
you need."
Unfortunately, that small set of scientific journals is what most researchers currently
have to rely on. Although some of the journals are online and searchable, the lag time
between discovery and release of the data is often six months or more. Traditional journal
publishing is also expensive, despite the fact that the research and peer review are
mostly done by scientists at public institutions or under public grants. Journal
subscriptions can run in the hundreds of dollars, and authors may pay as much as $1,000 to
publish their articles. By contrast, through PubMed Central, even a scientist in a
developing country will be able to present his or her work to the world, and in turn, have
access to a vast scientific resource for the cost of a computer and an Internet
connection.
Top university libraries, for those lucky enough to be near one, do stock many highly
specialized journals. But any given article can take hours to track down. In today's
bio-research environment, this is the equivalent of using the Pony Express to deliver
mail.
Brown envisions scientists using PubMed Central to construct personalized, continuously
updated journals with automatic searches and feeds for experiments related to their fields
of study. Much of the information they would use will be in Part I of the site, which will
consist of traditional, peer-reviewed journals that agree to participate.
Part II of the site is more controversial. For that section, a board of scientists will
establish criteria for publishing reports outside of the peer-review structure. The idea
is to lower the publication barriers to breaking research reports or even incomplete sets
of data - thus bolstering the new collaborative methodology.
Brown believes the review procedure for Part II should be "a simple screening
process that would only exclude stuff that on the face of it is not science." He
points out that scientists routinely rely on articles that are not peer-reviewed and that
those who would look at these papers are self-selected people who will evaluate the data
with a critical eye.
That doesn't wash with everyone, especially some of the journal editors and scientific
societies. They see PubMed Central as a threat to the time-honored process of careful peer
review of published articles - and a threat to their livelihood. A major shot across the
bow of the PubMed Central proposal came from the prestigious New England Journal of
Medicine last June in an editorial by one of its senior staff, Dr. Arnold S. Relman of
Harvard Medical School.
Relman wrote that he is concerned that health care providers will use unreliable
information from Part II and that free publication of papers on PubMed Central "would
very likely reduce the submissions, paid circulation, and income of most clinical journals
enough to threaten their survival."
Varmus believes the fear of clinicians' using unreliable medical data is misplaced.
"The results posted...that are most likely to affect health care will have been
subjected to critical review and editing by journals," he wrote in a discussion on
the NIH site.
The more likely outcome, advocates argue, is that PubMed Central will radically improve
physicians' practice by addressing one of the glaring faults in today's health care
system: the absence of "best practices" information in the hands of those
treating patients. By making information on the best clinical practices widely accessible
and by expanding "bench to bedside" collaborations between researchers and
clinicians, PubMed Central can help close the sometimes tragic gap between the state of
research and the state of practice.
The database could also work in the other direction: bringing the unparalleled daily
experience of clinicians into the raw data pool available to researchers. "Every
single day, detailed physiological measurements and careful observations are being made by
physicians about anything that's anomalous. It's a huge raw resource for scientific
discovery," says Brown, the Stanford researcher, who hopes PubMed Central will
encourage doctors to report their findings. "Most of [that data] just
vaporizes."
Moreover, journals' prestige and value may actually increase with PubMed Central as
they spend more effort analyzing and filtering the information in the vast database,
making it more useful to researchers, clinicians, and even members of the public.
Certainly, some of the kinks in PubMed Central will have to be worked out, especially
when, as Varmus hopes, it achieves greater coordination with international databases. The
site is designed for continuous improvement, and many feel that PubMed Central's benefits
will quickly outweigh its problems.
However the controversies and details are resolved, PubMed Central will surely
revolutionize the ease and speed with which research data moves around the world.
Balkanized data will be assembled and made accessible. And that, in turn, could have
direct and profound implications for treatment and outcomes.
