Comment on E-Biomed Supplement

A supplement to NIH's E-Biomed proposal appeared on June 20.
Articles in SCIENCE (June 20), THE SCIENTIST, and elsewhere
seem to indicate (A) the biomedical community has clearly
not embraced the proposal; and, (B) Dr. Varmus is intent on
going ahead with what might well be a monster more
destructive than Frankenstein's.

What is the rush, Dr. Varmus? The technology will not
disappear. There is no competitive international crisis, as
there was following Sputnik. No scientific study provides
evidence that E-Biomed will cure or even treat complaints
about the performance of NIH. Nothing in the new material
explains its urgency.

I see three issues of paramount importance that must be
resolved before E-Biomed goes further: 1.dangers posed by
economic disruption and lower standards; 2. the desire to
improve the productivity of NIH programs; 3. the appropriate
role of government, specifically NIH. Within these there
are legal issues that cannot be casually ruled out.

I also offer what I feel are better ideas.

1. Economic Disruption and Lowering Standards:

"Offering the international scientific community free, fast,
and full access to the entire biomedical literature," which
E-Biomed claims to be its most important goal, will have an
economic impact. The the proposal avoids this issue. NIH's
outline of "an electronic public library of medicine and
other life sciences" leaves major questions unanswered:

   Will E-Biomed hurt libraries and publishers?

If institutions' researchers have unpaid access to all the
literature on E-Biomed from their personal workstations,
why should any university, corporation, or other
organization continue to maintain libraries that purchase
journals and provide services far beyond, but closely
related to, their collections?

     Will associations of scientists survive?

If the entire literature is available on the internet
free, why would scientists and students join associations
and pay dues? It is true some wish to participate in
meetings and other activities, but publications have
usually been the primary benefit of membership.

The supplement appears to acknowledge that cost-free
distribution of articles would undermine the viability of
journals. It suggests, "the editorial board [of each
journal] would need to consider the means available for
recovering the costs of reviewing ... from annual
meetings, from workshops ..., or from increased annual
fees."  With these words, NIH cavalierly evades concerns
of the associations and its responsibility for the likely
economic devestation of E-Biomed. It dismisses the
dependence of journals on library subscriptions to cover
first-copy costs and more. NIH's solution is unrealistic,
perhaps even preposterous.

In the context of the "War on Faculty" [Chronicle of Higher
Education. April 16, 1999 B4], "The Treason of the Learned"
[Library Journal Feb.15, 1994 130-131], and similar
observations going back to the 1970s, it seems clear that
E-Biomed serves the agenda of university managers wishing
to shed the trappings of intellectual inquiry -- tenure,
libraries, academic freedom, associations, copyright,
publishers, and publications -- that interfere with
nonacademic interests in bureaucratic power. Over the
objections of academic senates universities have slashed
library spending. Since 1984 higher education spending on
libraries has been less than total unspent revenues. For
decades prior, library spending was greater. Universities
have never had so much money. The trouble is, they have
become greedy.

        What about the Sherman Anti-Trust Act?

"Every person who shall monopolize, or attempt to
monopolize, or combine and conspire with any other person
or persons, to monopolize any part of the trade or
commerce ..."

Based on the present evidence I would be inclined to vote
'guilty.' The supplement says the system, "would not be owned
by the NIH or any other component of the U.S. government." By
whom would it be owned? How would it not monopolize, through
predatory pricing, the scientific publishing industry after
330 years of private sector operation? How would it not
emasculate the hundreds of scientific associations that
are involved not only in publishing but in accreditation,
policy, and oversight of the various disciplines and
specialties.

   Will E-Biomed eliminate Index Medicus and Medline?

With its dependence on what appears to be a totally
automated search engine, E-Biomed appears to have dispensed
not only with libraries but with the NLM and its labor-
intensive services. I seriously doubt that an automated
search engine can adequately do the job of Medline, simply
based on the preference of many authors and readers for
unpredictable jargon.

   Wouldn't a better idea be to upgrade Medline?

Medline's coverage might be more sophisticated. I have a
report that indicates 10,000 Medline cites were screened
to locate about 400 articles related to "whiplash." More
labor-intensive indexing might eliminate the monumental
challenge that the literature makes to the researcher
before reading and evaluating a single article. The same
report indicates that less than 100 articles survived
tests for scientific merit applied by teams of specialists.

Medline's coverage might be wider. Medline, a true pioneer
in electronic publishing and dissemination, indexes over
400,000 articles annually. In spite of the impressiveness
of that number, it covers less than a fifth of over 22,000
periodicals received by the National Library of Medicine.
It ignores thousands of nonperiodical items received.

Medline got the blame for incomplete and inaccurate reports
of prior scientific work, real threats to the integrity of
science. Apparently, "searchers often go back only a few
years (in medicine, for example, three years -- the length
of time covered by Medline, the biggest reference data
base)." When searchers fail to cover the literature fully,
proposals, merit review, and the sponsors all suffer.
[Chronicle of Higher Education. April 21, 1995:B1-B2]

Many researchers tell me they prefer more specialized
databases that offer comprehensive coverage and analysis.
Maybe a better idea would be for NIH to foster more
specialized resources by improving research overhead for
libraries and improving the commercial market for
research information.

  Why is lowering standards of access good?

The system of fees and licenses is restrictive. So is the
formal publication process requiring peer review. It limits
the published literature and its readership to qualified
individuals. That is good, not bad, according to the most
authoritative comments posted so far. Dropping restrictions
as proposed may put E-Biomed closer to talk radio. The
reference to Paul Ginsparg's experience has little relevance.
Physics and math are relatively small fields. Errors in
physics theories are unlikely to affect anyone's health.
Snake oil cures and other quackery are typical of medicine,
not physics, creating a special need for rigorous oversight
by the Food and Drug Administration. The jurisdiction of
Federal "scientific integrity" policing extends only to
Federally financed researchers. Others are free to report
whatever they like, probably with the protection of the
First Amendment. Relying on researchers' integrity alone
does not seem to be realistic.

As the prologue to E-Biomed points out, electronics made
communication easier than ever before. Informal exchanges
via email have supplemented, perhaps replaced, more
traditional channels. This has reduced, not aggravated
chronic problems in science communications. No urgency has
been expressed by the research community for NIH's solution
as far as I can see. When an 'invisible college' wishes to
set up an exchange of informal documents in its own
specialty, perhaps protected by passwords or limited to a
qualified list, no one objects.

Many publishers and others (known as aggregators) provide
electronic access to their journals, including to backlogs
of unpublished accepted papers. I think that the
mainstream journals are nearly all available online.

True, journals are not free. Why is that problem? The
economy is good. Universities have more unspent revenue
than ever before! (according to U.S. Dept. of Education
and other statistics). If you are concerned that
researchers are spending direct grant funds on
publications, why not improve overhead reimbursement of
their libraries and insist that university upgrade their
collections?

This is a better idea. Federal research commands sixty
per cent of sponsored academic research expenditures.
Why does indirect cost reimbursement barely cover ten per
cent of library spending? Why was library spending cut
while research spending grows? If library spending had kept
pace with the growth of research since 1970, it would be
more than double.

2. Productivity and performance:

One of the most disturbing statements in the original
proposal is not mitigated by the supplement. It reads:

        The active E-Biomed process might be
        accompanied by a much-needed effort to
        convert material already published on paper
        to digital text and image format, with
        hyper-linked citations. This additional
        initiative would ultimately allow all users
        of E-biomed to move seamlessly through the
        entire body of reported information in
        biomedical sciences. And it would also
        enhance scientific productivity and reduce
        burdens on library facilities.

If NIH wishes to enhance scientific productivity, NIH must
improve peer review. Within the issue of productivity, the
fundamental weakness of E-Biomed is that it simply
generates millions of documents, forcing the reader to
read, comprehend, and decide. The addition of unreviewed
articles to the formal findings and other items indexed by
Medline annually only aggravates the problem.

The Achilles' Heel of merit review is the degree to which
the referee is not informed. Because of the growth of
science, a review may require several specialists. E-Biomed,
as proposed, asks every author and every referee to be
responsible for the entire primary literature. If NIH
wishes to improve productivity NIH must emphasize evaluation
of reported research, synthesis, and commentary. That would
probably  mean increasing the burden on library facilities
and increasing support for library research.

Aside from scientific productivity, the passage reveals
a careless attitude toward copyright. This comes at a time
when the U.S. government and industry are making constant
efforts to protect intellectual property abroad. Two thirds
of the science literature is authored outside the U.S. and
is protected by foreign copyright and international
copyright agreements. Of the balance of articles signed by
U.S. authors, a fraction -- perhaps 15 per cent of the
total literature -- has a direct connection to NIH.

In scientific publishing, the power of copyright to
encourage investments has produced an endless stream of
innovations. Even during the height of the Cold War, when
it is doubtful that any court would have heard a suit
for infringement, the Soviets' rights to translations of
their journals were respected by U.S. publishers. A
better idea than E-Biomed would be for NIH to encourage
researchers' institutions to upgrade their libraries and
to provide a commercial market hospitable to investment.

3. Ethics and policy.

The government built mass transit systems. Everyone
benefits when mass transit is used. Why not go whole hog
and give free rides at taxpayer expense? I think the answer
to this must also apply to E-Biomed.

Instead of a radical Soviet-style intervention, why not
foster existing publishers and libraries? The U.S. beat the
Soviets to the Moon and otherwise by encouraging private
sector initiatives in information (even though the Soviets
beat us into orbit 10 years earlier). The government
rejected investing in a 1958 plan similar in many respects
to E-Biomed.

The minute we got to the Moon, however, universities savagely
cut their library budgets, heading toward ratios common
before World War II. During the pre-War period, I am reminded,
mainstream science and culture was in Europe. The U.S. sat on
the post-Colonial fringe. The potential for return to this
era through the decimation of university libraries is a
problem that NIH would do well to study.

Who discarded the expectation by government planners of the
1940-50s that universities would conserve the knowledge
produced by investments in research? University library
collections are now inadequate, apparently with the tacit
blessing of science  agencies like NIH. Now NIH and LANL
wish to deliver the coup de gras.

Ginsparg mounted his experiment by fiat, in the absence of
public comment, as far as I know. Having proved that the
public will consume information distributed free at
taxpayer expense, Ginsparg can tell us little more than
we knew before. The Ginsparg precedent is dangerous in that
it may encourage other government agencies to interfere
with the private sector without reference to policy or the
public interest. Having finished its experiment, LANL's
role should end. It should turn its technology over to the
private sector and get out of the document delivery
business. If selling unreviewed physics articles remains
viable, it may yet prove its value.

The earliest code of medical ethics required physicians
above all to do no harm. One of the most distressing
aspects of E-Biomed is that the damage it will probably do
to libraries, publishers, and the research community will
be unrepairable. There will be no going back once the
universities have shut their libraries and publishers are
out of business.

Albert Henderson
Editor, PUBLISHING RESEARCH QUARTERLY
POB 2423 Noble Station
Bridgeport CT 06608
<70244.1532@compuserve.com>