T H E   N I H    C A T A L Y S T     J U L Y   -  A U G U S T  1999



by Fran Pollner


Regardless of size and scope, existing biomedical publishing enterprises will surely be changed should E-biomed materialize, said panelists hailing from Elsevier Science, one of the world’s largest scientific publishing houses, with 350 journals in the life sciences; Nature, one of the most heavily subscribed science publications and home of Nature Genetics, the world’s most often cited scientific publication; and The Journal of Immunology, the 85-year-old publication of the American Association of Immunologists.

In one of the most controversial moves in his six-year directorship of NIH, Harold Varmus has offered the world the idea of "E-biomed," the provisional name of a proposed electronic publishing system for the world’s biomedical and related research. Some—mostly the scientific troops—are hailing the proposal as visionary; others—including some journal publishers—seem more disposed to burn it at the stake.

E-biomed would use the Internet to provide universal access to all published reports in the reviewed scientific literature—not just the title or the abstract, but the full report—free, with no subscription, no fee, no license needed to enter. It would also allow scientists to post essentially unreviewed papers in a separate section.

Recently, as the E-biomed protagonist on a panel that included Mary Waltham, U.S. president of Nature, Karen Hunter, senior vice president of Elsevier Science, and Michele Hogan, executive director of the American Association of Immunologists (AAI), Varmus tried to ease some qualms.

"None of us believes journals will disappear in the near future. We will go to our graves carrying our journals," he said. "And we don’t view the [existing publishing] system as broken—but as suboptimally used," he added, speaking for himself and his collaborators in the E-biomed idea, most notably David Lipman, director of the National Center for Biotechnology Information at the NLM and the architect of PubMed and GenBank, and Pat Brown of Stanford (Calif.) University. "Within the E-biomed system," he maintained, "journal identity and hierarchy will persist."

"Money," said Hunter of Elsevier, which publishes 350 life sciences journals, including Lancet and Cell, "is at the heart of a lot of the concern. There’s really a lot of desktop access now. The distinction is whether it’s free or not." Elsevier’s journals, she said, have been available on the Internet since 1991—with more than 3.5 million users and nearly a million articles in its database.

Waltham observed that the scientific publishing community had already begun to embrace Internet publishing and its attendant licensing complexities and predicted that "with or without E-biomed," free, wide access would eventually emerge. "It was a slowly rolling ball," she said, but a "firm kick" has sent it flying into cyberspace—"hopefully in the right direction," she cautioned.

The "firm kick"—"E-biomed: A Proposal for Electronic Publications in the Biomedical Sciences"—was launched May 24 from Varmus’ NIH web site and within a month had compelled hundreds of people around the globe to respond. Some responses were terse, some elaborate; many bordered on hyperbole to describe their support, opposition, or reservations. (See <http://www.nih.gov/welcome/director/ebiomed/ebiomed.htm> for the proposal and links to posted comments and a June 20 addendum responding to repeating themes within the comments received.)

The panel discussion, which was arranged by a science writers’ group and held at Lister Hill the last evening in June, did not reflect the range of opinion expressed online, the greater part of which was exhilaration, praise, and offers to participate. Rather, many of the negative reactions and misperceptions posted at the web site—and answered online by Varmus and his colleagues—were echoed by panelists and members of the audience. These were largely concerns about the undermining of peer review and the quality of scientific literature, dangers attending unreviewed clinical research reports on the web, bankrupting existing publications, the costs of E-biomed, access disparities among "have and have-not" nations and individuals, and NIH’s assuming monopoly control over biomedical publishing.

The proposal itself commits NIH to provide "financial, technical, and administrative assistance to initiate" an E-biomed program. It emphasizes the need for extensive partnering with existing journals and for international collaboration.

Two Paths to E-Print

The proposal envisions speed, diversity, and expandability as concomitant benefits of a two-tiered system for getting research reports published electronically and disseminated worldwide. One route duplicates in cyberspace what currently exists in paper: Investigators submit their manuscripts to the journal of their choice (if the journal of their choice has indeed opted to participate in E-biomed); the journal’s editorial board then reviews the submission in its usual rigorous manner. Stringent editorial board review would apply as well in the case of any new electronic-only journals created specifically to exist in the E-biomed environment.

The second route involves no peer review—only minimal screening to ensure that the subject matter of the submitted report is appropriate for the web site and contains nothing pornographic, libelous, or otherwise outrageous. It would then pass into an unreviewed repository of biomedical reports. "Negative" results, technique refinements, and other writings not considered a priority for publication by many journals could thus be made available to an interested audience with otherwise limited means of exposure to the information.

During his panel presentation, Varmus emphasized the fluid nature of the medium as a boon to scientific reporting. Free of the space and two-dimensional constraints of print, investigators could embellish their electronic reports with color photographs, charts, movies, and extensive datasets—presented in layered fashion, so that readers with greater interest might with a simple click go deeper into the subject. Also, critical comments, revisions, and retractions from readers and authors could be appended to papers.

Dissemination and Control

Perhaps the greatest appeal of E-biomed to authors is that it affords the widest dissemination possible of their findings–the essential objective of science reporting. "Scientists don’t make money from publishing their results; they want their results to be seen by as many people as possible," Varmus observed.

Journals that decline to participate in E-biomed for fear of loss of revenues may find that authors opt to submit their manuscripts elsewhere. Smaller specialty journals that may be the mainstay of a professional society’s coffers will be placed between the proverbial "rock and a hard place," Nature’s Waltham later said. "If they join [E-biomed)], they’ll lose subscription fees; if they don’t, they’ll lose their manuscript stream."

Hogan of AAI, which has published The Journal of Immunology since 1915, warned that publications will "go under" and that NIH could become the "sole supplier of scientific content," with the consequent "loss of peer review independent of government" and "inconstant federal funding in lean times."

Varmus reiterated, "We are not the owner or publisher" of E-biomed. "No one," he added, "worries about ‘takeover’ with GenBank or PubMed—everyone appreciates them, and E-biomed would be the same thing."

The success of E-biomed depends on stringent peer review by participating journals, he said. He, noted, too, that the quest for international partners has met with considerable interest from the European Molecular Biology Organization, the European Molecular Biology Laboratory, and the European Bioinformatics Institute.

Repository Contradictions

The nonreviewed or repository component of the E-biomed proposal drew contradictory reactions. It was viewed, variously, as a nonthreatening convenience and as a potential source of "criminally inaccurate junk." Both appraisals stemmed from its similarities to an existing database established by physicists in 1981, based at the Los Alamos National Laboratory, and supported by the National Science Foundation and the Department of Energy (see <http://xxx.lanl.gov>).

An electronic archive and distribution server for research papers in physics and related disciplines, its contents are determined by the scientists who use it. Authors submit papers at will and may update them as they choose; users can register to receive notice of new submissions in their interest areas.

Several individuals pointed out that this "preprint archive" has been happily coexisting with peer-reviewed physics journals for nearly a decade to the benefit of the physics community and the detriment of no one. And panelists Hogan and Waltham declared that their publications would have no problem with the preprint, or repository, aspect of E-biomed, although Hunter observed–and others agreed–that the communities of physicists and biologists are different.

The nature of their research reports is also different, individuals noted both during the session and in responses sent to the E-biomed web site. Many considered unreviewed reports in the realm of clinical research a potential hazard to the public health, especially if they are seen to have the "NIH imprimatur." They also feared the repository could be used to promote commercial interests.

Varmus agreed that flawed information in clinical areas could be a problem, but noted that most scientists would not willfully associate themselves with inaccurate data and that the repository material–unlike much other health information on the Internet–would clearly state its source and be conspicuously labeled "nonreviewed." Moreover, by its very nature, the repository allows for speedy commentary and follow-up by other researchers and by the original authors—and, therefore, for online critical review by experts in the field.

Overall, the legitimate benefits of the repository outweigh its pitfalls, he said, and different measures might be taken in its pilot phases to minimize the pitfalls, perhaps by limiting the subject matter initially.

What Price E-biomed?

The bottom line and who would pay was perhaps the most contentious subject of the evening.

NIH currently pays more than $100 million for subscriptions to journals, library copying, and other costs related to "getting our results out to the public," Varmus said. He estimated that setup and operating expenses to be shared by NIH and its E-biomed partners would be about $2–$3 million.

It would cost between $200 and $1,000 to process an article in the E-biomed system, including the costs of distributing manuscripts to reviewers, collating comments, and coding and scanning for the web, he said. Authors could pay that cost—with a small fee paid to the journal upon submission and a larger fee upon acceptance. The author, ideally, would retain copyright.

Hogan maintained that moving from print to electronic publishing would not lower costs, as Varmus contended, but merely shift them. Hunter argued that not all nations are electronically advanced and not all authors can afford the fees. Varmus observed that a fee waiver would likely be part of the E-biomed apparatus. He argued that "Net connections—which NLM is helping set up [in other parts of the world]—are cheaper and faster."

Loss of library subscriptions and loss of licensing fees for access to electronic versions of journal articles was viewed, however, as a death knell for many publications, especially those with a narrow focus. And well-heeled journals that opt not to join E-biomed could conceivably sustain a large enough circulation decrease to affect advertising revenues.

"We’re heavily dependent on ads, which subsidize our low sub[scription] rates," noted Nature’s Waltham. "The journal price would have to go up if the ads go down—which is why we watch this initiative with interest."

What Now?

Waltham later said that Nature’s participation in E-biomed could be "negotiated," based on "conditions of fairness" she declined to elaborate. Until then, she said, "we can afford to be cautious. We don’t need to be the first ones in." The journal’s huge subscription base, multidisciplinary coverage, and "value added" front half—the news, views, and commentaries—will keep its authors and readers loyal, she said.

She surmised that E-biomed will indeed be launched but hesitated to predict how "biologists will behave" in the face of it.

Varmus told the group that he and his collaborators will proceed "relatively slowly," initiating E-biomed in the "next year or two." But he indicated that the project is in motion. "We’re talking to some adventurous souls who’d like to create a new E-biomed journal, and we’re also starting work on the repository," he said.

Suggested Composition of an E-Biomed Board
The E-biomed proposal enlists a "Governing Board" (later referred to as an "advisory board" in the addendum) that would oversee general E-biomed operations, policies, and rules regarding submissions of unedited reports into the E-biomed repository—not the operations of the individual editorial boards within E-biomed. Its composition and the boundaries of its authority are included among a list of issues to be resolved within the international scientific community. During his presentation at Lister Hill, NIH Director Harold Varmus suggested this blueprint for the board’s composition:
Geographic Representation Special Interests
United States 4 Scientific Societies 3
Europe 4 Commercial Interests 3
Asia/Pacific 2 General public and advocacy groups 3
Africa 1 Libraries 2
South America 1 Sponsors (like NIH): ad hoc


Return to Table of Contents