Will NIH's new Web site change medical publishing?
From the January 2000 ACP-ASIM Observer, copyright © 1999 by the American College of Physicians-American Society of Internal Medicine.
By Phyllis Maguire
As the National Institutes of Health (NIH) prepares to unveil its PubMed Central Web site this month, many are expecting an anemic version of online medical publishing that was at one time viewed as revolutionary. Fierce opposition from journal publishers last year led the agency to scale back its proposal to archive biomedical articles and to eliminate some of the project's more controversial features.
As the site prepares to go live, however, it looks like PubMed Central may still help transform the world of medical publishing. While the site's main intent is to serve as a repository for articles from existing medical journals, it is busy exploring new types of publishing activities. Besides helping create a new breed of online-only publications, the site will also host medicine's first "preprint" service for material that has not been peer-reviewed.
While PubMed Central is expected to start out by putting less than a dozen of the world's 4,000 plus biomedical journals online, it has generated excitement among researchers and physicians who want better access to cutting-edge clinical research, and among libraries struggling with exorbitant subscription costs. The big question now is how traditional medical publishers, which currently control the flow of most medical findings, will react.
New publishing models
When the original version of PubMed Central was first proposed last spring by NIH director Harold E. Varmus, MD, the concept was revolutionary: Transform medical publishing by posting all biomedical literature online. Articles would be submitted to the NIH to be peer-reviewed by specific journals and then posted online at the same time they appeared in print. The free site, initially named E-biomed, would publish studies on clinical and lab-based medicine and bioengineering, as well as on cell, genomic and molecular biology.
E-biomed would also create another publishing mechanism where articles could be posted directly on the Web without peer review, much like physicists and other scientists post research "preprints" on the Los Alamos National Laboratory server. Under this type of model, authors, not publishers, retain the copyright to their articles, allowing them to freely distribute their work.
While E-biomed was hailed by researchers, medical publishers quickly attacked the idea. Critics charged that while researchers in basic science may benefit from sharing unvetted research, posting clinical information that had not been peer-reviewed could hurt patients. The College urged the NIH to proceed with "great caution" and warned against bypassing the quality-control mechanisms used by existing medical journals.
The NIH allayed those concerns by recasting its site as the partner, not the competitor, of established journals. The new Web site, renamed PubMed Central, would not establish its own editorial board, nor could authors post their own material. Instead, the site would rely on existing journals to submit peer-reviewed material. Resistance from publishers seemed to fade.
Now, however, it appears that some of PubMed Central's original innovations remain. The site is now affiliated with the Community of Science (COS), a for-profit spin-off from Johns Hopkins University that hopes to help create a new type of virtual journal. COS plans to electronically manage all pre-publication processes—such as article submission, peer review and formatting—for new start-up journals or for existing journals that want to publish online exclusively through PubMed Central.
COS is targeting several distinct markets. One segment consists of small medical societies that now publish journals through commercial publishers. Their members have long complained about underwriting high printing and distribution costs with pricey subscriptions.
Another potential partner is any group of academicians, researchers or clinicians who want to form an editorial board. "We're hearing from smaller societies that haven't gone into full-text publishing because of printing costs," said a COS spokesperson. "Our system will allow them to publish."
The preprint debate
Another controversial PubMed Central feature that survived opposition from publishers is the preprints section. This feature will allow researchers to share data that are not traditionally published, like negative findings and failed clinical trials.
All preprints must be screened by NIH-approved editorial boards, which will not be limited to reviewers from only print journals. In addition, non-peer-reviewed material will be clearly delineated from PubMed Central's peer-reviewed component.
Critics still worry that the information will be misused. Some are concerned that such a repository could be flooded by marketing information masked as research data, or that the consumer media will seize upon sensational and unproven findings and create havoc among patients.
Supporters, on the other hand, think access to research that produces negative results—a new drug that does not help patients with heart disease, for example—is critical. "When you're dealing with clinically relevant problems, very often the negative results or failures are even more important than the successes," said Anthony S. Fauci, MACP, director of the NIH's National Institute of Allergy and Infectious Diseases.
Others believe that such access may revolutionize the ways in which scientific information is used. "Now, the whole scheme of depositing scientific literature is based on traditional publishing, with its limitations of size and pages," said Vitek Tracz, chairman of the London-based Current Science Group, which is posting primary biomedical journals on PubMed Central and organizing a new virtual endeavor called BioMed Central to review PubMed Central material. "With PubMed Central, you can deposit original data in a form that scientists can add to and work on." Mr. Tracz believes that the preprint service will particularly benefit an area of scientific publishing that is consistently inadequate: clinical trial data.
PubMed Central's proponents also believe the site might help accelerate the pace of clinical research by releasing study results more quickly. They also hope that practicing physicians will benefit from services—such as state-of-the-art evidence-based clinical guidelines—that can be developed from PubMed Central material.
Reaction from publishers
While the medical community debates the merits of PubMed Central's features, a larger question remains: How will existing print journals respond to the NIH venture?
While most medical journals are already on the Web in some form, they typically give access only to print subscribers. (Medscape's MedGenMed and the British Medical Journal are two of a small minority that post the full text of their content online for all to see.) PubMed Central's model clearly conflicts with the business model that most journal publishers use.
At stake are hundreds of millions of dollars that journals earn each year in advertising and subscription revenue. In a letter to the NIH, the International Committee of Medical Journal Editors last year acknowledged that such an electronic enterprise could "scoop" content and damage the value of their print journals and Web sites.
Commercial publishers are ignoring PubMed Central for the moment, although the NIH project may be inspiring some companies to develop more aggressive Web strategies. Twelve commercial publishers—including giants Elsevier Science and Harcourt's Academic Press—recently announced that they will develop links among their 3,000 journals later this year, allowing users reading one journal to get information from another publication with the click of a mouse. Those links will provide only abstracts, however, and users will still have to pay to see the full text.
The major medical journals, on the other hand, have showed a mixed response to PubMed Central. The Journal of the American Medical Association has not yet decided if it will take part in the site, while the New England Journal of Medicine has stated that it will not share material, citing concerns over the non-peer-reviewed section.
The Annals of Internal Medicine is seriously considering participating, although no final decisions have been made. "I think it would be silly to sit on the sidelines," said Frank F. Davidoff, FACP, the College's Senior Vice President and Editor of Annals. One possibility would be to post the full text of original research for articles that appear in Annals on PubMed Central. Other parts of the journal like review articles, editorials and updates, however, would still be available only to members through print and online versions.
NIH officials are taking a wait-and-see approach, saying that the Web site is designed to change according to the needs of researchers and physicians. "We really are not proposing a specific model," said David J. Lipman, MD, PubMed Central's ad hoc director and director of the National Center for Biotechnology Information within the NIH's National Library of Medicine. "Instead, we're opening it up and seeing how the medical community responds."
The project's sponsors hope that such flexibility will be the key to PubMed Central's success. Authors who now vie for space in the most prestigious journals may find that PubMed Central can streamline publication, getting their findings out fast and free to physicians, researchers and consumers hungry for medical news. They may thus come to prefer new virtual journals that PubMed Central can make available.
"That could be a possibility, of course," surmised Elizabeth A. Pope, a staff scientist with the National Center for Biotechnology Information who coordinates PubMed Central submissions. "That's something that the end user is going to have to determine."
Internist Archives Quick Links
Superior MOC Solutions from ACP
Meet your requirements with our approved activities. See details.
Making the Most of Your ICD-10 Transition
To help you and your practice make a smooth and successful transition to ICD-10 coding, ACP and ICD-10 content developers have created multiple resources available at discounted rates for ACP members.