In the elite world of medical research, Dr. Harold Varmus is at the top of the heap. He runs the government’s biggest health research engine, the National Institutes of Health, and won the 1989 Nobel Prize for his groundbreaking work on cancer genes.
Yet Varmus, 59, has proposed such a radical, power-to-the-people idea involving Internet publishing that the rest of the medical establishment is at his throat – partly out of self-interest, but partly out of genuine concern that Varmus’ plan would add dramatically to the chaos of medical information and misinformation already on the Net.
Aided and abetted by computer whizzes at the National Library of Medicine, Varmus wants to do the unthinkable: Put the full text, not just abstracts, of research papers on the Net in an expedited way, and – here’s the grabber – in some cases without the peer review that journals like the New England Journal of Medicine and the Journal of the American Medical Association have long used as a way to weed out solid studies from schlock.
This is cage-rattling stuff – not just because it might mean scooping journals on medical news, but because some scientists might choose to bypass the traditional journals altogether in favor of speedier publishing – with less rigorous review – on the Net, all under the NIH imprimatur.
Speaking by phone from New York City last week, Varmus, who will leave NIH at the end of the year to head Memorial Sloan-Kettering Cancer Center, said he is and always has been a believer in peer review. This is the process by which, before accepting an article for publication, medical editors send it to outside experts who pore over the data looking for errors in methodology, science or interpretation.
And initially, all of the research that Varmus envisions for the new site, to be called PubMed Central, just posted last Thursday and scheduled to be operational in January, will indeed be forwarded from journals that have vetted the work.
This research – in biology, medicine, agriculture, and plant sciences – could be posted after publication in a major journal, which nobody objects to. Or before it appears in print, which critics say could make mainstream medical journals obsolete.
“This would be a surefire method of killing ourselves,” says Dr. Marcia Angell, interim editor-in-chief at the New England Journal of Medicine. If the Journal took the time and care to review articles, then handed them right to PubMed Central, by the time the Journal published the work “it would be just an archive. . .we would be committing suicide.” As it is now, the Journal posts its new research on its own Web site the night before the print publication becomes available.
And then there’s the even more controversial, back-door route to getting research in the public domain fast that Varmus has developed with David Lipman, director of the national center for biotechnology information at the National Library of Medicine.
The idea, still a work-in-progress, is to have a “screening” process in which papers that might never be submitted to a mainstream journal would be posted on the Net if they were forwarded by a “certifying group.” This non-peer-reviewed material would be clearly labelled as such.
A still-to-be-created committee would figure out who would qualify as a certifying group, but initially, this would be any organization with at least three members who are principal investigators on research grants from agencies such as the NIH, the National Science Foundation, the Department of Energy, etc.
The point is to “develop a way to disseminate research results in a way that allows anyone access any time, any place,” says Varmus, who dislikes the “gatekeeper” role that major journals now play.
The goal, he says, is fast, “barrier-free” access to medical research. No passwords, no need to “register” on a Web site to read a research study, no fees to Web surfers. In fact, in PubMed Central, the costs of publishing could be picked up by the authors of papers – whose funding often comes from taxpayers anyway – instead of those who want to read the work online.
Some people love this idea. Among them is Barbara Lackritz, a St. Louis, Mo. cancer patient and patient advocate who says that to keep up with the latest research, she needs more than the abstracts most Web sites offer. “The more information we can get out there for people to read and understand, the better off we are all going to be,” she says.
Dr. Helga E. Rippen, a physician-engineer who runs a health information think tank at Mitretek Systems in McLean, Va., agrees. “I’m into as much information as somebody wants,” she says, though she adds there’s also a need to teach consumers how to evaluate and interpret complex, raw medical data.
But Dr. Sidney Wolfe , director of Public Citizen’s Health Research Group, a health research-based advocacy organization in Washington, D.C., is leery about rushing unreviewed research into the public domain. “Needless to say, the sooner people get information, especially information with public health impact, the better. But it’s only better if it’s accurate.”
Biased, incomplete or incorrect medical information can do serious damage, he notes. Though physicists have posted preliminary research results on the Net, there’s considerably more risk when medical researchers do so, he adds.
And Angell, of the New England Journal, is appalled. “I believe in the marketplace of ideas,” she says. “But you can’t just throw up everything – advertising, old wives’ tales, research studies – on the Net and hope people can figure out what’s valid and what isn’t.”
For his part, though, Dr. George Lundberg, former editor of the Journal of the American Medical Association, is gloating.
Now editor of Medscape, an online health information service with access to 25,000 research articles, and its spinoff, Medscape General Medicine, an electronic medical journal, Lundberg says it is possible to put peer-reviewed medical information on the Net without the long delays in standard medical publishing. (At the New England Journal, for instance, it takes seven months from the time a paper is submitted to the time it is published.)
In fact, in a way, he’s beaten Varmus to the punch. Medscape, he says, does peer review its articles, but fast – in three to five days, not the 28 that print journals take.
So far, only a handful of existing journals have agreed to supply material to PubMed Central, including the National Academy of Sciences, the Journal of the American Society of Cell Biology, the Canadian Medical Association Journal, and a British entrepreneur who publishes biomedical journals.
But that should change as scientists realize the potential for getting work out quickly, says Jo McEntyre, a fellow at the National Library of Medicine. “Scientists love it,” she says, “because they are the ones who do the research.”
Because of the sheer volume of medical research, many scientists can’t get their work published in high-profile journals, she says. This may be especially true for “negative” studies that find that a cherished hypothesis doesn’t hold up.
Angell, the New England Journal editor, disputes this, noting that when a truly important hypothesis yields negative findings, those results are likely to be published.
The real problem, Angell contends, is that everybody assumes that “we live in a sea of good research,” much of which never sees the light of day. In reality, she says, there’s not a lot of good clinical research out there.
The New England Journal, accepts only 10 percent of the articles submitted. “And we’re not weeping as we reject what we reject,” she says. “We sometimes weep [ about] what we have to accept.”
But Varmus is undeterred. “It will take time,” he says of the new venture. But “in the first year, we’ll show what we can do.”
To read about the government’s new health Web site, go to (http://www.PubMedCentral.nih.gov) or to www.nih.gov/welcome/director/pubmedcentral/pubmedcentral.htm. To read a critique of the proposed site, read a June 10, 1999 editorial in the New England Journal of Medicine at www.nejm.org.
Evaluating health data on the Net
When the government’s new health research Web site, to be called PubMedCentral, goes online in January, it will join thousands of medical sites already available to Web surfers.
Roughly 45 million American households now have online access and half of all US adults use the Net to search for health information, says Dr. Helga E. Rippen, a physician-engineer who tracks such things for Mitretek Systems, a think tank in McLean, Va. Health-related Web sites are constantly being created and abandoned. An estimated 10,000 to 25,000 are up now.
But the quality of information on these sites varies widely, from peer-reviewed articles posted by medical journals such as The New England Journal of Medicine to thinly-disguised advertisements for pharmaceuticals to utterly unsubstantiated claims by quacks or shysters.
A growing numbeer of health information specialists, including the Internet Healthcare Coalition (www.ihc.net), are concerned about the glut of misinformation on the Net. Here are some tips that may help:
- Look to see who wrote the information, checking the “about us” section of the site. If there’s no author listed, or no credentials for the author, be suspicious.
- Check attribution of the information. Research that has been published in a mainstream journal has been reviewed by experts; so has most information from big government agencies such as the National Institutes of Health or the Food and Drug Administration. Information from drug companies may be reliable, but remember, these companies are selling products.
- Try to determine whether the authors have a potential financial or ethical conflict of interest. Ideally, authors are upfront about this, but many aren’t. A more subtle bias can occur with information put out by patients’ groups, which may be so wedded to one point of view that they exclude others.
- Beware of anecdotal information. Emotional testimonials can be convincing – but also misleading or irrelevant to you.
- When you read a study, check the methodology. Was there a control group? How many people were in the study? How long were they followed? Was the study “double-blind”? (This means neither doctors nor patients knew until the study was over who was taking the real drug and who, the placebo.)
- Just as you’d get a second opinion from a doctor, don’t rely on one Web site. Read many sites and cross-check what you find.
- If a treatment seems too good to be true, it probably is.