The word “publishing” has taken on new meaning. For centuries, publishing required an intermediary, someone who had an editorial staff and a press — in short, a publisher. Today, anyone equipped with a computer, a modem, and a connection to the Internet can be a publisher. The consequences of this technical transformation should give pause to all of us in medicine. Direct electronic publishing of scientific studies threatens to undermine time-tested traditions that help to ensure the quality of the medical literature.
The first hint that electronic publishing by medical researchers might become a reality came from the field of physics. Research physicists have been communicating with one another electronically for approximately 15 years, sharing preliminary results (“preprints”) before sending their data for publication. Approximately four years ago, a physicist, Paul Ginsparg, began collecting these preprints of research papers in a host computer at the Los Alamos National Laboratory. That computer has since become a vast repository of physics preprints. Every day, thousands of physicists in 60 countries receive e-mail messages from the Los Alamos computer, and any of them can download the files from the host computer.1 The physicists submit and replace their research “papers” and make changes whenever they see fit. The system functions much like an electronic journal, yet because the individual communications can be constantly revised by the investigators, they are not fixed in time as are papers in print journals.
Some have suggested that this method of publication might be desirable for medical studies. One group of devotees of electronic publishing recently proposed that, as in physics, medical-research communications might be stored electronically and titles and abstracts distributed by the Internet daily.2 Papers would be considered works in progress that could be changed as the author wished. In this system an open process in which anyone could comment on any paper would replace the peer-review system. The quality of a paper would be assessed by the number of times it was cited and retrieved.
What are the flaws in this proposal? In our opinion, a study represented by an abstract or a presentation at a medical meeting is incomplete until it undergoes peer review, is revised accordingly, and is published. A study found to be badly flawed during peer review may be completely revised or never published. Publishing preprints electronically sidesteps peer review and increases the risk that the data and interpretations of a study will be biased or even wrong. Investigators cannot be expected to judge their own work dispassionately. They are usually enthusiastic about their hypotheses and may be unaware of flaws in the design of their experiments or of the insufficiency of their data to support their conclusions. They need independent experts to evaluate their data. Most journal editors rely on carefully selected reviewers whose opinions about the originality, validity, and timeliness of a manuscript count heavily in the decision for or against publication. The reviewers are an investigator's peers; they cannot be replaced by multiple unspecified users of the Internet. When a scientific study is assessed by majority rule, the result is likely to be highly unreliable. Such a process could also invite manipulation and even fraud.
Another source of concern is the fact that much information about health issues on the Internet, such as the risks of medications and the effects of various foods on health, is of uncertain parentage. (Surely, anonymity has no place in reporting medical research.) At present the Internet seems to promote medical rumors more than dispassionate scholarship. Papers in medical journals are technical communications intended for physicians, not for direct consumption by the public. Physicians should act on new studies only after carefully considering the strength of the research and how the new data relate to previous studies, and after assessing whether the population studied is relevant to the patient at hand. Public access to Internet preprints of medical studies might lead some people to use the wrong medications or to stop taking needed ones on the basis of inadequate information, as has happened from time to time when news reports failed to interpret adequately a presentation at a medical meeting.
According to our current policies,3 we do not consider a manuscript for publication if its substance has already been reported elsewhere. An explicit policy is needed for the Internet because of its enormous capacity to transmit information to a large number of people simultaneously. We have decided that electronic publication should not be regarded differently. Thus, posting a manuscript, including its figures and tables, on a host computer to which anyone on the Internet can gain access will constitute prior publication. On the other hand, sending manuscripts by e-mail to a limited number of colleagues — a dozen or two, let us say — will not. Such a practice is analogous to faxing manuscripts to a group of research collaborators.
For studies submitted to us that have immediate health implications, we will continue our efforts to see that dissemination of urgent medical information is not delayed by either peer review or the publication process. Accordingly, we will discuss with authors the need for expedited review and rapid disclosure of studies, as we have described elsewhere.3 We believe that these policies are the best way to minimize dissemination of uninterpretable information while permitting the free exchange of ideas and data among colleagues and collaborators.
Physicians have not been at the forefront of electronic communication, but many are beginning to use e-mail and to explore the Internet. Electronic communication is likely to become critically important in medicine,4 and it behooves physicians to become competent in using it. In the future, when the Internet is widely used by physicians, information about advances that may have an immediate effect on the health of individuals or populations may best be communicated through this medium. But medicine is not physics: the wide circulation of unedited preprints in physics is unlikely to have an immediate effect on the public's well-being even if the material is biased or false. In medicine, such a practice could have unintended consequences that we all would regret.
