Editorials

Netprints: the next phase in the evolution of biomedical publishing

BMJ 1999; 319 doi: http://dx.doi.org/10.1136/bmj.319.7224.1515 (Published 11 December 1999) Cite this as: BMJ 1999;319:1515

Will allow researchers to share their findings in full, for free, and fast

  1. Tony Delamothe, web editor,
  2. Richard Smith, editor,
  3. Michael A Keller, university librarian and publisher,
  4. John Sack, associate publisher,
  5. Bill Witscher, associate director
  1. BMJ, and editor clinmed.netprints.org
  2. BMJ
  3. HighWire Press, Stanford University, Stanford, CA 94305, USA

    Graphic Where the mind is without fear and the head is held high;

    Where knowledge is free;

    Where the world has not been broken up into fragments by narrow domestic walls;

    Where words come out from the depth of truth;

    Where tireless striving stretches its arms towards perfection;

    Where the clear stream of reason has not lost its way into the dreary desert sand of dead habit;

    Where the mind is led forward by thee into ever-widening thought and action—

    Into that heaven of freedom, my Father, let my country awake.

    from Rabindranath Tagore's Gitanjali

    This week we launch clinmed.netprints.org, an electronic archive where authors can post their research into clinical medicine and health before, during, or after peer review by other agencies. Resulting from a collaboration between the BMJ Publishing Group and Stanford University Libraries, it will allow researchers to share their findings in full, for free, and as soon as their studies are complete.

    Articles will be screened for breaches of confidentiality and libel before we post them. After posting authors may submit them to any peer reviewed journal that will accept submissions that have appeared as electronic preprints. The list of such journals extends far beyond those of the BMJ Publishing Group and is growing daily (see box) Researchers who have retained the right to post their research results after publication in a peer reviewed journal can archive their articles here rather than on possibly more ephemeral institutional or personal websites.

    Journals accepting submissions that have appeared on preprint servers

    • American Journal of Botany

    • Biophysical Journal

    • BMJ

    • British Journal of Ophthalmology

    • European Journal of Public Health

    • Genetics

    • Journal of Accident and Emergency Medicine

    • Journal of Biological Chemistry

    • Journal of Clinical Pathology

    • Journal of Cognitive Neuroscience

    • Journal of Epidemiology and Community Health

    • Journal of Medical Genetics

    • Journal of Medical Screening

    • Lancet

    • Molecular Biology of the Cell

    • Molecular Pathology

    • Nature Medicine

    • Nature

    • Neural Computation

    • Occupational and Environmental Medicine

    • Pre-Hospital Immediate Care

    • Proceedings of the National Academy of Sciences

    • Quality in Health Care

    • Thorax

    • Tobacco Control

    Authors should check with editors for the current status of their journals

    With at least 20 000 websites devoted to health and medicine1 why start another one? Arguments for and against electronic preprints have been rehearsed several times in the pages of this journal and on our website.25 In brief, we see electronic publication of a completed study before peer review as analogous to the presentation of results at a meeting, an important part of the scientific process. Not everybody can travel to scientific meetings, but most researchers have access to the world wide web. Often the eventual publication has benefited as much from feedback received at meetings as it has from conventional peer review. We will therefore be providing facilities for direct reader feedback (familiar to anyone who has seen the rapid response feature on the BMJ's website). In many ways the BMJ's new netprint server (covering both electronic preprints and electronic reprints) is more evolutionary than revolutionary because we have always regarded publication in the paper journal as not the end but rather only part of the peer review process. Every editor has seen published studies destroyed in the correspondence columns.

    The arrival of the world wide web has given us the chance to rethink the trade offs that have traditionally been made in scientific publication. Previously, we have accepted that the benefits of quality control have outweighed its costs. These include the long delays between the completion of research and its publication, the random scattering of articles on similar topics among journals of varying accessibility, and the loss of some articles (not necessarily the worst) from the system completely. Meanwhile, the price charged for this value adding service is rising faster than most customers can afford.



    Embedded Image

    After three international congresses devoted to peer review its claims as an efficient means of quality control are looking threadbare,6 while the disadvantages of the current system are becoming more apparent. For example, those attempting systematic reviews of all research on particular topics often find that many studies are unreported or unlocatable. Yet these results may be crucial, especially given the publication bias against negative findings Patient care depends on clinicians having easy access to unbiased evidence, and researchers and funding agencies want to avoid duplicating research. In these ways the BMJ's netprint server shares the same aims as the calls from the Lancet and BMJ for all clinical trials to be registered. 7 8

    Opponents of free availability place a high value on peer review, regardless of the contrary evidence. They argue that it efficiently winnows out the wheat from the chaff and that articles not yet through the mill of peer review may harm public health or even science itself. Yet many unreviewed findings already find their way into the public arena with the connivance of researchers and conference organisers, usually in an uninterpretable form. Far better, we believe, for a study to appear in full, with the opportunity for readers to assess its claims and append their criticisms. We are alert to the danger that members of the public may be misled by what they read on the website, but an explicit warning appears on its opening screen and is repeated above every article (see above right).

    A substantial proportion of researchers—our intended audience—seems in favour of the concept. In a survey by Bioinformatics Inc two years ago 46% of 1067 scientists and medical professionals rated the practice of publishing electronic preprints as very or somewhat valuable.9 Four months ago we took a straw poll of authors of research published in the New England Journal of Medicine, the Lancet, and the BMJ to see whether our target user group shared similar views. Although the response rate of 30% was too low for firm conclusions, the results suggested that they did: three times as many respondents were in favour of preprints as were against, providing prior publication of them did not prejudice acceptance by peer reviewed journals. With this proviso 24 of 43 respondents would consider posting their articles on a preprint server.10

    Our proposal is not a covert attempt to do away with peer reviewed articles or journals but offers instead another route to “publication” with a different set of advantages and disadvantages. We can't know what clinical researchers—as authors and readers—will choose until they have been given the choice. In high energy physics a thriving electronic preprint server coexists with traditional peer reviewed journals, where most of the preprints are eventually published. The BMJ's hope is that eventually the whole peer review process will be transformed from the current black box to an entirely open process.11 The whole world will be able to watch, perhaps helping to restore public confidence in science. The modern world is suspicious of closed processes.

    Other examples of preprint servers exist: the Lancet has one for research into international health and for articles accepted by the journal for peer review.12 More are planned in clinical medicine13 and the basic life sciences. We hope to share common standards of interoperability with other servers, and we support the open archives initiative.14 The initiative's ultimate aim is to allow readers to locate articles on different servers, as if the articles were all in one virtual public library.

    As the possibilities of the world wide web are being exploited, the old style scientific journal is coming apart at the seams. Each function it performed is being scrutinised, and how to deliver it is being rethought. Netprints is just the latest example of this new tendency We hope that researchers will use clinmed. netprints.org and let us know how it could be improved. It could take us one step closer to realising Tagore's vision of untrammelled access to knowledge.

    Footnotes

    • We ask all editorial writers to sign a declaration of competing interests (www.bmj.com/guides/confli.shtml#aut). We print the interests only when there are some. When none are shown, the authors have ticked the “None declared” box

    References

    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.