Intended for healthcare professionals


Making research papers in the BMJ more accessible

BMJ 2002; 325 doi: (Published 31 August 2002) Cite this as: BMJ 2002;325:456

We're developing ELPS and will soon publish papers shortly after acceptance

  1. Marcus Müllner, associate editor (marcus.muellner{at},
  2. Trish Groves, assistant editor (tgroves{at}
  1. BMJ
  2. BMJ

    In 1999 the BMJ started using a process called ELPS (electronic long, paper short) to publish shorter, more reader friendly versions of original research papers in the print journal.1 Because readability is just as important on the web, these shorter papers are also published on along with their original longer versions. To make this clear to readers of we mark the papers “abridged text” and “full text.” Like everything else on, both versions of each research paper are freely accessible to everyone. The long versions are indexed for Medline, and are published in full on PubMed Central. Readers and authors seem pleased with ELPS so far, and we are continuing to explore new ways to serve their needs.

    In our current ELPS process the long version of the paper, after acceptance, is edited according to BMJ style. We aim to eliminate language ambiguities and correct numerical errors not detected during peer review. After editing the paper is sent to the authors for approval. Then the long paper is shortened by a BMJ editor, and the short version is sent to the authors for final approval.

    We have had very few complaints about the way we have shortened papers, and we are conducting a survey among authors and potential authors to see if there's more we can do to make ELPS a useful service. The process of creating an ELPS article might slow down publication slightly because it introduces another step into the process—the few days it takes for us to shorten the article and send a second set of proofs to the author. But in the near future we hope to be able to compensate for this by publishing accepted papers on as soon as we can, and then adding the edited long and short versions when they are ready. This will provide a better service to authors, readers, research participants, funders, and patients, all of whom want important research findings to be available as soon as possible.

    Some authors have started to send us two versions of their papers at the outset. Please don't do this: it doubles the work at all stages of peer review. When first submitting an original research paper to the BMJ, please send us just one full version that fits with our advice to contributors (as described at Once your paper is accepted we will take care of the rest, though there are a few things you can do to streamline the process (box). And please don't worry, we will not impose changes that authors are not happy with.

    How long should the full version be? Ideally we should not impose word limits on these, particularly as space is not a problem on We consider the long versions to be the original papers and, accordingly, put considerable effort into coming up with the best product possible. But we believe that most medical research can be reported clearly within a limit of around 2000 words; except, say, when a randomised controlled trial is reported together with an economic evaluation, when up to 3000 words might be acceptable. Furthermore, our resources are limited: we know that editors, reviewers, and technical editors would be faced with an impossible workload if all BMJ papers were excessively long. Lastly, we know that few readers want to plough through pages and pages of material (even on So, while we hope authors will try hard to stick to 2000 words of main text, we will try to be flexible.

    If we accept your paper provisionally after peer review and ask you to revise it, please send back just one version. Of course, the offer of publication often comes with a request to cut the text considerably while also answering a long list of comments from reviewers, an editorial committee, and technical editors, and we know that this is frustrating. But thoughtful and creative use of tables, boxes, and figures, and sharp pruning of the introduction and discussion sections, usually do the trick. A structured discussion can be effective and succinct.2

    We are still exploring the best way to present shortened articles to our readers. Other journals are following with comparable strategies. Meanwhile, we are about to start a randomised controlled trial to test different formats and styles of shortened papers. In a few years' time the research papers in the print BMJ may look more like articles in a serious newspaper, while those on could be live documents with raw data and multimedia features. For many people, news websites and articles in the Economist or Newsweek are much more entertaining to read than original papers in the BMJ, and entertainment is one of our missions. More importantly, however, we want to inform, educate, and influence, and we want to make every word count. We think ELPS helps us to do that.

    To get your original research paper* published in the BMJ as quickly as possible, please

    • Write your paper in line with our advice to contributors at

    • Send us just one version of around 2000 words

    • Revise and return your paper within two months of being invited to do so

    • Send us only one revised version of around 2000 words

    • Email it with a detailed letter explaining how you have revised the paper

    • Check and return quickly the edited long and short versions

    • Avoid making more than a few marks on the proof.

    *Assuming, of course, that it describes well designed and executed research that matters.


    1. 1.
    2. 2.
    View Abstract