Online Firsts

Will help to reduce delays in publication of research findings

The BMJ is about to start posting its original research articles on bmj.com before they appear in the print BMJ. They will appear in a new section on bmj.com called “Online First.” By early next year we expect all the research articles we publish to appear online first.

Like many journals, we are doing this because the research community wants results of studies to be publicly available as soon as practicable and because electronic publication makes it possible to remove some of the delays associated with print publication. At present papers are edited and often there are inevitable delays before they are published. At the very least they wait, unread, during the nine days between the time that we start to put a week's issue together and the time the print version lands on UK readers' doorsteps. However, we also think that online first publication prefigures a world of continuous online publication. The BBC updates its website every minute, and some basic science journals are already updating theirs several times a day. At present bmj.com's rapid responses appear every day, but with Online First we are embarking on a journey that in a year or two will probably see much more of the BMJ's website being updated daily.

But for now we are not being that radical. We need to understand the implications and to find out what our readers and authors think. Our plan is to post research articles as soon as they are edited. We considered posting unedited manuscripts as soon as we had accepted them—which is what our sister specialist journals are planning to do when they start their Online Firsts next year—but decided not to for two reasons. Firstly, we conducted two small surveys among our authors, and, although almost half thought posting papers without editing was acceptable, a quarter preferred not to, and a further quarter thought this unacceptable. So for now at least we will continue to edit papers to our normal standards, with authors approving the edited version before we post them. Secondly, we foresaw that if we posted unedited manuscripts we might need to post corrected versions as we go through the editing process. Although we can do that, and will no doubt occasionally have to do so, we're not sure that readers are ready yet for constantly changing versions of an article.

For our ELPS (electronic long, paper short) papers, we will edit the long version and post that once the author has approved it. This will thus be the definitive version of that article. Readers of bmj.com can browse the Online First section, and Online First articles are also fully integrated into bmj.com—thus they will appear when readers conduct an online search of bmj.com or look in one of the website's 250 topic based collections. In addition, the posting of an Online First will trigger an email to readers who have asked to be alerted when articles on that subject are published. When we subsequently print the shortened version of that article in a weekly issue of the print BMJ, the long version will move out of the Online First section and appear together with the short version in that weekly issue—just as it does now. One advantage for print readers of earlier on line posting is that by the time we publish the print version we may also be able to print some of the initial reactions to the article alongside the print version (as our American version, BMJ USA, already does1).

The date of posting will appear on each Online First article. Electronic articles are also identified by a unique number—the document object identifier (doi)—and guidance on how to cite the article will appear in each article.

We know that authors want us to publish their articles as soon as possible. We're less sure of our readers' reactions—though faster publication of trials should help those who are doing systematic reviews and faster publication of systematic reviews should help clinicians by providing decent answers to clinical questions. If posting of research articles online first is a success we will aim to move to do the same to other sections of the BMJ.