Today’s publishing ecosystem offers a wealth of options with new and existing models in which to disseminate research. Is one model better than another, and what differences do they offer, if any? This debate piece presents a view of two of these models from:
- Robert Kiley (Head of Open Research, Wellcome Open Research) on the value of a new model and reasons behind it.
- Adrian Aldcroft (Editor-in-chief BMJ Open) on existing models and why traditional publishers are still important and offer value.
Robert Kiley is Head of Open Research at the Wellcome Trust where he is responsible for developing and implementing an open research strategy. Robert is a qualified librarian and an Associate Member of CILIP. He also sits as a Director on the ORCID Board.
For more than a decade, the Wellcome Trust has been at the forefront of the open access (OA) and data sharing debates. At Wellcome, we believe that to maximise our research investment, research outputs must be findable, accessible, interoperable and reusable (the FAIR principles).
To support this aim, we encourage innovation in scholarly communication, to improve the way research is communicated. We believe the publishing process can be faster and more transparent, and make it easier for researchers to provide information that supports reproducibility.
With these objectives in mind we decided to launch a new publishing platform, Wellcome Open Research. It uses a publishing model developed by F1000Research which allows our funded researchers to quickly publish all outputs from their research – everything from standard research articles and data sets to case reports, protocols, and null and negative results.
The platform doesn’t undertake any editorial function: every submission is published once it passes a series of objective checks.
The platform’s transparent peer review process (post publication) encourages constructive feedback from experts, whose focus is on helping authors improve their work rather than on making an editorial decision to accept or reject an article.
One of the key areas we looked at was helping to improve reproducibility and reliability, and the platform requires all supporting data be made available (or at the very least, explains how the data can be accessed).
The platform has further helped speed up publishing time. The median elapsed time from submission to publication is 22 days, whilst the median from submission to when an article has passed peer review (and is indexed in PubMed) is 69 days.
We chose to create a new platform, because we saw many distinct advantages. It aligns to our values on openness and transparency and further supports our long-held view that researcher assessment should be based on actual outputs.
Adrian Aldcroft is the Editor in Chief of BMJ Open, an open access medical journal that promotes transparency in research and publishing. Prior to becoming an editor, Adrian conducted research on the human visual cortex at The University of Western Ontario.
Writing in The British Medical Journal, Sir Austin Bradford Hill, famous for identifying the link between smoking and cancer, asked, “Why does the scientific worker write at all?”.1 His answer was that the scientist had an objective to communicate knowledge that didn’t previously exist. To best achieve this objective, Hill outlined the Introduction, Methods, Results and Discussion (IMRaD) structure of the research article that remains familiar today. As an author, Hill’s impact was enormous, but for his voice to be heard he needed a publisher.
Just as The BMJ supported Hill, publishers today have a responsibility to support authors in communicating their work. While no longer required for the dissemination of work, which can now be achieved at the touch of a mouse, publishers add value by providing and overseeing expert judgment of an article and by placing research into a context that can be trusted. Although the traditional process of peer review has been described as imperfect, potentially biased and time-consuming, it remains the best way we have to stress-test research before releasing it to a wide audience.
When communicating research, authors are full of fresh ideas and new perspectives mixed up with biases, blind spots, and conflicts of interest. For a work to become part of the established knowledge base, a lot needs to happen behind the scenes. Considering the importance of the task, along with the complications inherent in being an author, it could be risky to make a raw manuscript public, particularly if the work has the potential to become influential. Reviewers, editors, and publishers provide the critical service of improving a work, often playing a judicial role, balancing viewpoints, making compromises, and bringing a work in line with standards and values deemed important by the research community and enforced by the journal and its publisher.
Which isn’t to say that all research writing needs to pass a difficult peer review process. Different jobs require different tools. Preprint servers, like arXiv in the physical sciences, and the newly announced medrXiv, of which BMJ is a founding organisation, are an important, developing space for posting early versions of a work. Preprints offer a valuable opportunity to share a manuscript and receive feedback from a global community of researchers, but they are lacking the trust that is established through peer review and are not intended for general consumption by the media and the larger public.
Preprints have found success in the theoretical and basic sciences, but for more applied fields of study there are risks in drawing the general public’s attention to work that has not been scrutinized by editors and peer reviewers. This is particularly true for clinical medicine. Evidence-based medicine relies on published research to inform clinical guidelines and ultimately to determine spending on healthcare. It is crucial that we make every effort to ensure that when clinical results are presented to the public they are in context and accurate, and if we make a mistake we need to have the expertise to investigate and correct it. Professional publishers and editors closely linked to the journal provide such expertise.
At BMJ, our publishing choices are closely aligned to evidence-based medical models. The move on some of our journals to open peer review was not made to follow trends; rather, it was based on the results of randomized controlled trials comparing open peer review to more traditional systems to ensure we would not be sacrificing quality or accuracy.2 More recently, we have elected to include patients in the peer review process, so that those most affected by research can help shape it, and again this has been guided by a developing evidence base.3 It would be reckless to dispense with established peer review processes without amassing evidence demonstrating that changes would be an improvement.
It’s probably true that, when done well, the work of professional editors and publishers is so seamless that it goes unnoticed. It’s also true that today much can be accomplished with a laptop and an internet connection. But the value of publishers is largely foundational, working to improve and establish trust in the research we publish.
- Br Med J 1965;2:870. doi: https://doi.org/10.1136/bmj.2.5466.870
- van Rooyen S, Godlee F, Evans S, Smith R, Black N. Effect of blinding and unmasking on the quality of peer review: a randomized trial. JAMA 1998: 280; 234-7. doi:10.1001/jama.280.3.234
- Schroter S, Price A, Flemyng E, et al Perspectives on involvement in the peer-review process: surveys of patient and public reviewers at two journals BMJ Open 2018;8:e023357. doi: 10.1136/bmjopen-2018-023357