Shifting trajectories

Michele Avissar-Whiting outlines the course of post-pandemic preprinting 

At a press briefing on 21 January, Dr. Anthony Fauci briefly let the public in on academia’s open secret: preprints have enabled the breakneck speed of discovery we’ve witnessed during the Covid-19 pandemic.

The urgency of the crisis inspired researchers to skip the rigamarole of the standard publishing process and release their findings as soon they were in hand – so that others could read and build on them. Without question, prepublication data and full manuscripts made available in preprint servers and other open data repositories shaved months off the development of Covid-19 vaccines and guided clinical management of the disease and public health recommendations to control its spread.

Though preprinting has been a norm for researchers in the physics and math communities for decades, it is relatively new to biomedical circles, with only modest increases in uptake seen each year since 2015 – that is, until 2020. While most researchers in these fields still haven’t adopted the practice, the pandemic has more than doubled the monthly rate of biomedical preprints, according to the preprint advocacy group, ASAPbio. And this trend has shown no signs of slowing.

What we’re seeing now in the wake of this trial by fire is a sort of renaissance in academic publishing: a collective realisation that there may be a better way to share research and assess its legitimacy. Preprints would seem to be a panacea for the many problems that have plagued scientific publishing. They condense the time to dissemination from months to days, make access to science immediate and free, and allow for wider and more diverse scrutiny of the work before it gets a journal’s stamp of approval. But these benefits have some problematic counterpoints that have garnered significant attention during the pandemic. 

Free dissemination and open access means that most anything resembling a research article can find its way onto a preprint platform, and anyone with an internet connection can read it. This permissiveness has made the medium more subject to exploitation by bad actors: those seeking merely to advance narratives and peddle dubious treatments and prophylaxes. In response to this, the more well-established and better governed preprint servers have tightened up their existing screening processes to filter out alarming, potentially harmful, and fringe topics or those that don’t seem like sincere intellectual pursuits. Moreover, the leading preprint servers are actively engaged with initiatives driven by organisations like ASAPbio, aiming to build trust in preprints while guarding against their misuse.

Those studies that do make it past the screen (around 90 per cent on our platform, Research Square) are adorned with a red-lettered disclaimer alerting readers to their not-yet-peer-reviewed status. This disclaimer typically propagates out when a preprint is shared, but it doesn’t stop the occasional reader from sharing a preprint as though it has been vetted. This has been a reasonable source of worry for anyone concerned about the spread of misinformation. Ultimately, though, most of the research posted as preprints winds up published in journals, without dramatic changes to the conclusions, according to a recent study published in eLife. So if critical issues are not addressed during peer review – and too often, they aren’t – those studies continue to be shared and cited, but without caveat, under the premise that the research is sound. 

The issue of readers misinterpreting or intentionally misrepresenting the conclusions of preprints has also been pervasive during the pandemic. But that confusion and misuse is not unique to preprints - it continues with the journal-published version. It is a genuinely difficult problem that perhaps can only be addressed by authors (and reviewers and editors) anticipating misinterpretation and misuse and taking a more proactive approach to stave it off.

The system that has governed scientific publishing for decades is based on the notion that research needs validation by at least two unbiased subject matter experts in order to be shared. But as research outputs have ballooned, publication has become excruciatingly slow. The global emergency of the last year allowed us to witness a new system in action: one in which a study posted online could receive immediate scrutiny from tens or hundreds of peers. This experiment in post-publication community review, which took place on the most consequential preprints about Covid-19, showed us that a preprint could be discredited and withdrawn nearly as quickly as it had been posted. That said, most preprints do not draw this level of attention; many are not scrutinised thoroughly until they’re submitted to a journal. Indeed, until preprints are given some reliable and recognisable mechanism of endorsement, they won’t be afforded the same gravitas as journal articles.

My hope is that the preprint movement is the first in a series of shifts culminating in a monumental transformation in scientific publishing. The new model, if it’s to scale with the current volumes of research being produced globally, will need to involve some combination of standardised assessment (both automated and human-driven), post publication review, and a system of endorsement and curation likely assumed by journals. It won’t be an easy move for an industry so notoriously bound to tradition, but the pandemic has forced us to reconsider the status quo for so many aspects of human life. This may be yet another silver lining.

Michele Avissar-Whiting is editor-in-chief at Research Square