In academia, preprints are versions of articles that are freely available online before they are peer-reviewed and officially published in a journal. Articles uploaded by researchers to these sites are only preliminary reports with unverified results. Popular sites for biomedical research include bioRxiv and medRxiv. These sites contain warnings that the articles “should not be relied on to guide clinical practice … and should not be reported in news media as established information”. Since the onset of the pandemic, the number of articles being uploaded to preprint sites has surged; as of the 7th May, medRxiv has posted 2355 COVID-19 related papers. Preprint is a fast way to share novel findings, as the process of having work peer-reviewed and accepted to a scholarly journal can take months or even years. However, such a platform where unverified information is made public also poses many dangers.
Preprint allows scientists to receive early visibility, instead of having to wait years for journal publication, in what can be a bureaucratic process. It allows the opportunity for peer review by hundreds of experts online, instead of one or two. Peer-review is a vital part of the scientific process, and prevents biased and poor-quality work from being disseminated. The logic behind preprint is that bogus reports are quickly shot down, preventing weak and flawed papers from ever being published in a journal. It has been argued that use of preprints puts the focus back on the research itself, and less on the journal in which it is published. During the coronavirus pandemic, preprint has been invaluable to scientists from different research groups and countries, allowing them to collaborate and quickly share discoveries about treatments and vaccines as they are being made.
But due to the obvious extensive media coverage of the pandemic, this rapid publication also brings the risk that false or unreliable findings could be picked up by news outlets and sensationalised, contributing to panic and hysteria.
Whilst preprint certainly has its benefits within the scientific community, it has always faced understandable criticism. Before preprint, once a report was ready to be published in a journal, journalists would attend press releases and interview experts. This allowed verified results to be publicised in an informed manner. In the age of the Internet, journalists can go onto preprint servers and unintentionally spread flawed research to the public. History has shown that even once false scientific findings have been numerously refuted, much of the general public will not be swayed – and so the damage has been done. An age-old example of this is the MMR vaccine scare: in the early 1990s, findings were prematurely published of a link between MMR vaccinations and autism in children – which has since been extensively disproven. Despite this, the MMR-autism link is still cited by many anti-vax groups, presenting a tangible risk to public health.
Physicists and mathematicians have used preprint for years without any trouble – likely because news about electromagnets and vectors isn’t going to make big headlines, but research about the side effects of new drugs does. Critics have suggested that preprint is simply just not suitable for all fields of science. A topical example is the reported effect of ibuprofen on coronavirus symptoms. Towards the end of March, reports from a range of outlets began to emerge that patients self-medicating with ibuprofen suffered more severe coronavirus symptoms. People Magazine wrote “Experts agree that more research is needed, but that it is best to opt for paracetamols like Tylenol for now”. These reports led to panic buyers emptying supermarket shelves of paracetamol-based painkillers, whilst stocks of ibuprofen remained high. The Commission on Human Medicines has since stated that there is insufficient evidence supporting these claims. Ironically, UK doctors have since begun to trial different formulations of ibuprofen as a treatment for coronavirus. Nonetheless, for many members of the general public, the ship has sailed, and they will choose not to ‘risk’ taking ibuprofen when self-medicating symptoms. Whilst it could be argued that if the research had eventually been proven valid then the responsible thing to do was to warn the public immediately, such risks must be better mitigated to prevent so-called infodemics. There have always been risks to sharing dubious science, but these risks are much more acute when they affect public health and medical practice during a global pandemic.
Fortunately, preprint servers have begun to take more active measures against poor-quality coronavirus work. BioRxiv have now stopped taking papers that make predictions about COVID-19 solely using computational data. Albert-László Barabási, a computational scientist from Northeastern University, disagrees with these changes. Speaking with Nature in a recent article, he stated “The purpose of a preprint server is that we decide what is interesting, not the referees”. Whilst this is a compelling point regarding why preprint servers came about in the first place, this is where the potential for abuse of preprint arises. In the same article, one of the founders of bioRxiv stated, “We’re not going to peer review to work out whether the [computational] modelling they’re using has any basis. There are some things that should go through peer review, rather than being immediately disseminated as preprints.” Whilst it is reassuring that preprint servers are taking steps to reduce the spread of misinformation, are these steps enough? These conflicting attitudes reflect the on-going debate on the suitability of preprint articles for all fields in the scientific community. An open letter was published by Fiona Fox from the Science Media Centre, which concludes that “the changes being made to a part of the system that was not working are set to have profound knock on effects on another part of the system that works and serves science well. The challenge here is to fix one end without losing the gains we have made in reporting findings to the public in an accurate and measured way. Science is rightly proud of its reputation for informing the public based on strong evidence. Let’s not destroy that in pursuit of preprint – however great the prize.”
Whilst it is easy to scapegoat journalists with no specialist knowledge for the problems surrounding preprint articles and misinformation, the debate of preprint articles needs to involve both experts and non-expert journalists. Since so much scientific writing is not accessible to the general public due to complex terminology, the media are an important part in presenting discoveries to the public, as well as ensuring scientists remain transparent and accountable.
Since the onset of the coronavirus pandemic we have faced an infodemic of fake news, some of which could prove to be as damaging as the virus itself – whether that manifests in anti-vax groups, destruction of 5G towers or anti-Asian racism and xenophobia. It’s important to bear in mind that preprints are not the sole cause of infodemics: various tabloids are infamous for cherry-picking scraps of information and spinning them into provocative headlines, and this issue clearly predates preprint servers. Whilst ways of mitigating the dangers of preprint in the biomedical field must be improved, we shouldn’t let these risks cloud the benefits – particularly at a time when rapid global scientific collaboration is so imperative.