As COVID-19 became less novel and more a fact of life, society as a whole began to look at the “inadvertent consequences” of the pandemic—from the small, silly things like, “why do we stand so close to one another in the checkout line?” to the bigger, more serious ones, like the rapid, life-changing advancements of mRNA vaccine technology.
Now, in a new study published in PLOS Biology, two UK researchers are examining another inadvertent consequence: the unprecedented rise of preprint publishing—within the context of biosecurity and increasingly open science.
Of the 125,000 COVID-19-related scientific articles published from January to October 2020, more than 30,000 papers—or 25%—were hosted by preprint servers. In the first few months of the pandemic, that number was even higher, with 40% of COVID-19 literature shared as preprints. These shared critical data, including a medRxiv paper that showed most outbreaks of SARS-CoV-2 could be contained if upward of 70% of close contacts were traced and isolated—data that was subsequently cited in national policy guidance from Europe to Africa. Another preprint that showed the infectiousness of SARS-CoV-2 peaked up to three days before symptom onset completely changed quarantine and contact tracing policies for many countries. The paper appeared in Nature Medicine one month later.
While most scientists and peers are not arguing the value of preprints and open science in general, the authors of the new study do caution the risk of deliberate misuse of biological research.
“As open science becomes widespread, we must consider its impact on those risks and develop solutions that ensure security while facilitating scientific progress,” write co-authors James Smith and Jonas Sandbrink, both from the University of Oxford.
The paper identifies one risk mitigation opportunity for preprint publishing specifically, as well as one general recommendation for dual-use oversight. Notably, both hinge on the importance of oversight earlier in the research lifecycle.
Compared with traditional journals, there are only a few prominent preprint servers, namely bioRxiv, medRxiv, arXiv and OSF preprints. The authors think this is an advantage, presenting a unique opportunity for coordination. For example, if an article is flagged by at least one server as potentially concerning, other servers could agree not to post that article until it is traditionally peer reviewed.
For the oversight of dual-use research of concern, regardless of publication route, the study authors see preregistration as an important mitigation component. Preregistration means archiving a time-stamped protocol that can be referred back to once a project is completed. Clinicaltrials.gov, where details of the study design, outcomes, results and follow ups are provided for all clinical trials in the U.S., is an example of a form of preregistration currently utilized in the industry.
In high-risk fields, the authors suggest journals require the inclusion of a hazard assessment as a prerequisite for publication. They also show support for “registered reports,” a specific type of preregistration where the introduction and methods (stage one submission) are peer-reviewed before the work begins. If all is well, the paper is guaranteed publication. If there are concerns, however, a biosecurity expert could be assigned to risk-assess the article before publication. While this is similar to how some traditional journals already operate, Smith and Sandbrink argue that the format of preregistered reports and preregistrations is currently focused on confirmatory research—as opposed to exploratory work, which is usually more applicable to dual-use research of concern.
“There are proposals in the open science community to adapt preregistrations to be suitable for exploratory work. If implemented, early collaboration with biosecurity experts would be advantageous in ensuring that risk of misuse is one of the criteria considered,” the authors write. “Addressing dual-use risks at the early stages of the research lifecycle may be more effective than suppressing the dissemination of dangerous insights after work is completed. Therefore, interventions aimed at encouraging review at the conception of research seem particularly promising.”
As with most things, Smith and Sandbrink acknowledge that money plays a role in any of the proposed changes, specifically when it comes to funders and institutions. These organizations are in the best position to recommend—or encourage, mandate, incentivize, etc.,—that scientists follow a specific plan when it comes to risk mitigation efforts.
“In general, there is a need for ongoing, critical evaluation of incumbent and changing scientific practices, and consideration of the risks that such practices represent,” conclude Smith and Sandbrink.