4 August 2020

Open Research in the Time of COVID-19

Open Research in the Time of COVID-19

The impact of global lockdown

In fewer than 6 months the ways of working and communicating research have shifted at an unprecedented rate. In a matter of weeks there were funder requirements for data sharing to speed up research into COVID-19; collaborative international guidelines developed on sharing COVID-19 data in a range of disciplines; newspaper discussions about the sharing of modelling code; an explosion of preprints and accelerated peer review. Many of these changes have been discussed and advocated for a number of years but the pandemic has accelerated behaviour changes across research.

Changing – and continuing – open practices

As lockdown restrictions in the UK ease and the research sector moves into planning for the next phase of the pandemic, the University Library wanted to reflect on how changes in research communication have had an impact locally, which changes may remain in the long term and the lessons we can learn. There have been lots of examples of Sheffield researchers continuing or adopting new open research practices.

According to Dimensions.ai there are 110 publications by The University of Sheffield authors about COVID-19 across a wide range of disciplines, 10 of which are preprints shared prior to peer review to speed up dissemination. This early distribution of research allows findings to be scrutinised sooner and for them to be built upon faster, which is clearly crucial in the middle of a global pandemic.

Increased openness in research is not only occurring through traditional publications; funders and publishers have mandated that data relating to COVID-19 is shared “as rapidly and widely as possible”. Meanwhile, data visualisation has taken on an important role in explaining the spread of the pandemic through numerous discussions about ‘flattening the curve’.

Data sharing allows others to build on your research, and when accompanied by visualisations it can also be crucial for helping you and others to better understand your data.

Colin Angus in the School of Health and Related Research has been sharing visualisations of publicly available data on COVID-19 throughout the pandemic using Github to manage the codeTwitter as a dissemination tool, ORDA to archive the data and make it easily cited, and MedRxiv to share the subsequent collaborative paper. This has made the existing data more understandable and his work was recently cited, alongside that of Professor Alasdair Rae, in an Institute for Fiscal Studies report.

Professor Helen Kennedy has written persuasively about how public understanding of these visualisations is not well understood and there is potential for them to be used to mislead as well as to inform. It is more important than ever to be able to produce clear visualisations which represent the data responsibly, so the Library and IT Services have relaunched our data visualisation hub which provides tutorials, case studies and a space to discuss issues researchers are facing.

Open research workflows have also been exemplified by the work of the Sheffield COVID-19 group and colleagues who have recently published in Cell on their work looking at the mutation of COVID into a new, more infectious strain. This research analysed genome samples which were published openly on a data sharing platform, GSAID, and was initially posted on the preprint server BioRxiv before its formal publication in the journal Cell. This project demonstrates the power of open research to enable new questions to be answered and for research to happen at a faster pace.

What will become the new normal?

The UK is moving out of lockdown and, whilst it has rapidly become cliched to talk about the ‘new normal’, it is worth considering what changes will remain after the immediate crisis has passed, especially given the straitened finances universities are facing. Some publishers have already announced that the temporarily removed paywalls will be returning in the autumn, and the workloads involved in writing and processing the current volume of preprints on COVID-19 is unsustainable (see Richard Sever’s comment about BioRxiv staff working 7 days/week). So what will remain?

The UKRI proposals in their Open Access (OA) consultation suggest that OA policies in the UK are likely to be strengthened, and the recent government roadmap for Research & Development has come out strongly in favour of data sharing. Internationally, cOAlition S has released its rights retention strategy aimed at safeguarding researchers’ intellectual ownership rights and suppressing unreasonable embargo periods in support of the broader open access Plan S principles. Simultaneously researchers have adjusted to the absence of paywalls and logins for many electronic scholarly resources and this may embolden universities when negotiating with publishers in the coming months.

The discussions about research software have already produced considered commentary from the Software Sustainability Institute about the importance of resourcing and incentivising code sharing. Researchers at The University of Sheffield have access to an excellent team of Research Software Engineers who can help improve research software, a practice being adopted at many other universities.

An expansion of these types of roles may be one consequence of COVID-19, as there is greater recognition of the time, effort and expertise required to ensure the code underpinning research is reproducible and made open in a timely way.  At Sheffield, we are also lucky to have the recently appointed Research Practice Lead, Tom Stafford focussing on good research practice generally and reproducibility in particular.

The crisis has shown there is substantial scope for reforming peer review, but what shape this will take is unclear. Whilst there has been an increase in the posting of preprints, there have also been more checks applied to preprints and some journals have been accelerating their formal peer review processes for COVID-19 research. At the same time traditional peer review failed to identify substantial problems with the New England Journal of Medicine and Lancet papers on Hydroxychloroquine which were eventually retracted, so there may be calls for a more thorough, closed peer review process.

There are also discussions of crediting (financially or through other incentives) peer review, funders using open peer review on their open research platforms, and suggestions to expand the scope of peer review to include data and code. Peer review remains a contested area with lots of suggestions on how to fix it but seemingly little consensus.

It is not clear what scholarly communication will look like after the COVID-19 pandemic, but this period has been a large-scale stress test for a series of changes to open up research which have been discussed for many years. The challenge for researchers and those of us facilitating research is how we can learn from this.

Rosie Higman – Research Data Manager

Fore more like these, check out our Open Research blog pages!

Ask a question

Email: library@sheffield.ac.uk

Phone: +44 114 222 7200