Open Research & Scholarly Publications

Covid-19 highlighting issues in academic publishing

The recent retraction of papers published in the Lancet and New England Journal of Medicine, two highly prestigious medical journals, has elicited widespread publicity.

The retracted papers concerned research into the anti-malaria drug hydroxychloroquine and its potential value in treating Covid-19. They were withdrawn when it came to light that the large patient dataset on which their findings were apparently based could not be verified. This followed scepticism about the scale of the dataset when contrasted with known patient admissions and the legitimacy of the company, Surgisphere, which provided it. Professor Sharon Lewin, the director of an Australian Institute researching infectious disease, described the controversy as a ‘wake-up call’, and questioned whether standards had been compromised in the clamour to publish topical research. The credibility which a publication in a prestigious journal such as the Lancet is afforded can be seen by the fact the World Health Organisation suspended medical trials following publication of these now withdrawn publications.

The controversy does highlight two broader issues beyond the immediate issue of the Covid-19 pandemic putting strain on the usual quality control mechanisms which academic journals rely on.

Firstly, that the peer review process relies on trust and taking information at face value, and secondly, that there is a lack of transparency around the sharing of the research data which underpins research findings.

In relation to peer review, in the eyes of many this remains the ‘gold standard’ by which academic research is sifted and evaluated. Peer review is where experts from within the relevant academic field review a paper and make recommendations to the journal as to whether it is worthy of being accepted for publication, and seek to make improvements to the clarity and rigour of the paper prior to its publication. It has come to be understood in recent years that peer review has its limitations, a matter addressed in The Conversation back in 2018. Concerns have for instance been raised around peer review’s reliance upon free labour provided by time poor academics, and that there is a lack of understanding as to how it operates in practice. A point made in a recent open access article: ‘The limitations to our understanding of peer review’ (Tennant, 2020).

Concerning the second point, the failure of authors to share the datasets which underpin their research has contributed to a ‘reproducibility crisis’, with past investigations unable to be replicated and conclusions verified. There remain some hurdles, technological, legal and cultural, in the sharing of different types of research data. In the past several years many research funders have instituted policies concerning the retention and accessibility of research data. UK Research and Innovation (UKRI) for instance, which oversees the UK’s Research Councils, has developed a Common Principles on Data Policy which provides an overarching framework for those in receipt of their research funding.

However, progress has been slower in this area than in relation to open access to research articles. An editorial published in Nature as far back as 2009 lamented the fact that research data was subject to ‘shameful neglect’ and that it was vital for research that data be ‘preserved and made accessible’. This state of affairs continues to some degree up to the present time, with the Open Access policy in effect for the upcoming Research Excellence Framework (REF) focusing its attention on open access to research articles and conference publications. Similarly Plan S, a major new open access framework being adopted by several major research funders, does not include access to research data within its scope. As shown by recent events though transparency around the data being used to underpin research outputs remains a matter of great importance.

For more information on Research Data Management (RDM) principles and for funder specific requirements around RDM, please see the section of our libguide dedicated to Research Data Management.