Celebrating investigations: the UK Parliamentary report on research integrity

The UK Parliament Science and Technology Committee, chaired by Norman Lamb MP, reported last month on research integrity. Hindawi’s Head of Research Integrity, Matt Hodgkinson, expands on comments he made to Nature Index on the release of this report.

Politicians are often seen as being out-of-touch, especially when it comes to science, but the report by the Science and Technology Committee is well-informed and plain speaking on research integrity issues and how they affect the UK. British institutions could do more to prevent and address research integrity issues, but this report is a clear and constructive contribution to improving the situation. UK government departments joining the UK universities’ concordat to support research integrity is a great suggestion, and the committee resolving to flag issues in its own 2017 report on regenerative medicine in light of the Macchiarini scandal is another example of this committee “eating its own dog food”.

I'd like to see all research institutions go further and have specific research integrity officers... Prevention is better than cure. Click To Tweet

The report rightly recommends enforcing the research integrity concordat, including having a named senior contact and publishing an annual report. I’d like to see all research institutions go further and have specific research integrity officers (RIOs) and publications officers, as many Russell Group universities do already. They can provide training at all levels (not only to graduate students and doctoral supervisors, as the report recommends) and advice on data management, analysis and statistics because many academics work in comparative isolation.

It is surprising, as the report notes, that many UK universities report having had no investigations for several years, but even if this persists their RIOs will have plenty of work – prevention is better than cure. Having a specific office for investigations avoids confusion about where responsibilities lie and may limit COIs. The soon-to-be compulsory annual reports could be less empty if they included all claims raised to the university’s attention, not merely the number of formal investigations they opened.

Artificially separating questionable research practices (QRPs) and misconduct is a mistake this report avoids. The emphasis on statistical rigor and transparency is important, as it recognizes that QRPs are much more common than outright misconduct. A frequent QRP is publication bias and getting as much research published as is possible is key to preventing much misreporting of science; research institutions and funders should make this a requirement.

Another aspect of transparency is data sharing, which Hindawi has contributed to this year with our requirement for a data sharing statement for all our research articles. I’m worried by the proposal to limit access to research data to avoid possible misuse because beyond necessary limits (eg for patient privacy, environmental protection, or legal issues) this risks a perception of circling the wagons to shield research from scrutiny as seen with the PACE trial.

Another issue with the report was a focus on clinical/biomedical science; problems with basic science are not given as much attention by this report, or in general, but for example work by Malcolm Macleod and colleagues on reporting shows they are real. The report was understandably limited to STM, but similar issues may affect the humanities, eg a lack of data sharing among historians, and more research is needed spanning different scholarly communities.

It would be good to reach agreement across publishers on detecting errors and fraud using software, such as with image manipulation – as was previously done with Crossref Similarity Check for plagiarism – to introduce a common infrastructure for the prevention of publishing articles with research integrity issues. Many publishers are already discussing this and more structure to these discussions would be helpful.

The report suggests creating a new body – hopefully one with teeth – to oversee university research integrity and create standards and consistency in investigations. This would fill an important gap, but it is a pity it will not be part of UKRIO, who welcomed the report but do not want to take on a statutory role. In any case, it is surprising that many UK universities have not signed up to UKRIO and they should do so as quickly as they can! Australia has a similar organization and China is introducing oversight of investigations by their Ministry of Science and Technology: such a research integrity organization is needed for all countries or international regions. The UK could work with other countries, for example via the G7, to create global standards.

Publishers have a duty to correct the literature no matter how errors arose. Click To Tweet

The discussion of training and culture in the report is important. Investigations that explore how misconduct occurred and how it can be avoided in future are more constructive than those which narrowly focus on guilt, as these can fail to see the wood for the trees when exonerating or condemning individual researchers. Errors need raising early to the attention of journals because publishers have a duty to correct the literature no matter how errors arose and, equally, journals must report credible suspicions to institutions else serial offenders will keep operating.

As the report suggests, universities and publishers fixing errors and investigating misconduct should be celebrated and not seen as a badge of shame.

 


Matt Hodgkinson oversees publication ethics at Hindawi. He was previously a Senior Editor at PLOS and BMC.

Opinions in this blog post are that of the author, and not necessarily that of Hindawi. The text in this blog post is by Matt Hodgkinson and is distributed under the Creative Commons Attribution License (CC-BY). Illustration by Hindawi and is also CC-BY.