WORLD VIEW

Make reports of research misconduct public

Confronted with bad behaviour, institutions will keep asking the wrong questions until they have to show their working, says C. K. Gunsalus.
C. K. Gunsalus is director of the National Center for Professional and Research Ethics at the University of Illinois Urbana–Champaign.
Contact

Search for this author in:

During decades as a research-integrity officer, expert witness for misconduct investigations and consultant, I have been inspired — and I have seen inexcusable conduct. Even when investigations are exemplary and findings clear, universities rarely report them publicly. That secrecy perpetuates misbehaviour and breeds mistrust — as evidenced by the ongoing revelations of universities that failed to respond appropriately, sometimes for years, to allegations of sexual misconduct.

Science is fast becoming more transparent. So, too, should institutional practice. Open misconduct reports would create a virtuous circle. Institutions would learn from their own and others’ investigations. Leaders would be more likely to pay attention to reports that are subject to scrutiny. Honest researchers could see that although groundbreaking science is often uncertain, it is qualitatively different from the conduct that leads to misconduct reviews.

We are already seeing such a shift in health care. Last month, a study showed that mortality is lower in UK hospitals in which medical professionals feel that they can talk openly about problems without worrying about repercussions to their careers (V. Toffolutti and D. Stuckler Health Affair. http://doi.org/c6df; 2019). I often find that institutional investigators ask the wrong questions, such as: ‘We don’t have to report this, do we?’, ‘How could anyone think Dr X would do such a thing?’ or (to whistle-blowers) ‘Why would you want to cause trouble for your own research project?’ Investigators pin all the blame on one actor without examining the contributions of co-authors or supervisors of the flawed work. An investigation might stop abruptly if the subject of it resigns. A 2019 paper examining investigations by institutions after the retraction of 12 clinical-trial papers by one research group stated that although investigations lasted for between 8 and 17 months, they did not examine preclinical papers from this group even after receiving detailed, serious concerns about them (A. Grey et al. Res. Integr. Peer Rev. 4, 3; 2019).

Unfortunately, this situation seems not to have improved over time. It is three decades since a US working group of lawyers and scientists issued a 1989 report finding that institutions seem to have mishandled allegations in at least 8 of 21 research-misconduct cases that had received media attention. (There wasn’t enough information available to assess all of them.)

Since then, resources for preventing and responding to misconduct have proliferated. Professional associations for research-integrity officers exist on at least three continents. Yet we are not gathering reliable data about how often internal investigations are effective in responding to misconduct, or even the types and incidence of misconduct.

Last year, the UK Parliament reported that one in four universities failed to produce the required annual ‘narrative statement’ on whether and how many research investigations had been conducted. Only 58% could supply Internet links to statements from either of the previous two years.

It’s time to do better. Misconduct investigations will be more effective and potentially even lead to systemic improvements if they are framed in the right way, with questions, such as: ‘Do you want your reputation associated with an institution that countenances dishonest work?’, ‘How might this have been prevented?’, ‘What are our students learning about how to do research?’ and ‘Are other scholars depending on this work?’ Institutional procedures must not only protect the rights of whistle-blowers and those accused, but also protect students, colleagues, scholars and the advancement of knowledge. I propose a three-part approach: a checklist to strengthen investigations; the external peer review of investigatory reports; and the publication of findings. Last year, colleagues and I released a pilot checklist for investigations as a starting point (C. K. Gunsalus et al. J. Am. Med. Assoc. 319, 1315–1316; 2018). Across a range of disciplines, including surgery and air transport, checklists have been effective at improving processes and outcomes.

Even the best guidelines are of limited value without a mechanism for holding investigators accountable. We use peer review for research; we should do so for investigations, too. Consortia of universities could work together to peer review and certify reports before they are finalized.

And, when reports are complete, institutions should make them public. There will be many complex choices to work through to make this a reality. These concern protecting the vulnerable and the innocent, timing and confidentiality. Some current policies will require adjustment. Like-minded efforts are under way. Funders and researchers seek healthier working environments. Open science is spreading. Research papers are reporting more details about their methods and making more data available than ever before. Universities should apply similar principles to research-misconduct investigations. We need systems that are as rigorous, open and accountable as the research that institutions seek to produce.

One of the most inspiring misconduct investigations I know centred on a radiologist at the University of California, San Diego. He resigned after concerns about duplicate data were raised against two of his papers. But, after a faculty committee reviewing all 137 of his publications found that 60 were fraudulent or questionable, they published their findings and worked to correct the literature (R. L. Engler et al. N. Engl. J. Med. 317, 1383–1389; 1987). That was more than 30 years ago. It is time to start a fresh conversation. Checklists and greater transparency could make this kind of diligence the rule rather than the exception.

Nature 570, 7 (2019)

doi: 10.1038/d41586-019-01728-z
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up