Science is knowledge gained from reproducible observations or experiments. Yet in a commentary in Nature in May last year, researchers from biotechnology company Amgen reported that the findings in 90% of landmark cancer papers from academic laboratories could not be reproduced, even with the help of the original study authors.
Clearly, at least in the biomedical research sector, things are going badly wrong.
While part of this lack of reproducibility is due to incompetent or sloppy research, some is undoubtedly due to deliberate misrepresentation, fabrication or falsification of data - research misconduct. A review from October 2012 in Proceedings of the National Academy of Sciences (PNAS) showed most of the rapidly growing number of retractions of papers were due to research misconduct, rather than inadvertent mistakes.
In the US, the Office for Research Integrity (ORI) handles more than 200 allegations of research misconduct each year. It ensures that every allegation of misconduct in National Institutes of Health-funded grants is investigated, but rather than carrying out the investigations itself, it provides advice to institutions and checks to see they do the investigations properly.
The ORI also provides training videos, forensic image analysis tools, advice to whistle-blowers and institutions, analyses data, and – importantly – makes findings public.
By contrast, Australia does not have an ombudsman or an office for research integrity, and the few Australian cases of research misconduct that ever reach the public domain suggest they are not handled well.
To get on track …
Australia needs an ombudsman or an office for research integrity to:
- provide advice to, and oversight of institutions handling allegations of misconduct
- assist whistle-blowers
- collect data on research misconduct
- update the Australian Code on the Responsible Conduct of Research
- cover published research that is not funded by the National Health and Medical Research Council (NHMRC) or the Australian Research Council (ARC)
- provide an avenue for appeal
An ORI or an ombudsman would increase the public’s confidence in science, reduce wastage of research funds, and help protect the careers of responsible scientists as well as the subjects of clinical studies.
Because there’s no national body to promote research integrity, those individuals in institutions that are landed with the task of handling an allegation of misconduct are usually inexperienced, and therefore bound to make errors.
Indeed, the same mistakes are repeated in case after case, whether it’s failure to give support the person accused, failure to treat allegations seriously, superficial investigations by panels that lack experience or relevant scientific expertise, ignoring real or perceived conflicts of interest of the panel members, no accounting for inconsistencies between employer law and the Australian Code for the Responsible Conduct of Research, or failing to provide reports or retract papers.
In addition, the strong incentives to protect institutional reputations combined with lack of oversight can lead, at least hypothetically, to denial or cover-up.
It is unlikely that Australian researchers are any more honest or dishonest that those from other countries, as they are subject to the same short-term grants, lack of job security, high levels of competition for funding, demands to demonstrate constant productivity, and intolerance of experiments that don’t work.
Even without such powerful incentives, it’s inevitable some researchers will bend or break the rules. For institutions, and for Australian research as a whole, the issue is not whether misconduct will occur: it’s how misconduct is handled when it does occur.
Independence is key
Australia needs an ombudsman or an office for research integrity that is independent of the universities and research institutions, and has a mandate to consider allegations about any published research.
Its mission would be to protect the integrity of the scientific literature, and to promote integrity in the practice of science. It would not operate as an inspectorate, nor demand researchers attend training sessions and sign documents promising to “be good”; it would operate according to a “fire alarm” model, in which anyone can call for help if they have concerns about breaches of the research code.
If someone raised a concern with the office directly, it would inform the relevant institution, and if someone raised a concern with an institution, they would have to notify the office.
An Australian research integrity office would need to have teeth. Although it would not necessarily need to have the power to impose penalties itself, it would need to be able to stop grant-funding from universities or institutions that failed to co-operate.
This way it could ensure that institutional investigations were carried out fairly and thoroughly. It would publish regular reports showing how many cases were registered, and what the outcome was.
An attempt was made to improve research integrity governance in Australia with the establishment in February 2011 of the Australian Research Integrity Committee (ARIC), but that body can only consider questions of procedure, and not appeals based on the evidence or merits of a case. Since its inception, ARIC has considered four cases, of which only one was deemed to be within its scope.
Australia can - and must - do better. Establishing a proper ombudsman or an office for research integrity is a start.