Cleaning up the act

Conversations with dentists are always a strange, one-sided affair, and one practitioner is probably quite happy it stays that way.

If a patient did manage to get a muffled word in and ask Jon Sudbo how he ended up in his small country practice, the response, as he stuck the drill back in their mouth, would certainly raise an eyebrow.

Those familiar with the film Marathon Man might choose not to quote the findings of the inquiry into the former research scientist's record. "The bulk of Jon Sudbo's scientific publications are invalid due to the fabrication and manipulation of the underlying data material," investigators concluded.

The former Professor Sudbo's journey from the University of Norway to a dental practice in the Norwegian village of Seljord began when he admitted in early 2006 that he falsified data for studies published in The Lancet, the New England Journal of Medicine and the Journal of Clinical Oncology.

Now given limited permission to practise as a dentist in Seljord, he is an outcast from the scientific community, and his story provides a morality tale of how vanity can lead to research fraud.

Cases such as Sudbo's, along with those of Hwang Woo-Suk and Jan Hendrik Schon, make the headlines, although the sheer chutzpah of their deception perhaps made it inevitable that they would be caught and sent to the stocks to be pelted with the opprobrium of their peers. Yet there is concern that revelations about these mavericks have allowed the scientific community to congratulate itself on its ability to deal with misconduct. But such complacency is misplaced, as there seems little doubt that the integrity of research is still under threat.

There is widespread scepticism about Britain's system for dealing with research misconduct in which universities carry out their own investigations. The UK Panel for Research Integrity (UKPRI) issues guidelines on ethics and best practice rather than taking enforcement action. The problem, it is generally agreed, is that the current system is inadequate to cope not only with the sheer quantity of output but also with the close relationship between universities and industry.

At the beginning of August, Research Councils UK outlined proposals for a new research integrity body with powers to investigate allegations of misconduct and create a database of cases to allow vetting of would-be staff. Given that universities themselves are entrenched in their opposition to external regulation, devising a system that will please everybody could take some time.

In the meantime, there are many voices in higher education calling for an overhaul of the system for dealing with research misconduct. The loudest is probably that of Aubrey Blumsohn, whose bruising experiences as a whistleblower have been documented in the pages of Times Higher Education. In 2004, Blumsohn, a doctor and former senior lecturer at the University of Sheffield, alerted his managers to concerns about the conduct of his research unit's study of Procter & Gamble Pharmaceuticals' osteoporosis drug, Actonel. The drug company's research findings, he told Times Higher Education in 2005, had been published in the name of Sheffield academics even though the university's researchers had not carried out their own analysis of the trial data. This is known as "ghost-writing" and is perhaps the aspect of research in British universities that arouses most passion among critics.

Blumsohn was suspended from duties, eventually won a pay-off and is now a ferocious and prolific blogger on scientific misconduct. He is deeply unimpressed by the claim that universities are able to deal with misconduct internally.

"The long history of research misconduct all over the world tells us that university internal investigation is essentially a sham," he says.

"There are odd instances where internal investigations are plausible, but they are in circumstances where the system itself is not challenged - for instance, a PhD student who fiddles some results might be dealt with effectively or even brutally.

"But as soon as other issues are involved - the reputation of the institution itself and the reputation of powerful academics or commercial interests - these investigations tend to descend into farce."

He argues that the definition of research misconduct in Britain is wrong and needs to be widened beyond falsification, fabrication and plagiarism. "There are many people who have fought against this definition and argue that research misconduct is anything done intentionally that distorts or threatens to distort the scientific record.

"It's like defining murder based on the mechanism of death rather than the act of deliberate killing. You might define murder as stabbing someone with a knife, as hitting someone over the head or as running someone over with a car. Then suddenly you have a problem if someone kills by poisoning. That's not murder because it doesn't fit with your restrictive definition."

He claims that there is a gulf between aspirations for the UKPRI when it was founded and the effectiveness of the body that exists today. At the time of its formation in April 2006, Sir Ian Kennedy, UKPRI chairman, told Times Higher Education: "We looked at universities' procedures and many would not have survived attacks from anybody legally represented. They are not fit for purpose."

Blumsohn says today: "The individuals involved in the formation of the UKPRI have written clearly about the problems, yet we ended up with the formation of a body that is structurally incapable of addressing them."

Other campaigners also doubt if higher education has the will to police research misconduct effectively.

Peter Wilmshurst is an NHS cardiologist who has reported scores of research misconduct cases, including that of Anjan Banerjee, struck off by the General Medical Council in 2002 for falsifying scientific research a decade after he first admitted the offence. In the meantime, he had obtained a Master of Surgery degree from the University of London, using fabricated data.

"I am fundamentally opposed to self-regulation," says Wilmshurst. "It never works. This is partly because research fraud is prevalent and affects all universities. Universities A, B and C will not be too critical of university D today, because tomorrow it may be the turn of (university) A, B or C to have misconduct exposed, and they will be judged by the same standard.

"Usually the easiest thing to do is give the fraudster a good reference and get him to move on elsewhere. I know department heads and professors who were given good references to move on when they were suspected of fraud."

Wilmshurst has also written extensively about the relationship between industry (in particular pharmaceutical companies) and academia. He believes that universities are under enormous pressure to verify data that will allow the commercial exploitation of research.

"We are led to believe that university research is free of bias because that is what universities tell us. However, universities have to tell us that because that is their selling point and it is also how they sell their services to industry. What they say is, 'You should let us do the research for you because the public trust us and not you.'

"So industry gets universities to do its research and both sides know that industry pays over the odds. By paying well over the odds for the research, industry keeps the department on side."

The Committee on Publication Ethics (CoPE) was set up in 1997 as a forum for journal editors to discuss research integrity. It now has about 3,500 members, including the leading medical and scientific journals. Harvey Marcovitch, its chairman, is ideally placed to give an overview of how universities respond when confronted with allegations of misconduct.

"I get the impression as an experienced medical editor that there is often a concerted attempt to bury bad news. For obvious reasons, universities don't like to be exposed as dens of vice and corruption.

"We've also come across institutions that just don't know what to do. It's not a question of brushing it under the carpet because they don't have any brushes - or carpets. There is a sort of bewilderment and an ignoring of the problem in the hope that it will go away.

"One of the unsatisfactory things that CoPE discovered when we analysed eight years of work was that, in about half the cases, when an editor had reported a case to an institution, we never really got a resolution. It wasn't that they told us to get lost, it was just that the editors changed, letters got lost and so on. And that was disappointing."

Marcovitch describes the Norwegians' investigation into Sudbo as a model of how to deal with scientific fraud.

"Within six months they produced a report, they had exposed him as a fraud, they had stripped him of his right to practise, fired him from his jobs and written to all the journals to retract the articles, which I think was fantastic. I don't think anything like that could happen here."

Andy Stainthorpe, head of the UK Research Integrity Office (UKRIO), acknowledges that the system in this country is far from being a model of good practice, but insists that self-regulation is the only way forward.

"Where self-regulation is practised fairly and run on the right principles it can work well. There may be some who are critical of this approach, rightly or otherwise, because they consider that matters in which they may have been involved were not investigated appropriately. Other systems in use around the world are not perfect. The judicial system used to investigate misconduct by the Danes and the approaches used by the (US) Office for Research Integrity and National Scientific Foundation have weaknesses, too. In the Danish system, a High Court judge chairs the committee. They may be very knowledgeable about the law, but not about complex research.

"Such systems often have limitations about which allegations can be considered. This might be limited to research that is funded by the national research funder, and allegations can be presented only by researchers of a certain standing.

"What about other misconduct in research supported by other funders? Is this investigated to a different standard?

"The other part of the equation is taking things out of context, away from an organisation that has expertise in the field and greater awareness of local factors. It is important that an individual has a fair hearing."

Stainthorpe says that the UKRIO is very close to making available a standardised "toolkit" that universities can use to investigate allegations of research misconduct.

"The reputations of universities are based on having sound and fair principles for research integrity. Where an individual feels that the allegation has not been investigated effectively by a university, there are other options, including taking the matter to the UK Panel (for Research Integrity). In the past they might have felt that their only option was to take the case to the media.

"It might sound simplistic that by giving someone a manual you can facilitate a complex investigation, but I believe having the right procedure to follow will help organisations take investigations forward appropriately, particularly the ability of universities to access expert advice when needed."

Those at the top of higher education are also confident that they can deal with the problem within the sector. As Janet Finch, vice-chancellor of Keele University, puts it: "What has to be recognised is that each individual university is the employer of the people who do the research, and it is the university who holds the research grant.

"So I think it's the university sector that has to take responsibility for the issue. We have made huge strides in recent years when it comes to ethical approval of research projects before they start, and I see the move towards a more co-ordinated approach to research integrity as being the next step."

PLAGIARISM: A MURKY AREA IN WHICH SUSPICIONS ARE HARD TO ADDRESS

While falsification and fabrication in research is often an open and shut case, plagiarism is a grey area.

Lifting chunks of work from other sources is obviously wrong, but what about reproducing elements of your own work or copying the methodology of another researcher covering the same field?

"Plagiarism seems to be a pathetic crime perpetuated by the weak, because it's easy to do," says Harvey Marcovitch, chairman of the Committee on Publication Ethics.

"People must hope they can get away with it, and the vast majority of them probably do. We get regular complaints of plagiarised papers. Some of our members would say that sometimes plagiarism is acceptable."

Marcovitch believes there is a generation of scholars for whom the boundaries of research ethics are blurred. "We get the impression that sometimes younger researchers and clinicians are more puzzled by the concept of plagiarism because they've been brought up at school writing essays by using the internet. They see material on the internet as being freely available and that anyone can use it. The idea that there can be ownership of information is somewhat puzzling for some people."

There is also the murky issue of co-authorship where an academic's name is on a paper as a result of their seniority rather than as a reflection of their contribution to the work.

"But who's going to confront them?" asks Marcovitch. "Is the editor of a journal really going to phone up Professor X and say 'What did you really do?' - I don't think it's terribly likely.

"I remember telephoning a famous professor because I had been sent a paper for my journal with his name on, and there was something in it that puzzled me. So I rang and he said 'What paper?' and when I told him he clearly had no knowledge of it at all.

"What could I do? Take his name off? The embarrassment would have been enormous, so I let it go."

Although software that enables universities to crack down on student plagiarisers has been available for some time, programs to address academic plagiarism have only just become available.

A service called CrossCheck, using software developed by iParadigms, the makers of Turnitin, enables journal editors to check the originality of papers submitted to them, including those protected by access controls allowing only subscribers into online databases.

The British Medical Journal is currently deciding whether to buy the tool, which came on the market in June. But as Trish Groves, its deputy editor, explains, the use of plagiarism detection tools raises a whole new set of difficulties.

"It makes a lot of work. A human being still has to look at the percentage of material that overlaps and to decide how much of that is benign. The software just picks out words and phrases that match but doesn't express an opinion about them.

"We might use it when we think, 'Hmm, that looks a bit fishy', then it can be a tool you can use to do work that might otherwise take several hours."

A pair of US academics appear to have hit on a time-saving solution to the need to carry out lengthy comparisons between allegedly plagiarised material and the original.

Mounir Errami and Harold "Skip" Garner, both from the University of Texas Southwestern Medical Centre, run a website called Deja vu, which contains highly similar and duplicated material in the Medline database of biomedical research.

"We put the suspicious citations in the database," they say, "but we don't make the judgment call. It's up to the people using the database to decide if the material is plagiarised. We now have recruitment committees emailing us asking if applicants for positions have any entries."

UK PROCEDURE '20 YEARS BEHIND'

A comparison of the systems adopted in Britain, Denmark and the US for investigating allegations of research misconduct reveals considerable difference between approaches.

Here, it is up to universities themselves to deal with the problem and, as critics of the system point out, this can vary from institution to institution.

In order to help fashion a more systematic approach, Universities UK set up the UK Panel for Research Integrity in Health and Biomedical Sciences (UKPRI). The panel developed a programme for 2006-08 that aimed to provide guidance to those dealing with allegations and developing standardised procedures. The work is carried out by the UK Research Integrity Office (UKRIO). However, it has no powers to deal with any allegations of misconduct itself.

This contrasts with Denmark, which has a central body for dealing with research misconduct, the Danish Committees on Scientific Dishonesty. These consider allegations that individuals feel institutions have not dealt with properly, although they also hear from people wishing to clear their names.

The three committees cover health and medical science, natural, technological and production science, and cultural and social science. They are made up of researchers in the respective fields and chaired by a High Court judge.

Charlotte Elverdam, general counsel to the committee, says: "As we see it, the advantage of our system is that a judge has expertise in looking at evidence in complex cases. If a judge presides over the committee, then perhaps there will be less scepticism about its impartiality."

In the US, responsibility for dealing with allegations of research misconduct in biomedical sciences lies with the Office for Research Integrity (ORI), an agency of the Office of Public Health and Science. The National Science Foundation (NSF), a government agency that funds science and engineering research, deals with allegations in those fields.

Congress requires all agencies and research bodies to have policies for dealing with misconduct, which is defined in law. Universities and agencies funding research are required, under the legislation, to adopt procedures for dealing with it.

As soon as an institution begins an investigation, either the ORI or the NSF must be notified and the outcome reported. They have to be satisfied that the matter is dealt with properly.

Nick Stenek, an ORI committee member from the University of Michigan, says that universities initially made a case for self-policing to the politicians.

"Congress felt that universities didn't have an incentive to go out and investigate misconduct rigorously. There has now been enough research to suggest there is good reason to believe universities don't do as good a job if you leave them to their own devices."

Comparing the US system with ours, Professor Stenek said: "I think the UK ought to take a harder look at what needs to be done. I would say the UK is 20 years behind our system."

THE EXTENT OF MISCONDUCT

Like so many good ideas, the inspiration for doing a study on the extent of research misconduct in Britain was the product of an evening in the pub.

David Geggie, then a trainee Accident and Emergency (A&E) surgeon, was having a drink with colleagues when the issue of research came up.

"There is enormous pressure to do research to advance in medicine, and because of that pressure there could be the potential for bending the rules," he says.

"We had a discussion about how prevalent the problem was, and then I went to do a search to see what had been done on it. There were a couple of studies from the US, so I decided to see if the results were relevant in the UK among my peers."

His work remains the only study of the extent of research misconduct in the UK.

It went out to consultants at seven hospitals on Merseyside who had been appointed in the past five years. The results were published in the Journal of Medical Ethics in 2001.

Geggie, now an A&E surgeon working at Warrington General Hospital, found that 55.7 per cent of respondents had seen some form of research misconduct.

The most frequent were issues of authorship, with 29.9 per cent stating their names had been left off papers they had worked on.

In the case of personal misconduct, 5.7 per cent admitted to cheating to improve a grade, reporting results they knew were false, or plagiarism.