Lessons from the Pearce affair: handling scientific fraudBMJ 1995; 310 doi: http://dx.doi.org/10.1136/bmj.310.6994.1547 (Published 17 June 1995) Cite this as: BMJ 1995;310:1547
Belatedly, Britain should abandon its lax approach to scientific fraud
Last week Malcolm Pearce, a British gynaecologist, was removed from the medical register for fraud: he had published two papers in the British Journal of Obstetrics and Gynaecology describing work that had never taken place (p 1554).1 Less than nine months had elapsed between the whistle being blown on Pearce and his removal from the register. Outside observers might therefore conclude that, like other countries, Britain has established methods of preventing, detecting, and managing misconduct in research. They would be wrong. That the Pearce affair was handled well was unusual: the principal of Pearce's medical school knew what to do and was determined to do it—speedily and while protecting the rights of both the accused and the whistleblower. In most other medical institutions in Britain nothing would have happened2; the affair would have been brushed under the carpet, and the whistleblower would probably have been hounded out of his or her job.
Despite a report from the Royal College of Physicians,3 Britain has learnt little about handling fraud since the Darsee affair in the United States first brought the subject into prominence in 1983.4 This is despite a succession of other major scientific frauds in biomedicine.For example, the Office of Research Integrity, a branch of the US Public Health Service set up to investigate fraud, considered 73 cases in 1994.5 One particular abuse has indeed been tackled in Britain. Several general practitioners who engaged in fraud during drug trials have been struck off by the General Medical Council,6 but this has been largely because their frauds emerged through pharmaceutical companies' thorough auditing procedures and because the companies have taken an aggressive approach towards tackling fraud. The same has not applied within academia or the NHS. Until the Pearce case I knew of no other academic who had been investigated thoroughly. Within the NHS the postgraduate deans, nominated by the Royal College of Physicians' report to implement its procedure, know little about their responsibilities—or, indeed, about the report itself.
The Pearce affair has implications not only for the conduct of science and for how allegations of fraud should be tackled but also for medical journals. Commendably, the Royal College of Obstetricians and Gynaecologists commissioned a report into the role of its journal in this affair, and it has implemented the recommendations7: if followed more widely they should spell the end of amateurism in journals.
It is often unrealistic to expect a journal to detect fraud, but in this case the practices of the British Journal of Obstetrics and Gynaecology did not put as many barriers up to the publication of fraud as they might. Firstly, Pearce was an editor of the journal and the editor in chief was his head of department: this case shows how important it is for editors in that position to hand over consideration of the paper to someone else. Secondly, the journal did not review case reports at all, and, thirdly, the review of the clinical trial was clearly inadequate. Even with hindsight, the credulity in publishing the trial is reminiscent of that regarding Darsee, who claimed to have assayed 10different hormone concentrations twice weekly in blood obtained from rats' tail beginning at 1 week of age and continuing until death,8 or regarding Slutsky, who at one time was producing a paper once every 10 days.9 In this case a more disinterested editor might have questioned the fact that over three years Pearce purported to have collected 191 women with a syndrome so uncommon that a major referral centre was seeing only one or two new cases a month. Moreover, all of them had had a battery of complex tests, including karyotyping of both partners.
But, as the royal college's report makes clear, any journal can be the victim of fraud. Six years ago Drummond Rennie, deputy editor of JAMA, proposed a simple editorial audit on one in every 1000 papers submitted10: do the records exist, were the laboratory tests done, and what was the role of each “author”? Increasingly I warm to that idea. The cries that monitoring would discourage scientists from starting research have already been answered by Congressman John Dingell, the man responsible for making the American biomedical establishment take fraud seriously. “Scientists need to understand,” Dingell said, “that the best way, perhaps the only way, to avoid the threat of ‘science police’ is for scientists to show that they have the ability and will to police themselves. It is a matter of morality but also of self interest.”11
Another important aspect of the Pearce affair is the light it throws on gift authorship—the practice of treating authorship as something that is conferred as a benefit rather than earned through taking responsibility. For, as they were reminded in letters from the General Medical Council, coauthors have responsibilities to have done enough of the work to be called to account over it. Here none was qualified to be a coauthor: indeed, coauthorship was impossible since the work had not been done. Nevertheless, there were explanations why their names were included: the two junior authors had already been rebuked for askingquestions about details of the work and, as one of the them said at the council's hearing, they were “made to feel small.” The most senior author, Professor Geoffrey Chamberlain, who was also Pearce's head of department and editor of the journal, had twice asked for his name to be removed. Of all the abuses of scientific research, gift authorship is the most common and the most lightly regarded. Even the royal college's report, in comments that I disagree with, states that “Mrs Hamid's contributions…in the way of literature searches and writing of introduction and discussion components…justified her acceptance of coauthorship” and “Mr Manyonda's contribution … was at an intellectual level with significant contribution to the discussion … there is no ground for criticising Mr Manyonda for being a coauthor of the paper. He had accepted the existence of the case on trust from Mr Pearce.” This is an unusual attitude to authorship, at variance with accepted recommendations, which if followed will set the clock back.
Many people accept or confer gift authorship, detection is unlikely, and the rewards are obvious: tenure, promotion, research grants, and fame, especially in a society that measures worth by the weight of papers produced rather than their quality. Another reason why gift authorship is so common may be because the recommendations produced by the Vancouver group, an international group of medical journal editors, are difficult to understand12: the group should simplify them and also print the masterly table of legitimate and non-legitimate grounds for authorship produced by Ed Huth, a former editor of the Annals of Internal Medicine and member of the Vancouver group.13 Most importantly, however, we should revise our criterion of worth. As recommended by other bodies,14 15 appointment committees in Britain should follow the longstanding example of Harvard (requiring candidates for a full professorship, for example, to submit copies of only their 10 best articles).
Crucially, however, the Pearce affair raises questions of management. Firstly, we must accept that fraud exists, though with an unknown prevalence: estimates vary from 27% of scientists encountering 2.5 episodes over 10 years16 through 0.28% in audits of cancer trials17 to one new case per million population every year (P Riis, personal communication, 1995). Next, the universal lesson is that institutions are not good at policing themselves, so several countries have set up bodies specifically to do this for them, ranging from the Office of Research Integrity in the United States to central committees on scientific dishonesty in the Nordic countries and Austria. The latter committees teach good research practice, advise whistleblowers, are notified of all cases, and may undertake investigations themselves: moreover, they monitor every case and publish annual reports.
A central committee would also seem the most suitable pattern for Britain, particularly as a single body could acquire the necessary experience and skills that more peripheral bodies would lack. On Danish experience, three quarters of the work could probably be handled by the secretariat (disputes about who owns data and authorship, for example), but some would need “due process,” and for this reason the presence of a judge on the committee, as in the Nordic countries, would be important. Some link with the General Medical Council, which has statutory powers over doctors, and the statutory bodies would be inevitable. For this time the public outrage that patients might have been put at risk by Pearce's medical frauds means that the subject will not go away. Given its pioneering work, the Royal College of Physicians should seize the initiative again, convene another meeting of interested parties, and implement a workable solution. The time has come for Britain to abandon its lax approach to scientific fraud.