Education and debate

Dealing with research misconduct in the United Kingdom

An American perspective on research integrity

Conduct unbecomingthe MRC's approach

An editor's response to fraudsters

Deception: difficulties and initiatives

Honest advice from Denmark

Dealing with research misconduct in the United Kingdom

After years of inactivity over the problem of research misconduct in the United Kingdom, there is now an enthusiasm and drive to do something. But how should medical fraud be tackled? Representatives from medical journals (both British and American), the Medical Research Council, a medical school and a medical charity, and a member of the Danish Committee on Medical Dishonesty give their views on this important topic.

An American perspective on research integrity

Drummond Rennie, deputy editor (west), JAMA. 

Institute for Health Policy Studies, San Francisco, CA94109, USA

rennie{at}itsa.ucsf.edu

An allegation of scientific fraud can ruin the careers of both the accused and the accuser, divide faculties, bring a research institution's functions to a halt, provide a field day for the media, and, when the scientific establishment is unprepared, result in a loss of confidence in the entire research enterprise. Yet, despite repeated demonstrations that this is the case, scientists are still reluctant to face up to such an unpleasant problem. Three years ago, at a meeting on research misconduct held by the BMJ, I warned that many extremely embarrassing incidents at a variety of institutions would be required before anyone took any action in the United Kingdom. This seems to have been borne out. At a meeting organised by the Committee on Publication Ethics (COPE) in London, I was depressed that so few seemed to have paid the slightest attention to the rich, well documented and instructive experience of the United States, where an energetic attempt to face up to the problem has been made. Such parochialism may doom the UK to repeat the many mistakes already made by others.

To the American observer, the news from the UK about incidents of misconduct in research is, as baseball's greatest philosopher, Yogi Berra, remarked, "déjà vu all over again." I began to be seriously concerned about the problem in 1979 when, as deputy editor of the New England Journal of Medicine, I had to help sort out a serious case of fabrication and of plagiarism during review.¹ In the next few years, several other cases, for example, that of Darsee, involved that journal.² The US Congress reacted to the media attention to these and many other cases with more than a dozen highly publicised congressional inquiries, the first in 1983 under then Congressman Al Gore.³ The response of each research institution varied but was only too often characterised by circling the wagons, denial, and cover up. Under the eyes of the press, each institution would hurriedly patch together its own process, assembling ad hoc panels, sometimes with glaring conflicts of interest. The results were frequently slow, bungled, idiosyncratic, and unfair to almost everyone.³

View larger version (77K): [in this window] [in a new window]

Against this background, several meetings were held in the 1980s, jointly arranged by the American Association for the Advancement of Science and the American Bar Association, to frame some rational response to the growing public perception that fraud in science was rife. Senior scientists kept insisting that this perception was false, but since their assertions were made in the absence of evidence they appeared self serving and unscientific. In 1989, federal regulations were issued governing research sponsored by the US Public Health Service and the institutions where this research was conducted. The rules provided a definition of misconduct and a process that institutions had to follow when an allegation was made. An office was set up, now called the Office of Research Integrity, to oversee and enforce the institutions' compliance. Since the Public Health Service sponsored most publicly funded biomedical research, its definition and process became the standard adopted by research institutions in the United States.

The past nine years has seen a few high profile cases (for example, the Gallo case and the Imanishi-Kari or "Baltimore" case) thrown out, but a good many others have been concluded without too much fuss. A universal definition and a set of procedures to be applied to research conducted under the aegis of every US government agency, from the National Institutes of Health and the National Science Foundation to the National Aeronautics and Space Admininstration, have still not been agreed. Heated argument still continues, and is unlikely to subside when the White House committee charged with putting together government-wide procedures reports this summer. We can, however, reach a few general conclusions.

	An assessment

When the Office of Research Integrity adopted a "scientific dialogue model," evaluating cases according to the way scientists look at the evidence, judgments would be challenged and resolution would be slow and incomplete. Scientists are not trained in conflict resolution; their intuitive response is usually wrong and they tend to set up shaky ad hoc procedures that do not guarantee the accused notice of all the charges, the opportunity to respond to all the charges and to the evidence, and a decision based on rigorous standards.⁴ When the Office of Research Integrity changed to a more "legal" method, following the procedures of administrative law, cases would be handled more expeditiously and were less prone to challenge. A Commission on Research Integrity, of which I was a member, examined the issue. Its report in late 1995 broadly followed an earlier attempt by the National Academy of Sciences. ^{5 6} It refined and extended the definition, basing it on the principle of telling the truth, and suggested a whistleblowers' bill of rights and responsibilities. As happened with the original regulations, the commission's definition has been widely reviled by a scientific community that still has difficulty coming to terms with the basic fact that together with the privileges of a profession come responsibilities.

All sorts of other issues remain unsettled. ^{7 8} For example, the use by whistleblowers of a law dating back to the Civil War that permits their bringing action to recover misspent government money. However, the generally quieter tone in the United States seems to reflect an understanding that matters are now being dealt with fairly routinely and competently. It also reflects the fact that since 1992 the Democrats have been in a minority in the Congress, and Congressman John Dingellwho pursued the issue so remorselessly for so long, and kept it on the front pages of the papershad to relinquish his powers to less interested Republicans.

	Lessons for Britain

The idea that the situation in the United States is uniquely bad rings hollow in the face of growing numbers of cases in Britain, Germany, and elsewhere. It has not been shown that scientists in Britain differ importantly from those in the US. Institutional denial in the United Kingdom is therefore no longer a sensible option. It would help if we were all to stop registering shock and recognise that the bestowal of a scientific or medical degree is not accompanied by a guarantee of honesty. The only useful approach, therefore, is to assume that, in common with other crimes, a certain proportion of our colleagues will plagiarise, fabricate, and falsify the evidencein other words, that scientific misconduct will occur infrequently but regularly. This routine approach requires that a definition of misconduct be agreed on and promulgated, not least because it is unfair to accuse people unless they have had the chance to know what is unacceptable. A corollary of this is that scientists must be taught about good and bad research practices and about research ethics. Courses in research ethics are proving useful in the United States, but my bet is that if senior scientists make efforts to become closer mentors to their juniors, this will raise standards considerably.

When allegations arise, research institutions must have an effective procedure in place. The requirements for such a procedure are listed in the box.

The legitimacy of institutions, whether or not they are funded by the public, ultimately depends on public confidence, and the public interest requires that the process and the resolution of cases be made public. Morale in the institution where the problem occurred can be devastated. An essential step in rebuilding trust is to show that justice has been done.

It makes no sense to leave the regulations governing research misconduct to be developed by individual research institutions, not least because some central oversight to ensure compliance is wise and necessary. The credibility of the process is greatly enhanced by having universally applicable rules, developed and supported by prestigious scientific and medical bodies. In addition, the central body, which must have the power to review cases and to sanction institutions that do not comply, or which fail to protect responsible whistleblowers, should publish their experience.

Other improvements would help. For example, we should stop being led astray by pretending we know the cause when in fact we can only speculate. We should ignore whining about the supposedly awful pressures of "publish or perish" when we have little credible evidence on what motivates misconduct, nor on what motivates the conduct of honest, equally stressed colleagues. Laziness, desire for fame, greed, and an inability to distinguish right from wrong are just as likely to be at the root of the problem. There is an urgent need to encourage investigation in this area, including confidential experimental audits.⁹ We must recognise that scientists have expertise in the interpretation of scientific evidence but little training in the dispassionate adjudication of cases, so we need help from lawyers. For example, we tend to absolve dishonest colleagues because their fabricated results are "correct," even if invalid.⁸ And we all tend to condemn as crooks those who are merely "uncollegial" and to condone the real crimes of those whom we like. We forget that the legal profession has had a great deal of practice with sorting out guilt from innocence, and they are the first people we should consult when putting together regulations to ensure that they will work and withstand legal challenge.

	Where to begin?

Implementation will take a long time, and whatever is decided on will offend some group. I would start with meetings modelled on those of the American Association for the Advancement of Science and the American Bar Association, and pay great attention to experience in the United States. Though I would model my approach on the one taken by the National Academy of Sciences and the Commission on Research Integrity, ^{5 6} I do not presume to tell Britain what model it should adopt, whether American, Danish, or some new one. But I do know that this represents a great opportunity, and the sooner you bring representatives of scientific and medical societies together with representatives of research institutions, commercial laboratories, pharmaceutical companies, members of the bar, and politicians, the sooner you will get a definition and procedures generally approved. And the sooner that happens, the sooner those involved will get justice; the sooner those high up in institutions will stop looking foolishly unprepared; the sooner the public will feel its concerns are taken seriously; and the sooner this initiative, started in Britain by far sighted medical editors, will be realised.

	References

1. Broad W, Wade N. Betrayers of the truth. New York: Simon and Schuster , 1982.
2. Relman AS. Lessons from the Darsee affair. N Engl J Med 1983; 308: 1415-1417[Medline].
3. Rennie D, Gunsalus CK. Scientific misconduct. New definition, procedures, and officeperhaps a new leaf. JAMA 1993; 269: 915-917[Medline].
4. Gunsalus CK. Preventing the need for whistleblowing: practical advice for university administrators. Sci Eng Ethics 1998; 4: 75-84.
5. Commission on Research Integrity. Integrity and misconduct in research. Report of the Commission on Research Integrity to the Secretary of Health and Human Services, the House Committee on Commerce and the Senate Committee on Labor and Human Resources. Rockville, MD: US Department of Health and Human Services, Public Health Service , 1995.
6. National Academy of Sciences. Responsible science. Ensuring the integrity of the research process. , Vol II Washington, DC: National Academy Press , 1993.
7. Parrish DM. Improving the scientific misconduct hearing process. JAMA 1997; 277: 1315-1319[Abstract].
8. Youngner JS. The scientific misconduct process. A scientist's view from the inside. JAMA 1998; 279: 62-64[Free Full Text].
9. Rennie D. Accountability, audit, and reverence for the publication process. JAMA 1993; 270: 495-496[Medline].

Conduct unbecomingthe MRC's approach

Imogen Evans, research strategy manager. 

Medical Research Council, London W1N 4AL

imogen.evans{at}headoffice.mrc.ac.uk

Scientific misconduct is taken seriously by the Medical Research Council. In its work as as an employer and substantial provider of research funding, two responsibilities are the management and training of researchers. If research is not conducted with integrity, the results cannot be trusted and the implications for the wider scientific community are both important and unacceptable. To reassure the scientific community that this message is more than a laudable intention, we have devised a specific policy and procedure for handling allegations of scientific misconduct.¹ The policy formally covers all staff employed in MRC research units and institutes, as well as visiting scientists, but researchers in universities and elsewhere who are awarded MRC grants will also be expected to operate under similar policies.

	MRC procedure

Having reviewed extensively the existing European and US guidelines on the management of misconduct cases, we decided to introduce a separate procedure to deal with allegations. These had been addressed previously under our normal disciplinary procedures. For this purpose scientific misconduct means fabrication, falsification, plagiarism, or deception in proposing, carrying out, or reporting results of research and deliberate, dangerous, or negligent deviations from accepted practice in carrying out research. It includes failure to follow established protocols if this results in unreasonable risk or harm to human beings, other vertebrates, or the environment and also the facilitating of misconduct by collusion in, or concealment of, such actions by others. Misconduct does not include honest error or honest differences in the design, execution, interpretation, or judgment in evaluating research methods or results of misconduct (including gross misconduct) unrelated to the research process.

We devised an essentially stepwise approach that sets out a sequence of stages for the investigation of an allegation. This procedure, which is shown in the box, was designed to achieve a number of aims, including appropriate confidentiality (particularly should an allegation prove groundless), protection of whistleblowers, and natural justice towards those who are the subject of the allegation. The director of the relevant MRC establishment normally has primary responsibility for adhering to the procedure. If he or she is not perceived as being impartial or is the subject of the allegation, however, the responsibility falls to the executive director of council.

The design of the procedure ensures that scientifically expert assessors evaluate the evidence and draw conclusions. There are also clear commitments both to inform the scientific community, sponsors, and other interested parties in the event of any proven allegation of misconduct in relation to published work and also to restore the reputations of those subject to ill founded, even potentially malicious, accusations. In this last instance, it is then a matter of principle that the MRC will pursue action against the complainant.

	Importance of good practice

Notwithstanding its seriousness, scientific misconduct is an extreme and unusual occurrence. Of greater day to day importance to the MRC is the need to ensure that standards of good practice are maintained in our establishments. We are therefore currently preparing a guide to good research practice to be published later this year. The intention here is to provide information and guidance to staff and visitors to our establishments on the key components of the contemporary research process, including supervision and training of researchers; the scientific method; research data, including gathering, storage, and retention; and publication of results, including authorship and methods of publication. As with our scientific misconduct procedure, we plan to distribute the information widely in the hope that the MRC approach will interest the broader research community.

	References

1. Medical Research Council. MRC policy and procedure for inquiring into allegations of scientific misconduct. London: MRC , 1997(MRC ethics series.)

An editor's response to fraudsters

Michael J G Farthing, professor. 

Digestive Diseases Research Centre, St Bartholomew's and The Royal London School of Medicine and Dentistry, London E1 2AD

Fraud in biomedical research is alive and well, and apparently flourishing. Despite increasing publicity during the past 25 years, its magnitude is unknown and its detection largely serendipitous. ^{1 2} Research fraud is committed by general practitioners, the young and inexperienced, and those at the very top of the profession.³ Fraudsters often arouse suspicions for some time before they are detected externally or a whistleblower feels secure enough to make his or her suspicions known. The most serious cases involving doctors are drawn to the attention of the General Medical Council, and the guilty invariably lose their place on the medical registrar. In Britain, where no agency exists to deal with less serious cases of research misconduct by medical practitioners and fraud in non-clinical scientific disciplines, the fate of those who are dealt with by internal institutional review is unclear. The Royal College of Physicians made recommendations on how academic institutions might handle suspected research misconduct,⁴ but there has been no national review of the implementation of these recommendations nor have institutions been invited to report on their activities in this area.

	What can editors do

I first came face to face with research misconduct as a part time editor of a specialist journal. I reviewed the cases detected during my first year and found examples of overt plagiarism, "salami slicing" of one piece of research to create as many articles as possible, duplicate publication, and the submission of manuscripts that had not been approved or even seen by coauthors.⁵ It might be argued that none of these cases amounted to serious research fraud; indeed, in every case the ultimate crime was prevented since all were detected before publication.

View larger version (52K): [in this window] [in a new window]

From an editor's point of view, doing one's duty is simplified if fraudulent material is actually published. In this case, retraction or explanation is required, the matter is in the public domain, and the offenders face public disgrace. It is up to others to decide whether there is a case to answer elsewhere, such as before the General Medical Council. The dilemma arises when there is clear evidence of research misconduct, but the information remains on the editor's desk. What is the editor's duty then? Is a standard rejection letter sufficient? Should an additional paragraph be added to explain to the authors exactly why the manuscript has been rejected? Under these circumstances an editor would need to be extremely certain of his or her grounds to avoid the threat of libel. Or should the editor write to the head of department (often a coauthor), or a dean or vice chancellor, explaining the concerns and perhaps requesting a full internal review?

An editor has no mandate to investigate suspected research misconduct. For overt plagiarism the case is usually secure and can be quantified by calculating the proportion of the manuscript that has been taken from elsewhere. It is often extremely difficult to investigate one's suspicions about "the perfect study" in which the data presented do not seem to have been generated in a "biological system." Although the opinion of a statistician can be helpful, uncertainty often remains. Examination of original research records is usually required, and generally these would need to be obtained at extremely short notice. It is unlikely that an editor would be able to achieve this, and anyway, is it really an editor's job?

	Committee on Publication Ethics

Last year about 20 frustrated editors got together to form an informal group, the Committee on Publication Ethics (COPE).⁶ This group had no pretensions that it was formed to stamp out research fraudit was a "self help" group for editors to discuss some of the dilemmas raised above and to seek advice on how they should be handled. In its first year the committee examined 17 cases. These included examples of plagiarism (one case involved several examples), suspected data fabrication, a serious conflict of interest between the reviewers and authors, and ethical issues relating to human research studies. All cases are brought anonymously, although we keep accurate notes of our meetings and plan to produce an annual report which will include the cases discussed. In 1997, the committee organised a meeting for editors entitled "Research misconducthow should editors respond?" and a second one on detecting fraud is planned later this year.

Many frustrations remain. An editor has no mandate to investigate suspected fraud and is therefore unlikely to be able to present a fully investigated case to an author's institution. There may be only suspicions. Is it right that the matter should then be allowed to rest? Members of the committee feel strongly that this should not be the case but recognise that this is not a job for the group. When the case is clear cut, editors do sometimes take the matter into their own hands and punish authors who indulge in duplicate publication by refusing to consider articles from that author for a statutory time period, say three or five years.⁷ We believe that this will only deal with the tip of the iceberg. Research fraud should be detected and reported well before its products land on an editor's desk.

	Need for an independent agency

Is it really such a difficult problem? Other countries are actively managing research misconduct and have left Britain way behind. The United States set up an Office of Scientific Integrity in 1990. This was replaced by an Office of Research Integrity in 1992 and was soon followed by similar agencies in Denmark, Norway, Finland, and Australia.³ These agencies rely on expertise provided by scientists, clinical investigators, and other academics, but they function independently of individual academic institutions, funding agencies, or other professional regulatory bodies. Surely the time has come for the speedy establishment of a similar agency in Britain? It is absolutely vital that we act promptly if research fraud and other forms of misconduct are to be prevented and detected. All of us who are involved in the many aspects of research and publication ethics must have access to an independent agency with which we can air our concerns when suspicions are raised.

Whistleblowers, possibly the most important tool for detecting research fraud, must, at least initially, have anonymity and full protection. Since working with the Committee on Publication Ethics I have been approached by a number of whistleblowers from various institutions, each asking for advice. My experience suggests that these people are not treated appropriately by their own institution. They are sometimes discouraged in pursuing their claims and are even threatened with career disruption or dismissal if they fail to keep quiet. Similarly, editors will be reticent about making accusations to deans and vice chancellors unless the case is secure; as discussed previously, full investigation is often impossible or inappropriate. "Do we need research police?" asked Professor Geir Jacobsen.⁸ If the policing means prevention and detection of research crime then the answer is unequivocally "yes."

	Part of public health

Some would argue that this is all a fuss about nothing. Most research misdemeanours are minor and cause no harm other than adding a few inconsequential inaccuracies to the biomedical literature. I would argue that the preservation of research integrity is just another aspect of public health. We have a drinking water inspectorate to protect domestic water supplies. We are about to have a food standards agency, an independent watchdog to ensure that the food we buy and eat is safe. Surely public concern about the entry of erroneous material in biomedical publications on health and disease is at least of equal importance. Fortunately, the regulation of clinical trials of new drugs is generally of high quality, but doctors still try and fudge the data, usually for pecuniary gain; and there is always the risk that an ineffective or possibly dangerous drug might be used to treat patients for many years before its lack of efficacy is detected. Similar concerns might surround the dishonest reporting of the safety of new surgical procedures by selectively discounting the cases that did not go quite so well, and inflating the efficacy of new diagnostic test, again by data selection.

Although editors may be regarded as custodians of biomedical publication, their ability to preserve the nation's research integrity is limited. There is an urgent public need for an independent agency to formalise the maintenance of research standards and the detection and prosecution of fraudsters. Clearly such an organisation would need to work closely with other bodies that are responsible for maintaining professional standards such as the royal colleges and the General Medical Council. One way forward would be for the government to commission a report along the lines of the James report on food standards⁹ and then use similar mechanisms to establish an independent Research Standards Agency perhaps with a parallel structure to that proposed for food.

	References

1. Kohn A. False prophets. Oxford: Blackwell , 1986.
2. Lock S. Misconduct in medical research: does it exist in Britain? BMJ 1988; 297: 1531-1535.
3. Lock S. Research misconduct: a resume of recent events. In: Lock S, Wells F, eds. Fraud and misconduct in medical research. London: BMJ Publishing Group , 1996Royal College of Physicians of London.
4. Royal College of Physicians. Fraud and misconduct in medical research. London: RCP , 1994.
5. Farthing MJG. Research misconduct. Gut 1997; 41: 1-2[Free Full Text].
6. Smith R. Misconduct in research: editors respond. BMJ 1997; 315: 201-202[Free Full Text].
7. Doherty M. The misconduct of redundant publication. Ann Rheum Dis 1996; 55: 783-785[Free Full Text].
8. Jacobsen G. Do we need a research police? J R Coll Physicians Lond 1997; 31: 8-9[Medline].
9. James P. Food standards agency report. London: Cabinet Office , 1997.

Deception: difficulties and initiatives

Cyril Chantler, principal, ^a Shireen Chantler, secretary, Scientific Advisory Committee. ^b

^a United Medical and Dental Schools of Guy's and St Thomas's Hospitals, Guy's Hospital, London SE1 9RT, ^b Foundation for the Study of Infant Deaths, London SW1X 7DP

Correspondence to: Sir Cyril Chantler C.Chantler{at}kcl.ac.uk

That fraud and misconduct occur in research is not in doubt.¹ Nor is there any question that they continue to pose a problem, despite recommendations to detect and eliminate them. The General Medical Council is clear that research misconduct is wrong and, in most cases brought to its attention, amounts to serious professional misconduct.² Nine of the 10 doctors who appeared before the conduct committee in the past five years have been suspended or removed from the medical register.³ We do not know, and it is probably impossible to know, how prevalent research misconduct is. Relatively few cases are reported in relation to the increase in medical research, though there are suspicions that it is more common than these cases suggest.⁴ But this is to miss the point: every single case reduces public confidence, abuses the use of public and charitable funds, and causes insult and frustration to the vast majority of careful, honest workers.

	Recognising the difficulties

Firstly, we need to recognise the difficulties. Much research in medical schools and the NHS is carried out by people who are not members of the medical profession, and many are not accountable to a regulatory body. Routine audit of scientific activity by internal or external mechanisms would be difficult. The scope of inquiry is so great that a large panel of experts would be required, the expense would be great, investigators would be removed from their prime endeavours, and the efficacy would be doubtful. Even financial audit does not prevent fraud.

Suspecting a case of misconduct or fraud is different from providing evidence to prove the case beyond reasonable doubt to a regulatory body, to a university inquiry, or to the courts of law. Often the circumstances are not clear cut and depend on the interpretation of actions alleged to have taken place. When an investigation is started under the institution's disciplinary procedures, the miscreant may resign before the process is completed. If the individual then applies for another post, the suspicions may not be passed on to the new employer because of the risk of legal action. Thus, some individuals may move to another department, where the process may be repeated.⁵

	Guidelines

Guidelines for the prevention and investigation of complaints were published by the Royal College of Physicians of London in 1991.⁶ This report defines scientific misconduct as including piracy, plagiarism, and fraud and provides a description of each. It includes guidelines for investigators in scientific research that were prepared by the Harvard Medical School and guidelines on authorship drawn up by medical journal editors. Guidelines also provide students with information on plagiarism.

View larger version (58K): [in this window] [in a new window]

	GMC initiative

The GMC, aware of the widespread concern about research fraud that was reflected in editorials in the BMJ and Lancet, ^{7 8} convened a meeting with representatives of the medical royal colleges and heads of medical schools to discuss a way forward. As a result of that meeting, a committee has been established. It has set in train a review of the Royal College of Physicians' guidance, drawing on advice from medical editors and others, with a view to producing clear advice supported by the colleges, universities, and the GMC. All university medical schools are being asked to submit their procedures to this committee so that best practice can be recommended. It is likely that those schools which have faced problems will tend to be the ones with the most developed procedures, but all should be encouraged to review their procedures now. Certainly all should have a named person or persons to whom a complaint can be addressed, in complete confidence, and whistleblowers must be respected and protected. Doctors have a responsibility to take action if they feel a colleague's conduct, performance, or health may place patients at risk, and a similar responsibility to report concerns about scientific research has been placed on them in the review of good medical practice recently undertaken by the GMC.

	Good practice

Clear guidelines on good practice in the conduct of scientific research should be available to all who undertake it. All should receive formal education on ethics and good practice in research, and the fact that they have done so should be recorded and audited. Some commercial organisations require all primary data to be recorded in bound volumes (not loose leaf volumes) with numbered pages. All alterations and deletions have to be signed and dated, and the printout results from scientific equipment have to be pasted into the books. The books have to be inspected and signed off regularly by the head of the research group (who has to be knowledgeable about the work), and when they are complete they are securely stored. Obviously, this requirement is necessary for patent and commercial reasons, but it does establish a clear audit trail. The lack of such a trail can impede investigations into misconduct and the investigators are left to choose between the veracity of different accounts and the different perceptions of the same event.⁶

	Editorial input

The editors of medical and scientific journals, who have done much to draw attention to the problem, could perhaps do more to help eliminate it. Rather than simply rejecting articles they find suspicious, they should be encouraged to express concerns to the author or contact the named designated person in the organisation (see above) that employs the lead author, or both.

	Investigation and inquiry

After a complaint has been received, the responsible individual to whom the complaint is addressed should invite the person making the allegation to submit a detailed statement in support, while guaranteeing his or her anonymity. If the allegation is frivolous, unsustainable, or unfounded, it should be dismissed and the person making it informed accordingly. However, the nature of the complaint and the action taken should be recorded, and steps should be taken to ensure anonymity. If there is prima facie evidence to support the complaint, or there is insufficient information, an inquiry should take place.

It has been suggested that an office for investigating scientific fraud should be set up, as in the United States. As noted above, the wide scope of scientific inquiry would require a large bureaucracy of uncertain efficacy to support this, but obviously it is a suggestion that merits consideration. An alternative view is that the responsibility for dealing with complaints has to rest on the employer, be it a pharmaceutical company, university, hospital, or whatever. Each organisation should be required to set up a scientific misconduct committee with external representation. If, as the first step, the responsible person decides that the complaint is trivial and no further action is to be taken, the committee should review the decision and record that it has done so. Where there is a case to be answered, the committee would undertake this task, and all cases must be concluded, even when the individual about whom the complaint has been made has left the organisation.

	National audit

A small national body could be established for audit purposes. All organisations undertaking medical research would be required to report regularly all complaints received and the action taken. The national office could then audit this informationin other words act as the agency for quality assurance in this area. It could also act as a resource for advice on good practice and how to deal with specific complaints where necessary.

	Repercussions

Finally there is the problem of what to do when misconduct has been proved. Presumably gross misconduct will lead to dismissal from employment and, in the case of doctors, referral to the GMC. For non-medical scientists the problem is more difficult to resolve. Information on those found guilty of misconduct could be kept by the national office and employers could check with the office before offering employment to non-medical scientists.

	Other initiatives

There is an increasing willingness now to do something about research misconduct. As well as the deliberations of the college of physicians' committee, the National Academic Policy Advisory Group led by the Royal Society is also undertaking a comprehensive review of the problem and will make recommendations. These initiatives are welcome and no doubt will lead to further debate which, hopefully, will not be protracted. We need to develop systems that inspire public confidence, protect the integrity of medical research and of individual researchers but are not overly bureaucratic and intrusive.

	References

1. Lock S, Wells F. Fraud and misconduct in medical research. , 2nd ed. London: BMJ Publishing Group , 1996.
2. General Medical Council. Good medical practice. Guidance from the General Medical Council. London: GMC , 1995.
3. Dyer C. Doctor admits research fraud. BMJ 1998; 316: 647.
4. Lock S. Fraud in medical research. J R Coll Physicians Lond 1997; 31: 90-94[Medline].
5. Dalton R. International recruitment highlights need to track scientific behaviour. Nature 1996; 383: 107-108[Medline].
6. Royal College of Physicians. Fraud and misconduct in medical research. London: RCP , 1991.
7. Time to face up to research misconduct. BMJ 1996; 312: 789-790[Free Full Text].
8. Dealing with deception. Lancet 1996; 347: 843[Medline].

Honest advice from Denmark

Povl Riis, professor. 

Ministry of Science, DK-2900 Hellerup, Denmark

The United Kingdom faces the same problem as a number of other European countries. The responsibility for unmasking and preventing research misconduct within medical science has been, and still is, part of the remit of a number of administrative, political, and scientific bodies. As a consequence, the natural law of shared responsibility comes into force, according to which the sum of shared responsibilities rarely or never amounts to the whole.

The latest figures from the Nordic countries, which have social structures and resources similar to Britain, show that 1-2 cases per million inhabitants are referred to their national systems annually. For Britain this would mean 60-100 cases each year, which is not a frightening prospect, especially since the more serious cases represent only 20-25% of the total. In addition, it must be remembered that one serious case of misconduct not dealt with fully, or at all, by a national system creates a media explosion that damages severely not only relations between society and biomedical research but the atmosphere within the scientific community.

	Independent national system

Not surprisingly, my suggested solution is the creation of an independent national system covering all health sciencesmedicine, dentistry, pharmacyand the drug industry and agencies. Such a national committee would be detached operationally, but not by membership, from the universities and other research institutions, thereby breaking away from the concept of total self government of institutions.

The committee should have a judicial chairman, for instance a High Court judge. Scientific misconduct most often takes place in the grey zone between legislation and unwritten guidelines for good scientific standards. The rules and guidelines of the system must be able to secure fairness for whistleblowers and accused scientists alike. Furthermore, experience suggests that the prestige associated with having a judge as chairman reduces the likelihood of subsequent court trials.

Members of the national committee should represent bodies such as the universities, scientific societies, research ethics committees, and government research institutes. The membership must be kept low, at eight to 10, and substitution of members should be possible. Membership could be considered a professional duty and consequently non-salaried, except for the chairman and vice chairman.

Whistleblowers should be able to contact the committee directly, not through the governing body of the university or the research institute. If the committee takes up the case, the institutions will obviously participate in the inquiry and the identity of the whistleblower will become known. I would not recommend a procedure whereby a complaint has to be made through the institutions at which the alleged misconduct has taken place. There are all too many examples of undue biased involvement by institutions. Furthermore, local resistance may well be strong in some academic circles, as was the case when research ethics committees were introduced.

	Procedures

A national committee should divide its procedural operations into two phasesthe inquiry and the investigation. Decisions on whether or not an investigation is required depend on the results of the inquiry. Ad hoc investigative committees can be internal to the independent national body or partly external, with an internal chairperson and an added number of independent experts accepted by both whistleblower and the accused person. The final report from the national committee should contain the committee's own conclusion, based on the premises of the ad hoc committee.

Punishment is best left to the institutions employing those found guilty of misconduct, but these should be obliged to report back to the national committee. However, the weight of reprisals should be determined centrally to avoid too heavy a punishment being meted out by an institution that wishes to demonstrate its commitment to purity.

	Creating a committee

In creating a national committee all interested partiesthe Medical Research Council, the Royal Colleges, the universities, the Department of Health, the professional associations, and the hospital authoritiesshould hold preliminary discussions. Even if one group should refuse to participate, the others must press on. A UK committee on scientific misconduct will need administrative offices and a secretariatperhaps in one of the royal colleges, the MRC, or the Department of Health. The budget could be covered by joint funding for a pilot period of, say, three years.

	Support for "authorship"

In addition to securing general prevention of fraud, a UK committee could create the necessary support among scientists for the endeavours of medical journals to restore authorship to its original position and validity. In this way such a committee could tackle the most prevalent "crime" in the dishonesty spectrum.