Letters

Beyond conflict of interest

What is truth as it relates to albumin?

BMJ's editors should publish their own conflicts of interest regularly

Perhaps authors should list "interests"

Competing interests still exist

Sponsored drug trials show more-favourable outcomes

Biomedical journals need a concerted response against influence of tobacco industry

Why do people affiliated with and paid by a particular industry hold a particular view?

Non-financial conflicts of interest are more serious than financial conflicts

What is truth as it relates to albumin?

EDITORThe editor of the BMJ is so busy pointing the finger at others' conflicts of interest¹ that he ignores one lurking at his own front door. I refer to the continuing controversy over albumin. Editors have to sell their journal, although giving a copy free to every member of the BMA does create a captive market that other editors do not enjoy. The conflict here is the editor's need to feature in the media competing with readers' need to be told the truth. What is truth and what is tendentious opinion masquerading as gospel? Offringa tried to clarify his original message² when he said that by writing "halted" he meant that the use of albumin should not be halted.³

Is he to be criticised for over-egging the pudding, or is the editor to be criticised more for producing a mouse and calling it a mountain?

Oliver Dearlove, Consultant anaesthetist.  .
Department of Anaesthesia, Royal Manchester Children's Hospital, Manchester M27 1HA o.dearlove{at}man.ac.uk

1. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292[Free Full Text]. (1 August.)
2. Offringa M. Excess mortality after human albumin administration in critically ill patients. BMJ 1998; 317: 223-224[Free Full Text]. (25 July.)
3. Offringa M. Consider validity, clinical relevance, and applicability of albumin for critically ill patients. BMJ 1998; 317: 343[Free Full Text]. (1 August.)

BMJ's editors should publish their own conflicts of interest regularly

EDITORThe BMJ has now joined several prominent medical journals that have clarified or revised their conflict of interest policies.¹ I hope that it fares better than its counterparts in evenhandedly enforcing its revised policy.

In one well publicised case, a single day's consulting to a panel of the Food and Drugs Administration was considered such a profound violation that the New England Journal of Medicine changed its rules about conflict of interest.² In the case of an editorial in that jounal which flatly dismissed the carcinogenic potential of environmental oestrogens, however, long term funding from the chemical industry to the author of the editorial went undisclosed, and when disclosed was addressed by the editor's comment that "the line has to be drawn somewhere."³ The executive editor of the New England Journal of Medicine has also written an editorial in the Wall Street Journal suggesting that receiving grants from the National Institutes of Health can be a serious conflict of interest.⁴ Such lack of consistency is a continuing source of frustration among scientists.

None of the recent clarifications and revisions has adequately addressed a new issue concerning conflict of interest: the practice of paying scientists to write letters to the editor or editorials that present medicolegal positions of industry. Documents uncovered during tobacco litigation show that this practice has become widespread. Similarly, a recent front page article in the New York Times reported a $5m (£3.1m) plan to recruit previously "neutral" scientists to create doubt about global warming.⁵ Such initiatives corrupt the scientific literature and require aggressive measures by scientists and editors. A neutral scientist might escape editors' attention. Thus the BMJ's plan to send authors of letters "a questionnaire to complete only if we suspect that authors might have competing interests" appears inadequate.

An important omission from the BMJ's revised policy involves editors themselves. The executive editor of New England Journal of Medicine often appears at legal or media events sponsored by organisations representing the legal interests of the medical device industry. The editor of a journal devoted to women's health serves on the board of directors of an international medical device manufacturer. When these editors write editorials, review papers, or speak to the press, do they represent the interests of their journals or their "competing interests"? An easy solution is available: editors of medical journals should publish their own conflicts of interest regularly (including speaking, consulting, and medicolegal activities). Will the BMJ lead on this issue, serving as an example to both its authors and other editors?

Robert F Garry, Professor. 
Department of Microbiology and Immunology, Tulane University School of Medicine, 1430 Tulane Avenue, New Orleans, LA 70112, USA rgarry{at}tmcpop.tmc.tulane.edu

1. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292. (1 August.)
2. Angell M, Kassirer JP. Editorials and conflicts of interest. N Engl J Med 1996; 335: 1055-1056[Free Full Text].
3. Tye L. Journal fuels conflict of interest debate. Boston Globe 6 Jan 1998:A1.
4. Angell M. Tuskegee revisited. Wall Street J 28 October 1997:A18.
5. Cushman JH Jr. Industrial group plans to fight climate treaty. New York Times 26 April 1998:A1.

Perhaps authors should list "interests"

EDITORI welcome the discussion on conflict of interest and wholeheatedly agree that disclosure and transparency are the keys.¹ Although the change from listing conflicts of interest to listing competing interests is in the right direction, it still implies that this might be something avoidable or harmful. The British Journal of Psychiatry simply lists authors' "interests" and lets readers judge whether they are conflicting, competing, or whatever. I wonder if the BMJ might consider doing the same in the hope that this will encourage disclosure and make readers realise that few people are truly disinterested.

Liz Wager, Medical writer. 
Janssen-Cilag, PO Box 79, High Wycombe, Buckinghamshire HP14 4HJ lwager{at}jacgb.jnj.com

1. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292. (1 August.)

Competing interests still exist

EDITORWithin a week of publication of the Cochrane Injuries Group Albumin Reviewers' systematic review of giving human albumin to critically ill patients¹ I was invited to attend a meeting of an international expert panel advising Bayer on albumin. The meeting was in California. My expenses would be paid and I would receive an honorarium of $1500 (£940). Bayer manufactures most of the albumin used in Canada. I read Smith's editorial and declined.²

Ian Roberts, Director. 
Child Health Monitoring Unit, Institute of Child Health, University College London, London WC1N 1EH Ian.Roberts{at}ich.ucl.ac.uk

1. Cochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients: systematic review of randomised controlled trials. BMJ 1998; 317: 235-240[Abstract/Free Full Text]. (25 July.)
2. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292. (1 August.)

Sponsored drug trials show more-favourable outcomes

EDITORSmith expresses concern about the influence of conflict of interest on publications.¹ A recent finding in a Cochrane review supports the concern that drug company involvement in clinical trials affects outcome.²

The searches undertaken for the systematic review of clozapine versus "typical" antipsychotic drugs for patients with schizophrenia identified 29 randomised studies in 2490 participants. Sixteen of the trials reported some kind of connection with the manufacturer of the compound. These studies were assumed to be sponsored studies. Peto odds ratios and 95% confidence intervals were calculated for the primary outcomes of relapse, clinical improvement, and leaving the study early.

The odds of relapsing were significantly in favour of clozapine in the sponsored trials (odds ratio 0.5 (95% confidence interval 0.3 to 0.7); 13 trials, 980 patients). Non-sponsored studies reported equivocal findings (odds ratio 0.4 (0.1 to 1.4); 10 trials, 783 patients). Both sponsored and non-sponsored studies suggested that clozapine mediates a clinically important improvement in patients with schizophrenia when compared with the older drugs, but again sponsored trials were more positive than trials not clearly supported by industry (random effects odds ratio 0.4 (0.2 to 0.7) in nine trials in 1126 patients and 0.3 (0.1 to 0.7) in eight trials in 743 patients, respectively). Sponsored studies also reported that significantly fewer patients taking clozapine left the study early when compared with patients taking drugs such as chlorpromazine and haloperidol (odds ratio 0.5 (0.4 to 0.7); 14 trials in 1245 people). Non-sponsored trials (12 trials in 950 patients) did not show this degree of effect (odds ratio 0.6 (0.3 to 1.2)).

The observation that drug industry sponsorship is associated with more-favourable outcomes is of concern. This finding emphasises how important transparency regarding drug company sponsorship is in the assessment of trial outcomes. Those undertaking meta-analyses of drug treatment should investigate for sponsorship bias by using sensitivity analysis.

It is of concern that licensing authorities make decisions mainly on the basis of trials performed by industry and therefore do not have information from independent researchers. Of course, the problem of conflicts of interest may be even greater outside industry, where trialists may have considerable investment in their own particular brand of community care package or psychological intervention.

Kristian Wahlbeck, Editor. 
Cochrane Schizophrenia Group, Department of Psychiatry, University of Helsinki, PB 320, FIN-00029 HUCH, Finland Kristian.Wahlbeck{at}huch.fi

Clive Adams, Coordinator. 
Cochrane Schizophrenia Group, Oxford OX2 7LG

1. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292. (1 August.)
2. Wahlbeck K, Cheine M, Essali MA. Clozapine vs "typical" neuroleptic medication for schizophrenia. In: Cochrane Collaboration. Cochrane Library. Issue 4. Oxford: Update Software, 1998.

Biomedical journals need a concerted response against influence of tobacco industry

EDITORWe welcome Smith's discussion of conflicts of interest in scientific writing and research.¹ The ethical challenge represented by the covert influence of the tobacco industry on academic research and publication needs a concerted response from all biomedical journals. The Journal of Epidemiology and Community Health will be joining this debate by publishing a fuller exposition of our arguments in an editorial.²

Tobacco industry papers published on the internet by Action on Smoking and Health expose how the tobacco industry has tried to influence research into tobacco and health (www.ash.org.uk/). For example, Philip Morris fostered controversy about the effects of passive smoking on health and countered authoritative review articles by establishing an international network of paid scientific consultants. Their activities included writing critical letters, publishing review articles, establishing a learned society on indoor air quality, and researching into other causes of lung cancer (such as keeping pet birds). The success of the strategy is shown by Barnes and Bero's study reported in Smith's editorial.¹ Internal industry documents also describe how the US Tobacco Institute paid $2500 to $10 000 (£1560 to £6250) to authors of letters criticising the Environmental Protection Agency's report of 1993, which declared environmental tobacco smoke to be carcinogenic.³ Some letters were revised by law firms before publication.

Letters to scientific journals should be rigorous and unbiased and not advocacy for sectional interests.⁴ Editors have a special responsibility to guard against bias in correspondence columns, especially when letters are not peer reviewed. The International Epidemiology Group recently circulated a code of practice for epidemiologists (www.dundee.ac.uk/iea/). This states that researchers should declare all actual, apparent, or potential conflicts of interest to ethical review committees and acknowledge publicly all sponsorship of research, as well as stating that epidemiologists should judge their work and that of colleagues impartially. This code is in accord with Smith's recommendations.¹

Medical editors and eminent medical scientists should agree a code of practice for medical and epidemiology journals. Actions could include keeping a register of scientists with interests related to tobacco; asking authors of papers (including letters) related to tobacco to declare conflicts of interest; checking authors' credentials using the register of interests; and vigorously naming and shaming those exposed as covertly funded by the tobacco industry.

Proctor of British American Tobacco argues that we should trust in the integrity of scientific researchers and the peer review process as guarding against results from research funded by the tobacco industry being presented in a skewed manner or so as to confuse the public health message.⁵ The internal industry documents show that the right to such trust has long since been abrogated. The credibility of science as a whole, and medical and epidemiological research in particular, is too important to allow the tobacco industry's efforts to go unchallenged.

Richard Edwards, Lecturer in public health medicine. ^*
P.R.Edwards{at}ncl.ac.uk

Raj Bhopal, Professor of epidemiology and public health. ^*
Department of Epidemiology and Public Health, School of Health Sciences, The Medical School, University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH

1. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292. (1 August.)
2. Edwards R, Bhopal R. The covert influence of the tobacco industry on research and publication: a call to arms. J Epidemiol Community Health (in press).
3. Wadman M. Dilemmas for journals over tobacco cash. Nature 1998; 394: 609.
4. Bhopal RS, Tonks A. The role of letters in reviewing research. BMJ 1994; 308: 1582-1583[Free Full Text].
5. Proctor CJ. Tobacco research: collaboration, not confrontation, is the best approach. In: Should industry sponsor research? BMJ 1998; 317: 333-339[Free Full Text].

^* Dr Edwards is (unpaid) chair of Northern Action on Smoking and Health. Professor Bhopal is a non-executive director of the Health Education Authority.

Why do people affiliated with and paid by a particular industry hold a particular view?

EDITORThe issue of conflict of interest, which Smith discusses in an editorial,¹ brings up a chicken and egg situation. Do people hold a particular view because they are affiliated with and paid by a particular industryfor example, the tobacco industryor are they affiliated with and paid by that industry because they hold a particular point of view?

Daniel J Highkin, Internist. 
Vancouver Clinic, Vancouver, WA, USA klutehighkin{at}earthlink.net

1. Smith R. Beyond conflict of interest. BMJ 1998; 317: 291-292. (1 August.)

Non-financial conflicts of interest are more serious than financial conflicts

EDITORI think that conflicts of interest should not be limited to financial considerations. Non-financial conflicts are much more serious barriers to the fair conduct, reporting, and criticism of research studies. I have encountered four types of such conflicts.

The first is fanaticism about a single issue. Certain causes attract people who become so committed that they cannot ever evaluate research fairly. Smoking and salt are two of these. I know of at least two senior academics who would never put their name to a research report describing any beneficial effect from smoking. Just as researchers, entirely appropriately, should declare any funding from, say, tobacco companies or food or drink companies, so researchers should also declare whether they are members of antismoking or antialcohol pressure groups or any other relevant organisations.

The second is political commitment. I know of one Marxist academic who could never admit that any action of any pharmaceutical company was beneficial. Antagonism to capitalism or to the pharmaceutical industry should be declared, as should share holdings and funding from industry.

The third is philosophical bias. This often governs attitudes to research. For example, some medical scientists cannot admit that nutrition is beneficial to health, while others believe that nutrition can do everything. Similar biases are associated with most controversial or new treatments.

The last is commitment to a particular theoretical framework for solving a problem. This is by far the commonest and in my view the most important source of bias and conflict of interest. How pleasing it would be to see at the end of a paper something like: "I am delighted by these results since they justify the 25 years I have spent following this line of research" or "I am thrilled about the negative outcome of this study because this seriously damages the case proposed by my opponent, Professor X, with whom I have been fighting."

Perhaps science might seem more human and more believable if we all agreed that conflicts of interest are everywhere. The only defence against them is complete openness, not just about money but also about political beliefs, fanaticism, and research commitment.

David F Horrobin, Chairman and medical director. 
Laxdale, King's Park House, Laurelhill Business Park, Stirling FK7 9JQ