Subtle hand of interest

The New England Journal of Medicine, probably the world's leading medical journal, will not publish articles written by authors who have a financial conflict of interest. As a consequence, it finds it almost impossible to publish editorial on some subjects. Dermatology is one example. Every possible author either does research funded by a pharmaceutical company or is a consultant to a company, or owns shares in a company.

The journal introduced its policy because of abundant evidence that financial conflicts of interest influenced how doctors behave and think. Most of this occurs subconsciously. Doctors will insist that their professional judgement is not distorted, just

as most of us are convinced that we are not influenced by advertising. But we are.

In the US doctors now begin lectures by describing their links with pharmaceutical companies. In Britain we have lagged behind, but both the British Medical Journal and the Lancet have now introduced policies on conflict of interest. Recent articles in Independent focused on the activities of the Portman Group and the funding of the Alcohol Research Group in Edinburgh by the drink trade, but this is not a unique case.

At the BMJ we include the source of funding for all research papers and ask authors to sign a form telling us of any conflict of interest. Reviewers of our papers are also asked to sign the form. We are interested not only in financial conflict of interest but also in any personal, political, academic or religious conflict that may influence a person's judgement. We do not aim to eradicate conflict of interest. That, we think, is impossible. Rather we aim to disclose important conflicts. Reade r s will then make their judgements on the author's arguments.

We are not saying that an author's arguments are rendered worthless by a conflict of interest: if we were, there would be no point in publishing the piece. What we are saying is that conflict of interest works in a subtle way. Researchers know that bias is all-pervasive. That is why they insist on "double blind randomised controlled trials" - where the treatment and a placebo are allocated randomly and where neither the patient nor the doctor knows who has received the treatment. Doctors who evaluate treatments that they also developed are likely to find the treatment effective irrespective of whether it is or not, if they know which patients have received the treatment. They are not being dishonest: it is the effect of bias.

We were led to our policy by the American example, by the accumulating evidence on how conflict of interest influences medical judgements, and from our own experience. In the year before our policy began, for example, we were sent a review of treatments for a particular condition. The reviewer could not at first understand why so much emphasis was given to one particular treatment. Then he realised that the authors were funded by a pharmaceutical company. In another case a BMA report on alcohol and the young, written with support from the drinks trade, analysed the evidence in a partial way (the report was improved before publication). The association later turned down money from the Meat Livestock Commission to produce a report on diet and health, funding the report itself.

Conflicts of interest commonly arise with letters to the editor. One eminent gastroenterologist told me his first thought on reading a letter in a medical journal or newspaper is "What are the author's financial connections?"

We have now collected a formidable file of conflicts of interest and published many statements of conflict. One of our most interesting examples concerned Sydney Brandon, emeritus professor of psychiatry in Leicester. He wrote us an angry letter about a report on people committing crimes under the influence of Halcion: he thought the whole thing exaggerated. We sent him a conflict of interest form, and he then realised how his extensive contacts with Upjohn, the manufacturer of Halcion, might h a ve influenced his view. "It is evident," he wrote in a letter published in the BMJ, "that I do have a conflict of interest but wrote my letter with no consciousness of the fact."

My own conflicts of interest are that I was a member of the committee of the Royal College of Physicians, which produced the safe alcohol limits that have been criticised by the Edinburgh group. But I am also a member of the Wine Society and the Scotch Malt Whisky Society, have known the two researchers from the Edinburgh group named by Independent for many years, and have spoken with them at the Edinburgh Science Festival in a session funded by the Portman Group: my expenses were paid, but I received no fee.