Attending
Sandy McCall Smith
Richard Smith
Derick Wade
Ann Sommerville
Liz Wager
Alison Tonks
Jeffrey Tobias
Tom Wilkie

Apologies: Peter Singer

We agreed that the committee was not equipped to deal with individual grievances or complaints against the BMJ. There are other avenues that would serve complainants better, and Richard Smith will write an editorial explaining what they are. They include: The press complaints commission and the journal committee of the BMA. The chairman of the joumal committee can approach the ethics committee for advice about grievances with an ethical component.

A general practitioner wrote to the BMJ reporting his treatment successes with a fish oil dietary supplement. He claimed that it was effective in a variety of psychiatric disorders if taken in high doses. RS wrote to him expressing his concern about the ethics of this sort of experimentation. He replied that he was not conducting experiments or research, just offering people a dietary supplement when he thought it might help them.

TW, JT, DW, and LW felt that his indiscriminate prescribing of Omega 3 fish oil may be misguided but was within the realms of acceptable clinical experimentation. Particularly since fish oil is a dietary supplement and not a prescription only drug. JT said he thought the author was naďve about the placebo effect, and possibly deluded about the effects of the fish oil.

RS and AT and AS disagreed. It’s not clear what he told the patients about this supplement. If he told them he had no evidence that it worked in their condition but they were welcome to try it, he has done nothing unethical. It’s unlikely that he put it in those terms, however. RS is keen to debate in the journal the grey area between acceptable clinical experimentation and unacceptable research. Mostly it boils down to the information given to affected patients. He will find a way to do this soon.

Decision (unanimous): Take no action against general practitioner. Write or commission something on this issue for the BMJ.

A systematic review done originally for the UK’s National Institute for Clinical Effectiveness and submitted to the BMJ in abridged form. A previously published Cochrane review on the same subject was not mentioned, and five trials included in the Cochrane review and the NICE review were omitted from the review sent for publication. Was the omission deliberate or just incompetent? Did the advisory group for the NICE review peer review this version before submission, as claimed by the authors?

AS: I suspect the omission was deliberate. They have not responded to direct questions about pre submission peer review by the advisory committee

TW: If they were deceitful, they were very bad at it. They must have known that one of the BMJs reviewers would spot that a Cochrane review had been left out

DW, LW, AT: The authors may have been working within tight constraints laid down by NICE. It’s likely that their hands were tied about what they included and when they submitted the manuscript.

JT, AT: The review may have been written a long time ago-before the Cochrane review was published. They may have been in too much of a hurry to update once they had been told to submit for publication.

SMcS and RS: We cannot make a judgement about their motives. The question is, should we ask a higher authority to look into it? SMcS: No, we haven’t reached that point yet.

DW: The authors should be made to consider the ethical implications of submitting a systematic review that is not a systematic review just because NICE has put them under pressure.

RS They must be clearer about their contract with NICE and the role of the advisory group.

Decision and action (unanimous). We agreed that the committee was not in a position to judge whether or not the authors had been deliberately dishonest. We have not yet reached the stage where we refer the case to some higher authority for investigation. BMJ editors have had one round of correspondence with the authors asking them to explain why the Cochrane review was omitted and to explain the precise role of the advisory committee. The authors have responded to one letter, but not to RS’s subsequent letter asking for more details. They should be contacted again and urged to respond.

A randomised trial showing that a nutritional supplement can cause a dramatic improvement in one aspect of the health of elderly people. The trial was a follow up of one published eight years ago in an international journal. We were left wondering why there was so much delay and why these results had not been reported in the previous paper. We also noted that there was only one author and the results were very dramatic if true. One of the BMJ’s statistical reviewers thought the results were fabricated. He wrote "This paper has all the hall marks of being entirely invented", noting a dramatic digit preference and significance tests that were unbelievable for the kind of health problem reported. RS wrote to the President of the author’s university in asking him to investigate.

SMcCS: There’s little more we can do here until the president of the author’s university responds

DW: Some of the test results he reports are simply not possible

RS: If we get nowhere with the president of the university the next stage might be to notify the international journal that published the previous paper. They might want to investigate that paper too.

SMcS: Another alternative would be to approach the national regulatory authorities. We have enough evidence here to suspect misconduct.

Decision and action (unanimous). Wait for a response from the university’s president, and talk to the editors of the other journal if none is forthcoming soon.

The BMJ published a paper followed some years later by two brief retractions - on the instructions of the university where the work took place. The university gave few details about what had gone wrong and who, if anyone, had been punished. One of the paper’s authors contacted RS because he felt the retractions made all authors look equally guilty, and in fact most of them were innocent of any wrongdoing. RS wrote to the dean of the medical school where the work was done, asking for more details of the case so something more satisfactory could be written in the journal.

He responded that the UK’s General Medical Council were in the middle of an enquiry, but gave no further information.

SMcS: We should let readers know that a hearing is going on. Perhaps we could publish the name of the author under suspicion. That would clear all the others. Alternatively we could report that "no action is being taken against the following authors" then list them. JT supported this view.

TW: We can’t do that without implying guilt which would be wrong. We need a detailed account from the university, which would clear the innocent authors.

AS: We need to find out when the GMC hearing will be completed, so we can publish the result.

AT: The retractions in the BMJ are inadequate, and I’m surprised the BMJ didn’t demand a fuller explanation at the time. We need a detailed editorial telling the whole story.

RS: In a sense all the authors are implicated in any misconduct. Being an author means taking shared responsibility for integrity of the results. Asking all authors to guarantee a paper may be unsustainable, however, which is why the BMJ is moving over from authors to contributors. The transition is not yet complete.

Decision and action. RS to ask the university for more details again. If they are forthcoming the BMJ will publish an account of the investigation and clear the innocent authors.

The journal will also revisit the debate comparing the notion of contributorship with authorship, and the status of guarantors.

• 002. A case of cheating at medical school. Prior disclosure of topics to be examined in a final Objective Structured Clinical Examination.

The committee received a letter and a report on the university’s response to the disclosure of a cheating episode in final exams. Academic staff at the university had suspected cheating and taken steps to minimise its impact including: Writing to all final year students before the exam, and modifying the OSCE in question. No further action required

• 004. A case report of a novel surgical procedure which was considered dangerous by a BMJ reviewer

The committee received a response from The British Association of Plastic Surgeon’s immediate past president. He acknowledged a problem with the procedure but could only offer the names of experts to ask for further advice. The committee felt there was little point in pursuing the matter further. RS suggested we use the procedure to illustrate a debate on the boundary between innovative practice and unacceptable experimentation.

• 001. A lesson of the week from a tertiary referral centre describing a patient’s poor outcome after delayed referral from another hospital. One BMJ reviewer suggested the patient had grounds to sue the original hospital for negligence.

The committee received a response from a professor at the tertiary referral centre whose juniors had written up and submitted the case. He explained that any implication of wrongdoing at the referring hospital was entirely unintended. The patient had received good care there, and eventually had a good outcome. The reviewers and the committee were misled by the submitted account which did not do justice to the complexities of the case.

The ethics committee agreed that asking for clarification from authors is an important first step whenever suspicions are aroused. The authors were able to clear this one up quickly, and convince the committee that the "bad practice" picked up by reviewers was a side effect of the reporting, not real.

Committee received the latest letter from RS to the spokesperson of this company. It urged the company to draft letters for publication so the debate about editorial bias could be conducted in the open. It also said that the journal received, and therefore published, far more high calibre articles arguing against this company’s products, than in support of it, which may have given rise to the spurious appearance of bias.

• 007. A phase one study in which subjects with a dangerous communicable disease were treated by innoculation with another dangerous communicable disease.

The committee received two referees reports about this paper. Both suggested that the BMJ should reject it on scientific not ethical grounds. The paper was rejected, and the case closed.

• 003. A paper describing a new theory of disease that made no scientific sense. The author reported wide ranging success treating himself and others with supplements based on his theory.

The committee received a letter from the author expressing his disappointment at being referred to the national regulatory authorities.

We discussed how far the committee should go to chase up the authorities, and agreed that we should at least find out whether or not the referral letter had been received.

Action: RS to ask the President of the national medical association for an acknowledgement.

Is "opt out" consent acceptable for a randomised trial, even if the intervention is non clinical? Can editors question the decisions of ethics committees?

This paper is due for publication in the BMJ. The authors sought and got approval from a research ethics committee in the UK to conduct the study.

TW: Opt out consent for a trial is not acceptable, although in this case the intervention is unlikely to have done anyone any harm. AT agreed.

AS: We are not in a position to judge the research ethics committee’s decision to let this trial go ahead. They had a long and detailed protocol to look at, we only have the abridged final version. It might, however, be worthwhile asking the chairman if the research ethics committee members realised that consent was "opt out".

LW: I’m not unduly concerned about this form of consent in this trial. Participants have not been harmed. Participants’ notes were scrutinised by a nurse, but she could have been the practice nurse with legitimate access.

Decision and action. Publication will be delayed until AT commissions two accompanying commentaries-one for and one against opt out consent for trials.

The committee agreed unanimously that the journal is compounding a wrongdoing by lifting cases and pictures from books or journals with less stringent consent policies than its own.

Action. SMcS to draft a more just policy on publishing patient details that are already in the public domain.

Decision. The BMJ asks authors to mention ethics committee approval in all research papers. It’s unclear whether or not they all do. It’s an important issue and the journal should investigate the proportion of papers with explicit reference to ethical approval at the protocol stage.

Committee received correspondence between an author and RS about a letter for publication. The letter refers to a paper in another international journal, so the BMJ decided not to publish it. The author believes his letter should be published and wants to complain to the ethics committee about it.

RS terminated their correspondence after a lengthy explanation of the journal’s policy on letters that do not respond to a recent paper in the BMJ. No further action required.

The above paper (published in the BMJ in February) was billed as an audit and the authors report that they did not seek approval from a research ethics committee. Several rapid responses criticised this aspect of the study as it involved performing liver biopsies in some participants.

The committee had a brief discussion and agreed that the best way to debate this issue was via the rapid response facility on the BMJ’s website. Further, broader discussions about ethics committee approval for audit were deferred to the next meeting.

Items for the next agenda include:

• Speeding up progress with declaring our own competing interests

• Ethics committee approval for audit

• Developing a programme for reviewing BMJ’s existing policies, such as consent.