Forskningsartikkel om Sudbø-saken

Da bomben om den kjente Sudbø-saken, som hadde utgangspunkt i kreftforskeren Jon Sudbøs juks, i artikler han hadde publisert i anerkjente medisinske tidsskrifter, sprang vinteren 2006, ble jeg kontaktet av en gammel kollega, forskningsredaktøren i det tyske aktualitetsmagasinet Focus, med spørsmål om å gjøre en grundig, analytisk reportasje om saken, med utgangspunkt i EBM (evidensbasert medisin), et prinsipp vi begge hadde arbeidet etter da vi arbeidet sammen.

Siden tyskferdighetene er begrenset til skoletysk, fikk jeg klarsignal til å levere saken på engelsk. Wikipedia henviser også til saken min, med utgangspunkt i en resirkulering i nå avvikleder – en nyhetstjeneste jeg drev i det aktuelle tidsrommet.

Artikkelen er ikke ment som forsvarsskrift for Jon Sudbø, men peker på hvor enkelt det er for medforfattere og forskere å unndra seg ansvar for falsifisering av viktige forskningsresultater.

Intervjuene ble gjennomført på flere språk, og oversatt til engelsk – med unntak for engelskspråklige kilder, naturligvis.

Sudbø case: Co-author responsibilities disregarded?

The medical research communities in Norway, Sweden and the USA were seriously rocked last month, as it became clear that renowned oncologist Jon Sudbø’s Lancet-published study “non-steroidal anti-inflammatory drugs and the risk of oral cancer” was a piece of fraudulent research. One week later Sudbø admitted fabrication of data in two more articles, in equally respected journals. No wonder people start asking whether medical research is trustworthy or not.

By Jarle Petterson

The question is justified. Both The Lancet, The New England Journal of Medicine and the Journal of Clinical Oncology publish studies assessed by a panel of dependable specialists, also known as peer-review. Short after the shocking news, The Lancet’s editor, Dr Richard Horton proclaimed that “The peer-review process is good at picking up poorly designed studies, but it is not designed to pick up fabricated research”, according to the BBC, adding “Just as in society you cannot always prevent crime, in science you cannot always prevent fabrication”. But the question remains; if you cannot trust the world’s principal journals in the field of medicine, whom can we trust? Are serious slip-ups of this magnitude excusable in leading and presumably reliable medical journals? Are there ways of avoiding recurrences?

Dr Sudbø worked with some of the world’s top cancer researchers. Several of them had been listed as co-authors of the articles Sudbø now admits fabricating, so the scandal affects their credibility as well, even though Sudbø himself recently exonerated them completely.

Several of the specialists who co-authored research with Dr Sudbø said they were stunned by the disclosures. “This is as big a shock for me as for everyone else,” colleague Dr Jon Mork told Norwegian newspaper VG. “I don’t think any of the other co-writers were aware of this”.

Dr Atle Klovning, MD, Associate Professor at the University of Bergen, Norway, and one of Europe’s leading authorities on EBM (Evidence Based Medicine) isn’t quite as happy about vindicating the co-authors altogether. “Like the chief author, the co-writers have to adhere to the uniform requirements for manuscripts submitted to biomedical journals,” he says. “These are outlined in the guidelines defined by the Vancouver Group, later evolved into the ICMJE (International Committee of Medical Journal Editors). The co-authors are responsible for individual evaluation of idea, protocol, the study’s conduct, interpretation of data and their results, and to carry out discussions, in the light of other published data in their particular field, if in existence, which is usually the case,” Dr Klovning explains, clarifying further by adding “Colleagues accepting co-authorships, will have to vouch for both applied methodology, data and results. In sum, the co-writers’ role should be far more important than that of the peer-reviewers.”

Dr Klovning warns against perceiving co-authorship as a means of remuneration, or bribery, even. “People tend to regard participation as important elements on a resume, but there are obligations to consider, which are often forgotten, I’m sorry to say. Critical appraisal of studies published in many of the respected journals, would divulge worryingly high numbers of incidents worthy of concern,” Klovning argues. “A journal’s reputation is of little importance. Sound criticism is all that matters, here as in every other aspect of life.”

According to Klovning, EBM may assist in bringing fraudulent research to light, e.g. by applying the check-list specified by ICMJE. “All the comparative results needed, are at hand in databases such as The Cochrane Library, PubMed and EMBASE,” he proclaims, emphasizing that the co-writers in this particular case probably didn’t live up to their responsibility according to the Vancouver rules.

In light of Dr Klovning’s information, we sought a comment from Dr Sudbø’s co-author, Professor Scott M. Lippman at the M.D. Anderson Cancer Center in Texas, USA, which proved to be difficult. We did however receive a formal statement from the centre.

“As you know, there are several established journal guidelines. Therefore, while not familiar with the Vancouver rule, we’ve reviewed the rule and agree with and adhere to guidelines as specified in the Vancouver rule,” M.D. Anderson’s press-spokesperson Nancy Jensen says in an email, elaborating:

“[…] In both instances, M. D. Anderson’s role was analysis and interpretation of survey or clinical data provided by the lead author in Norway. Neither study involved a clinical trial, but drew conclusions based upon the supposed records of persons with histories of smoking (The Lancet) or oral leukoplakia (NEJM).” Some time later Jensen added:

“Although we cannot speak for every co-author on the Lancet paper in question, in the case of the U.S. co-authors, we met every criterion specified by the Vancouver rule. The U.S. co-authors contributed substantively to (1) the conception and design of the study and data analysis and interpretation, (2) drafting the article and revising it critically for important intellectual content, and 3) final approval of the version that was published. No U.S. co-author was responsible for acquiring data or supervising data acquisition, which was done in Norway and was the responsibility of Jon Sudbo. As in many scientific collaborations, we worked as a team and divided the work load to accomplish the project. Unfortunately, the U.S. co-authors learned about the fabricated data after the article was published in Lancet. After reviewing the Vancouver rule, it is a good rule, but like every guideline or rule, it does not prevent or eliminate intentional misconduct by certain individuals.”

The experts we have consulted seem to agree that Doctor Sudbø’s co-authors were severely exploited, but consensus also appears to be that closer observation of the Vancouver (and other) rules would hamper future attempts at publishing fabricated material.

In order to fully understand the severity of Sudbø’s scam, we need to take in the following:

He invented more than 900 individuals as the basis for his research on the correlation between taking anti-inflammatory drugs, such as paracetamol, and oral cancer. 250 of his sample of 908 people had the same birthday. The Lancet article, published last October, concluded that individuals who took anti-inflammatory drugs were less likely to develop the disease.

“He faked everything: names, diagnosis, gender, weight, age, drug use,” Stein Vaaler, director of strategy at Oslo’s Radium hospital, said. “There are no real data whatsoever, just figures he made up himself. Every patient in this paper is a fake”.

Professor Anders Ekbom, head of the Clinical Epidemiology Unit at the Karolinska University Hospital in Stockholm has been called to supervise the commission appointed to investigate Jon Sudbø’s work.

“It’s early days yet. We’ve only just started gathering data from the studies in question, leading up to a conclusion on 1 April,” Ekbom explains in a comment to Focus Magazin. “The material we’re assigned to examine, is of such a scale that we’ll need every day until then, exploring the articles themselves, the co-authors’ role and any institutional links involved. Last, but not least, we need to ascertain whether or not Dr Sudbø’s papers have caused direct or subsequent damage in clinical medicine,” he says.

One of professor Ekbom’s main concerns – and mandates – is looking into ways of preventing fraudulent research from future publication, which has turned out to be something of a perpetual problem in science at large.

According to Norwegian daily Aftenposten, Sudbø’s solicitor, Mr Erling O. Lyngtveit stressed on behalf of his client, who’s since been admitted to hospital himself and is officially on sick leave, that none of the fraudulent articles have had any consequences for patients. He said other data and studies presented by Dr Sudbø over the years are genuine. The solicitor called the fabrications “a catastrophe” for Sudbø, “both personally, professionally and socially. He’s taking one day at a time. The most important thing for him now is to prevent his work as a doctor from coming into doubt.”

Opposite Focus Magazin Mr Lyngtveit upholds Dr Sudbø’s previous statement, declaring as follows: “Apart from what we communicated to the press last week, we have nothing to add until professor Ekbom’s commission and the Norwegian Board of Health have drawn their conclusions. I will however confirm that Dr Sudbø claims his 2001 doctorate paper 100% valid and reliable.”

Although Dr Sudbø has pledged that approval, rather than financial motivations, inspired the misconduct all along, scientific dishonesty remains. So does the fact that two years ago, the American National Cancer Institute contributed USD 10 million, in order for him to conduct a study aimed at the treatment of 350-400 patients, with the two drugs celocoxib and EKB569, alone or in combination, rendering Sudbø liable for economic, as well as scientific fraud. A crucial uncertainty is, however; should he carry that burden alone?