Lancet study faked

Investigation to probe all research conducted by scientist accused of fabricating results from 900 research participants

Jan 16, 2006
Stephen Pincock
A leading cancer hospital in Norway has begun appointing members of an independent review commission to investigate all the research conducted by one of its top scientists, Jon Sudbo, after claims emerged that he had completely fabricated a paper published last year in The Lancet associating long-term use of non-steroidal anti-inflammatory drugs (NSAIDs) with a lower risk of oral cancer.The Norwegian Radium Hospital in Oslo has appointed Anders Ekbom from the Karolinska Institute to lead the investigation, which will probe everything published by Sudbo, who was the lead author of the paper in question. In recent days, management at the hospital's Cancer Clinic has secured all data from his previous research for the purposes of the investigation."The commission "is supposed to start its investigation on Wednesday," hospital spokeswoman Trine Lind told The Scientist. "For safety's sake we have to look into everything."Last week, the hospital began investigating rumors that Sudbo had invented more than 900 individuals who served as the basis for a Lancet paper published on October 15, 2005. The study was an analysis of data from a population-based database, and found that long-term use of NSAIDs was associated with a lower risk of oral cancer (even in active smokers), but a higher risk of death from cardiovascular disease.After looking into the fraud claims, the hospital's management came to the conclusion that the data were fabricated. "We have reason to believe that it is true," Lind said, adding that Sudbo has admitted as much to his superiors at the hospital. British newspaper The Guardian quoted Stein Valler, director of strategy at the hospital, as saying "he faked everything: names, diagnosis, gender, weight, age, drug use."On Friday, the hospital contacted The Lancet, which is now conducting its own investigation, including contacting the 13 co-authors on the paper, including several researchers from the United States.Richard Horton, editor of The Lancet, said the issue was complicated by the fact that the hospital only had a verbal admission of fraud. "They don't have anything in writing, and neither do we," he told The Scientist.As a result, the journal would be contacting co-authors on the paper tomorrow and letting them know that, pending the investigation, The Lancet would issue an expression of concern. "Basically it means we have very good reason to think the work is unsafe, but not a cast-iron admission," he said. Once the investigation had been completed, a retraction would be issued."What I've been told is that he sat in front of his computer and made the whole dataset up and convinced his co-authors it was genuine," Horton said. "It's completely inexplicable."Over the weekend, the management at the hospital also reported the fraud to the Norwegian Board of Health. "The consequences of possible irregularities and fraud in earlier scientific works are so formidable that we need assistance from the Norwegian Board of Health," the hospital said in a statement. "We also have to collect all necessary data so as to enable the external investigation committee in carrying out its work."According to the paper, Sudbo and three co-authors contributed equally to the paper, which was partly funded by grants from the Norwegian Cancer Society and the U.S. National Cancer Institute. One co-author has consulted for Pfizer, maker of celecoxib (Celebrex); no other conflicts of interest were declared.Links for this articleRadiumhospitaletAnders EkbomJ. Sudbo, et al, "Non-steroidal anti-inflammatory drugs and the risk of oral cancer: a nested case-control study," The Lancet, October 15, 2005. PubMed: 16226613J. Sudbo