The aftermath of fraud

The admission of widespread data fabrication by Eric Poehlman, a well known nutrition researcher, as part of a plea bargain last week has sparked a discussion of about how journals can handle such impropriety. Poehlman's work was cited widely; the top 10 most-cited of his more than 200 papers have been cited an average of 125 times each. One of those top 10, a 1995 Annals of Internal Medicinepaper that was retracted last year was cited more than 150 times. "The realization that this work included false statements will force me to re-evaluate how the findings from my research are interpreted," said Wayne W. Campbell of Purdue, in Lafayette, Ind., who has cited Poehlman and called the situation "very disturbing." "I have followed his research with great interest over the years and used it as part of the foundation for some of my research," Campbell told The Scientist via E-mail. "We are in the midst of conducting a study that includes evaluations of energy metabolism and body composition in pre- and post-menopausal women. We had planned to use his research findings to help us understand the significance of our work, but that is not prudent now." "Those who coauthored and collaborated with Dr. Poehlman are in an unfortunate situation," said Walter DeNino, who worked with Poehlman as a research assistant and blew the whistle on him. Now a post-baccalaureate student at Columbia University in New York and about to apply to medical school, DeNino told The Scientist that "undoubtedly all of Dr. Poehlman's work will come under question, some of it unnecessarily." "The Annals of Internal Medicine paper has been cited in support of hormone replacement therapy repeatedly, [but] to my knowledge, these results have never been duplicated in [their] entirety," DeNino said.Ronenn Roubenoff of Tufts University, Boston, said his team was never able to replicate Poehlman's earlier work on various hormonal determinants ofresting energy expenditure. "In our hands, body composition has always been the major determinant of caloric requirements at rest, and we could not find significant effects of estrogen, epinephrine, and the like, which Eric had been publishing," Roubenoff told The Scientist via E-mail. "However, we ascribed it to use of different body composition techniques; now it may have a more sinister cause."Roubenoff, who cited Poehlman's work, also cowrote the editorial that accompanied Poehlman's 1995 in the Annals of Internal Medicine on changes in body composition throughout menopause that was subsequently retracted. Roubenoff said he doesn't think a letter about his own paper is needed, as it's just one citation and doesn't change their own data or the conclusions they drew. "But the comparison we thought we had to the effect of menopause from [his] work is no longer valid, so the interplay of papers that makes a scientific discipline move forward has been weakened," Roubinoff said.But a researcher who worked with Poehlman said it is still too early to evaluate the long-term consequences of the admission of fraud on his own work. "I myself managed and verified the databases I used," André Tchernof, now at the Université de Laval, told The Scientist via E-mail. "I am convinced that many other publications by other former members of the group are in the same situation. I would be saddened if the scientific community had the reflex to put everybody in the same bag."Yvonne Linné, of the Obesity Unit at Karolinska University Hospital in Stockholm, cited Poehlman, but hadn't yet heard about the admission of fraud. Although her research didn't depend on that data, "certainly it is sad to have cited a colleague in good faith [and then] find that data have been fabricated," she told The Scientist via E-mail. "I guess that not everyone who has quoted Poehlman over years can be expected to submit some commenting letter to journals."The downside for medical journals is "really substantial," said John Hoey, the editor of the Canadian Medical Association Journal, who noted that Poehlman's 1995 Annals of Internal Medicine paper was cited three times this year, even after it was retracted.The Uniform Requirements for Manuscripts of the International Committee of Medical Journal Editors (ICMJE) reads, in part: "It is not ordinarily the task of editors to conduct a full investigation or to make a determination; that responsibility lies with the institution where the work was done or with the funding agency."Editors, Hoey told The Scientist, don't do an investigation "because we don't have the time or expertise to investigate [situations which can be fraught with] charges and counter charges."Hoey said that it's up to universities to carry out such investigations, as the University of Vermont did in this case. "Given the way science works at the moment, I don't see how anything else could have happened in our setting. Personally, in my opinion, science shouldn't work any differently," said Russell Tracy, the university's senior associate dean for research and academic affairs.ICMJE member Frank Davidoff, editor emeritus of the Annals of Internal Medicine, told The Scientist that it was "really troublesome" that his journal had published the paper that was subsequently retracted. Davidoff said that a separate editorial might help spread the word in such cases, noting that "sometimes a retraction can get a bit lost, tucked away in a journal." "But I've never really heard a whole lot of discussion beyond formal notification," Davidoff said. "A retraction, after all, ends up on PubMed, but I suppose you could argue that editors might want to undertake an examination of the papers of those who cited a fraudulent paper and perhaps notify them by way of a letter. But, again, depending on the paper, there might be hundreds or even thousands of citations." "I think there is always going to be a question as to why the editor and the reviewers didn't discover a fraud, but we work with an expectation of honesty and trust among researchers," Davidoff said. "They are not looking for fraud; they are looking for other kinds of scientific inaccuracy."

The aftermath of fraud

In wake of Poehlman affair, academics, journals assess damage control

Interested in reading more?

Become a Member of