An Indiana University student researching the molecular physiology of diabetes admitted to a string of misconduct -- falsifying data in two papers, a successful grant application to the National Institutes of Health, and her defended PhD thesis.

Image by Tomomarusan via Wikimedia

The Office of Research Integrity (ORI) announced the misconduct finding this week in the Federal Register.

Emily Horvath, who got her PhD from Indiana University in December of 2008, admitted to falsifying data during her time as a grad student at the school, where she was mentored by cellular physiologist Jeffrey Elmendorf. According to the Federal Register, Horvath said she falsified original data in order to reduce the magnitude of errors within groups, thereby increasing the statistical power of the findings.

Research misconduct cases don't get too much more open-and-shut than this: The entire investigation took about three months, from discovering her malfeasance to the ORI issuing the official misconduct announcement, according to an ORI spokesperson. Indiana University spokesman Larry MacIntyre declined to elaborate on the incident, citing protection of Horvath's privacy under the Family Educational Rights and Privacy Act (FERPA).

"We've decided that because this did involve a student disciplinary process, it's covered by FERPA, which sharply limits anything we can say," MacIntyre told The Scientist.

MacIntyre did, however, say that he thought that the NIH-funded research project upon which Horvath was working was still "underway." The $369,000 R01 grant, about molecular mechanisms of insulin resistance and ways to reverse it, still appears to be active on the NIH's website.

According to the ORI, Horvath admitted to falsifying 5 figures in that grant application to the NIH's National Institute of Diabetes and Digestive and Kidney Diseases, which was awarded in March 2009. She also admitted to falsifying 5 figures on a separate NIH grant application, on which Elmendorf was listed as the principal investigator, to the NIH's National Center for Complementary and Alternative Medicine. That application was withdrawn last May.

According to the ORI spokesperson, it is somewhat rare for misconduct cases to be resolved so quickly. The key to that quick resolution was a "three-way agreement" between Indiana University, Horvath, and the ORI, the spokesperson said. In this agreement, Horvath agreed to make a fulsome and complete admission of wrongdoing, with the university overseeing that process. For its part, the ORI agreed to accept Horvath's admissions without launching its own investigation, which can take years to complete. "Both parties have the objective of clarifying the literature as well as making sure respondents who have committed misconduct are prevented from spending (Public Health Service) funds for some period of time," the spokesperson said.

The two journals that published Horvath's fraudulent figures are in the process of considering retracting the papers that contained those figures. "It's under evaluation," said Maggie Haworth, managing editor of Molecular Endocrinology, the journal that in 2008 published a paper describing the effects of chromium on the plasma membrane's sensitivity to insulin with 4 fraudulent figures (2C, 5, 6D, and 11). Haworth declined to comment on how Horvath's manipulations slipped through the peer-review process at Molecular Endocrinology.

"We are processing it so that the papers will be retracted," added Scott Herman, managing editor at Endocrinology, the journal that published Horvath's other fraudulent paper, which detailed how cells become more insulin-resistant with increased glucose flux through a particular biosynthetic pathway. Horvath admitted to fudging only one figure (2C) in that paper. When asked if he thought the misconduct compromised the broader conclusions or implications of the research, Herman responded: "I don't know. I think it just sort of skewed some of the results." Both journals are published by the Endocrine Society.

The Scientist was unable to locate Horvath, and calls and emails to coauthors on the two journal articles were not returned. Phone and email messages sent to Elmendorf were not returned.

If she seeks to continue a career in science, Horvath faces stiff suspensions. She cannot serve on any advisory board or committee at the Public Health Service (PHS), which oversees NIH, for three years. During that time, she also must essentially identify herself and restate the findings of her misconduct case if she is to be a part of any grants coming from the PHS.

MacIntyre declined to comment on any whether further disciplinary actions, such as stripping her of her PhD, would be levied against Horvath.

Editor's Note (16th April): Jeffrey Blaustein, editor in chief of Endocrinology, responded to The Scientist's question about how peer reviewer's at that journal could have missed Horvath's data manipulations and OKed the publication of the paper. In an email sent today, Blaustein wrote: "Without being specific about the particular situation about which you have written, the short answer to your general question of how reviewers/editors miss data manipulation is that while, reviewers/editors often catch plagiarism, instances of fabrication/data manipulation are unfortunately much more difficult to detect. The reason that ethics training has become so important in science is that science relies on trust that others are engaging in ethical behavior. If an individual (trainee or PI) changes a value, for example, in a table, how could a reviewer or editor know? Replication is such an important component of the scientific method, because this is how we are all protected in the long term against spurious or fabricated data. Unfortunately, that sometimes takes time. In the end, we hope that science is self-correcting."


Related stories:

Student fudged immunology data
[21st August 2009]

Images faked by UCSF postdoc
[16th February 2009]

Postdoc censured for fudged images
[3rd December 2008]

Grad student falsified data
[5th August 2008]