Research Misconduct

On June 28, Eric T. Poehlman, a research professor at the University of Vermont, was sentenced to serve 366 days in jail for fabricating and falsifying a decade's worth of data on aging, menopause, and obesity. Poehlman's jail term marks the first time an academic scientist has been incarcerated for research misconduct.

Poehlman has admitted to using fraudulent data in 10 scientific papers and in 17 grant applications, which resulted in nearly $3 million in federal grant awards. In addition to his jail time, Poehlman is barred from applying for or receiving federal funding for life, has to correct the literature, and has to repay the government nearly $200,000.

Although this case is a mesmerizing story, few of the research misconduct cases opened each year at federal agencies involve criminal wrongdoing. Most misconduct cases involve only noncriminal transgressions that are handled by the agency or the alleged offender's institution.

Federal agencies deal with research misconduct allegations according to a government-wide policy developed by the White House Office of Science & Technology Policy. The policy sets the basic guidelines that federal agencies should follow but allows them to customize their own policies.

Central to the federal policy is the standardized definition for research misconduct, which is the "fabrication, falsification, or plagiarism in processing, performing, or reviewing research, or in reporting results." The policy also requires that, for the action in question to be found to be misconduct, it must be a significant departure from normal activities of the local research community, be done intentionally, and be supported by sufficient evidence.

Federal agencies are given ultimate oversight authority under this policy, but the research institutions that receive the grants are named as the primary bodies to prevent and investigate research misconduct and to take appropriate disciplinary action. In addition, the guidelines state that an allegation of misconduct should trigger an inquiry to ensure that there is substance to the charge, an investigation to collect evidence (if substance is found), and an adjudication or settlement phase where a decision of guilt or innocence is made and any necessary disciplinary action is set.

From this framework, federal agencies have developed their own policies to handle research misconduct cases against their external grantees. The differences in agency processes and procedures vary with which office deals with misconduct cases and with who carries out the inquiry and investigation phases.

One place research misconduct cases can be delegated to is an agency's Office of the Inspector General (OIG), which is an independent oversight body within the agency. That's the arrangement at the National Science Foundation, which focuses about half of its investigative staff on noncriminal cases involving research misconduct. These cases are referred to as administrative, because if an allegation is substantiated, it typically results in administrative actions by the agency or institution rather than civil or criminal prosecution by the Department of Justice.

"Unlike for criminal investigations, the research misconduct component of our office is not inherent in OIG's role, it's delegated to us from the agency," explains James T. Kroll, head of administrative investigations in NSF's OIG. In serving this role, OIG typically has about 50 to 60 active research misconduct cases, with the bulk of them involving alleged plagiarism.

The process used by NSF directs all allegations of research misconduct directly to OIG. This segregation from other parts of NSF keeps the allegations from having any influence on research funding decisions.

"You don't want division directors, program officers, or peer review panelists to make funding decisions based on allegations that may or may not turn out to be true," explains Kenneth L. Busch, an investigative scientist at NSF's OIG. If the allegations lead to misconduct findings, then appropriate action can be taken, he adds.

With an allegation in hand, the office opens an inquiry to determine if there is any substance to the alleged misconduct. The first step is to go to the person alleged to have done something wrong and get his or her side of the story.

"Whenever possible, we go to the subject first, because they can sometimes clear up the allegation," says Scott J. Moore, an investigative scientist at NSF's OIG. The exception to this policy is in misconduct cases where OIG believes that contacting the subject could seriously compromise the case.

If the inquiry phase, which regulations limit to 90 days, turns up enough evidence to support the allegation, an investigation is opened. OIG typically defers its investigation while it refers the matter to the institution where the alleged misconduct occurred. OIG considers the institution to be in the best position to evaluate the alleged wrongdoing relative to the local research community's standards, something that is required for a finding of research misconduct.

Institutions can also receive allegations directly and do their own inquiry without telling the funding agency. If the process moves to the investigation phase, however, the institution is required by NSF to inform the agency. The institution would then continue its investigation.

"Once a case is referred to a university, they are responsible for the investigation," Busch says. He adds that his office does not call the university for weekly updates but instead waits until the university makes its findings, assesses any necessary penalties, and submits its report to NSF. Regulators give the institution 180 days to complete the process, although the time frame is flexible.

As for the actions taken by the university, "we try to make it clear when we refer a case that any action taken needs to protect the interests of the university," Kroll says. "We tell institutions not to try to do something that they think will impress NSF."

After an institution has completed its investigation and taken the necessary action, it submits a detailed report to NSF's OIG. That office reviews the report, conducts any additional needed investigations, and writes its own report with its own recommendations, which it sends to NSF's deputy director, who makes the agency's adjudication. This step can take up to six months.

If a research misconduct finding is warranted, there are several levels of action that OIG can recommend and that NSF can accept to protect the government's interest. These actions are administrative in nature and are similar to those used by other agencies.

The first level involves a letter of reprimand to the guilty party. Beyond a letter, the individual could be required to certify that anytime he or she submits something to NSF, it is accurate and contains no inappropriate material. The individual also may be asked for assurances; that is, the individual would have to have his or her work reviewed and cleared by another person at the pertinent institution prior to submitting it.

In the case of both certification and assurances, the actions are between the individual and OIG. Documentation and forms go directly to OIG, and NSF's program officers and peer reviewers don't see them.

The most serious sanction bars the individual from receiving funding or performing research on any federally funded program for a period of time. This ban holds across all federal agencies and is used only in the most serious cases.

The Department of Defense also uses its OIG to handle research misconduct cases. These cases are directed within DOD's OIG to the Defense Criminal Investigation Service (DCIS). Although this office focuses primarily on criminal investigations throughout DOD, it deals with six or so cases involving grant fraud and misconduct out of its total of about 1,600 cases per year.

"There are three prongs to our process at DCIS: criminal, civil, and administrative," says Richard C. Beltz, director of investigative operations at DCIS. "If there is no criminal prosecution, the department has potential civil remedies. If there are no civil remedies available, there are administrative remedies that are available through department offices such as the Office of Naval Research," he explains.

Beltz acknowledges that his office does not see that many misconduct cases, but when allegations in this area are made, they are taken seriously. "Suspension and disbarment are very often used as a remedy against folks who commit fraud against DOD," he says.

"The last thing universities and companies that rely on federal research grants want is to have their staff suspended or disbarred," Beltz adds. He says that these groups work quickly to fix misconduct problems when they are identified.

The situation for handling research misconduct cases at the Department of Energy is slightly different. Research misconduct cases involving criminal activity would be handled by its OIG, although there haven't been any such cases in the past 10 years, according to an OIG spokeswoman.

Instead of setting up a centralized office to deal with research misconduct cases, the DOE policy charges the individual institutions with handling allegations and completing any necessary inquiries, investigations, and adjudications. "The department decided that is was just not a good use of the taxpayers' dollars to have a centralized office to deal with the one complaint or so that comes in per year," explains William J. Valdez, director of the Division of Planning & Analysis in DOE's Office of Science.

Valdez notes that the institutions are not required to tell DOE about these cases unless there is a problem where agency help is necessary or one of the parties in the case believes that justice was not served.

If help is needed, the requests would come to Valdez's office, but that doesn't happen often. "I can't think of an instance where we've had to intervene in a local investigation in the past two or three years," he says.

Valdez points out, however, that while DOE has not had much of a problem with research misconduct, that doesn't mean it's not happening or that it shouldn't be taken seriously. "There is an assumption that it is happening," he explains, but universities are able to handle the situations on their own.

One agency that does not use its OIG for its research misconduct process is the National Institutes of Health. All of its misconduct cases are sent to the Office of Research Integrity (ORI) in NIH's parent agency, the Department of Health & Human Services (HHS).

Although ORI also handles cases from other HHS entities such as the Centers for Disease Control & Prevention, it mostly deals with cases stemming from NIH grants and involving research fabrication and falsification. ORI has about 50 to 70 cases open at any given time.

But unlike some other agencies, ORI does not handle the inquiry or investigation phase in-house. Instead, it relies on the institution to work through the allegations on its own and requires that the institution inform ORI only when an investigation is opened. The institution must also submit a report once the investigation and adjudication phases are complete.

"The institution gets the first shot to basically define what the problem is and to make a determination about it," explains Chris B. Pascal, director of ORI. "ORI does a detailed oversight process that is basically a double check of everything the institution has done," he says.

"We don't always agree with the findings of an institution," Pascal notes. He explains that this may be because there is not enough evidence or the analysis done by the institution is flawed. In such cases, ORI can decide to close the case, or it may go back to the university and ask that the case be reopened.

ORI may also recommend a finding of research misconduct that wasn't made by the university to the Office of Public Health Services—the office charged with adjudicating research misconduct cases for HHS. For example, in the Poehlman case, which involved ORI, the University of Vermont made 30 findings of misconduct, but after reviewing all of the evidence, ORI added another 20 findings, Pascal notes.

The penalties imposed by ORI are similar to those levied by other agencies, with federal-wide disbarment being the big stick. Other "rehabilitation sanctions" include requiring the individual to develop a supervision plan at his or her university or to retract or correct affected literature.

For cases at ORI where research misconduct is found, about 10 to 14 per year, the office makes public the individual's name, his or her institution, and other details about the case in its annual report. According to Pascal, this release of information is done to make it more difficult for individuals to move to a new institution unnoticed and to let the scientific community know that research misconduct is occurring and that there are consequences for such actions. Other government agencies, however, do not include the identity of the guilty party when they report their misconduct findings.

In the end, no matter how seriously federal agencies take research misconduct, it comes down to researchers making ethical decisions in their work. "Not every wrong piece of data out there is research misconduct," points out NSF's Busch. "It's just a matter of what you do with it."