BACK TO THE FEATURE INDEX

Conducting scientific research is, in essence, a quest for knowledge and truth--or is it? High ethical standards are expected of everyone involved in the research enterprise, but can they be achieved? Genuine scientific curiosity is likely the driving force for most scientists. However, the pressures to succeed and the temptations of recognition and influence can be corrupting forces.

Defining what exactly constitutes scientific misconduct has been difficult even for international experts, but most scientists will recognize unethical conduct when they encounter it. But how to proceed once alleged dishonesty has been identified? How could the fraud have been prevented? Who shares responsibility?

The case scenario presented here addresses issues of both due process and responsibility. Let's examine the latter first.

Terry, a first-year graduate student, was working under the "guidance" of postdoctoral fellow William after William and the PI, Sara, had laid out the project idea. When Terry obtained dramatic results within the short period of 1 year, postdoc Tina helped to edit the manuscript before Sara reviewed it. This scenario suggests that Sara was not intimately involved in the review of Terry's primary data. Two PIs from other institutions contributed research material and accepted co-authorship on the high-profile publication. When pharmaceutical company scientists failed to reproduce Terry's results, they discovered, as a result of Sara's willingness to share Terry's notebooks, strong evidence for possible fraud committed by Terry.

How could the fraud go undiscovered for so long? Who is responsible?

All authors on a publication share responsibility for its content. That's why only those involved in the conceptualization, conducting, interpretation, and/or writing of the paper should be included. As far as we know, the external co-authors in this case (Mary and Toby) only contributed reagents; most likely, they had no direct knowledge of Terry's primary data or significant input in the data interpretation. If so, they should not have accepted "ghost" co-authorship and have themselves acted improperly. If, however, they were more intricately involved, they bear some responsibility for not detecting the fraud. Regarding the potential public health impact, perhaps Mary and Toby should have demanded extra scrutiny of Terry's results.

As the PI and corresponding author, Sara is accountable both to her institution and to her funding agency. She also bears responsibility as a mentor, supervisor, and group leader. Did she regularly review and discuss Terry's original data with Terry and William? After all, he was a first-year graduate student! Or did she rely too heavily on William as Terry's "guide"? Did she mentor her postdoc on how to be a "substitute" supervisor? Although it is not inappropriate to ask a more senior lab member to mentor a student to a limited degree, that does not release the PI from her duty to supervise personally a student whom she accepted into her lab.

A number of other questions spring to mind: Should, or could, William have known the true results from direct observations in working alongside Terry, especially since the data seemed so important, instead of being deceived by "too-good-to-be-true" data? Or was William too overwhelmed by the task--surely he had his own project to worry about? What about the other lab members, especially Tina, who edited the manuscript? How well did they know Terry's work? Did Sara encourage open discussion and honest critique among the lab members? Should Sara have insisted on extra scrutiny to verify the results, since they were so potentially influential? How brilliant did the alleged deceit have to be to go undiscovered, especially since Bob Smith (the "whistleblower" in this case) easily revealed the allegedly fraudulent pencil data entries?

The case study does not provide enough information to determine the exact role each person played. But the resulting questions highlight that neither improper conduct nor honest mistakes happen in a vacuum--the way we organize our labs and interact with one another can have a huge impact on the relevance, and accuracy, of our research. Lessons from the case study apply to any young investigator on the brink of independence and at the transition from mentee to mentor.

In research, we teach and are taught by example as an intricate part of our work and our student-adviser relationships. However, I believe that young aspiring scientists should also be given the opportunity, in a formalized setting, to reflect on how best to achieve and maintain an ethical research environment. This can help us make deliberate choices, especially in situations that are characterized by apparently conflicting interests, as exemplified in this instance by the relationships among PI Sara, postdoc William, and graduate student Terry. We need to recognize that, for example, the delegation of responsibility, while providing opportunity for growth, must be balanced by appropriate guidance, and that trust in a colleague's work needs to be balanced by reasonable professional skepticism. Without a conscious decision on how we want to shape our own lab environment, the tough road from postdoc to independent PI could get even bumpier.

How should the investigation of the alleged misconduct proceed?

The need for due process following misconduct allegations was recognized in the United States about 20 years ago. Since then, the Office of Research Integrity (ORI), part of the U.S. Public Health Service (PHS), has developed a set of rules on how respond to allegations of research misconduct. European countries are also moving forward with formal plans for implementing due processes [Science, 17 December 1999, p. 2258; Lancet 354, 57 (1999)]. Institutions that receive PHS support (and Sara, the PI here, likely does) need to comply with the ORI rules and establish institution-wide policies for misconduct inquiries.

According to ORI guidelines, the whistleblower, Bob Smith, should formally notify the relevant office at the institution about his allegation. If concerned about conflict of interest within the university, Bob might also contact ORI directly, which would then inform the institution. Under the circumstances--that Sara cooperated in resolving the discrepancy in experimental outcomes by providing Terry's notebooks--it seems fair that Bob directly informed Sara. Because her reputation is at stake, she may or may not choose to join in alerting her institution of the allegation against Terry. Due to the high impact that a finding of misconduct would have not only on Terry, but also on Sara and her university, the diligent implementation of due process is especially important.

During the initial inquiry into whether a full investigation is warranted, the accused (i.e., Terry, and possibly Sara, for negligence) must be notified and given the opportunity to respond. If initiated, a full investigation should proceed in a timely fashion. The perceived public importance of the research paper warrants the inclusion of independent investigators from outside the institution to minimize a potential conflict of interest. At the same time, and especially because of the potentially damaging outcome of an unsubstantiated accusation, the investigational proceedings should be kept confidential. Because patients are not at risk, this is both justifiable and warranted by ORI policy. If and when misconduct is determined, all involved funding agencies, the journal that published the article, and--in this case--the public, should be informed. Thereafter, the institution and PHS may take further disciplinary actions, such as expelling Terry and debarring Sara from future funding.