Conflict of interest in all its many forms requires institutional guidance
Policies must address myriad issues
The editors of the prestigious British medical journal Lancet recently issued a public statement acknowledging a failure to disclose conflict-of-interest concerns about a 1998 study they published connecting autism to childhood vaccines.
On Feb. 20, Lancet editors announced the results of their investigation into allegations of research misconduct against the authors of a paper detailing possible links between the combination measles-mumps-rubella (MMR) vaccine and autism and inflammatory bowel disease in children.1 Publication of the study, and the resulting publicity, has been cited as a major factor in declining vaccination rates in the United Kingdom.
Now, journal editors have learned that the lead investigator, British researcher Andrew Wakefield, was conducting a related research endeavor at the same time the autism project was under way. The second project, about which editors claim both they and Wakefield’s co-authors were unaware, involved Wakefield collecting information to determine whether there was clinical evidence to support a legal action by parents claiming the MMR shot had harmed their children.
"We regret that aspects of funding for parallel and related work and the existence of ongoing litigation that had been known during clinical evaluation of the children reported in the 1998 Lancet paper were not disclosed to editors," the statement by editor Richard Horton said. "We judge that this information would have been material to our decision making about the paper’s suitability, credibility, and validity for publication."
The paper’s key authors also responded in writing to the journal’s statement, with Wakefield denying that his participation in the second project compromised the integrity of the initial research or the journal publication.
"The clinical and pathological findings in these children stand as reported," Wakefield states. "They have now been confirmed independently by reputable physicians and pathologists. . . . My colleagues and I have acted at all times in the best medical interests of these children and will continue to do so."
The Lancet debate highlights the difficulty that many research institutions — universities, medical journals, and medical centers — have in managing potential conflicts of interest.
In the United States, there are no clear, universal standards for what constitutes a conflict of interest for a researcher, research sponsor, or institution, says Robert S. Bienkowski, PhD, executive director of research at Iowa Health Des Moines, a three-hospital health system.
"There is not much language in the federal regulations beyond that stating that conflicts of interest must be disclosed and managed," he says. "And there is no simple way to say that something is a conflict or it is not. What might be a serious issue at one institution might never come up at another one."
Research institutions must develop their own policies and procedures about research conflict of interest (COI), including what it deems to constitute a potential conflict, policies and procedures for discerning and disclosing potential conflicts to participants, and what institutional processes will be involved in managing potential conflicts so that they do not impact the ongoing research.
Institutions might want to start by modeling their own policies after that of another institution, but the development of a policy should be seen as a process, he cautions.
Some larger universities may see types of conflicts that another institution might not deal with, and detailed procedures dealing with that type of conflict would be unnecessary.
For example, some research institutions have developed separate, for-profit corporations to market commercial uses of the results of some of its sponsored research. A researcher might have a dual role, doing work both as an investigator and in developing markets for the separate corporation.
That institution would need policies about what types of research the investigator could and could not participate in, and how the dual relationships would be disclosed to subjects in projects the investigator would be involved in.
At the same time, as a smaller institution gains experience in research, it may discern areas of potential conflict unique to that setting.
"It’s important that this be a process that can continue to evolve," Bienkowski says. "And it’s important that it not be just the IRB that is involved in developing conflict of interest policies."
Representatives from the institution’s administrative and executive offices should, if not directly participate in policy development, be at least very aware and supportive of the ongoing development of institutional policies on COI, he notes.
At many institutions, a separate committee or task force is charged with developing and institutional policy.
Although the stereotypical COI scenario involves an investigator with a monetary interest in an entity sponsoring research — for example, a clinician who holds stock in a company that later wants to sponsor research in the clinician’s practice area at his or her institution — it is important to realize that not all COIs are financial.
"Money is a frequent issue, but you have to look at other potential conflicts as well, that might not be as obvious," Bienkowski says.
The potential of certain research findings to be published in a prestigious journal, for example, or a particular clinician’s strong interest in a particular disease or condition, could influence his or her conduct in a research protocol.
And potential conflicts exist at the institutional as well as the individual level, he adds.
An institution’s mission to treat a particular illness or disease, for example, could produce an institutional bias in favor of pushing the envelope in clinical research.
For example, a noted cancer hospital would be an ideal setting for oncology research, but how would the institution ensure that appropriate protections existed to balance research progress with protections against coercion and undue influence on patients to participate?
When is a COI too much to manage?
IRBs will usually be the entity deciding whether a potential conflict of interest can be addressed by implementing certain safeguards, or whether a particular investigator should not be allowed to participate in or conduct certain trials.
Again, different institutions may make vastly different decisions with regard to what kinds of COI they feel they can comfortably manage, says Bienkowski.
Where one institution may simply prohibit an investigator from conducting a particular project, another institution — possibly with more experience in that particular area — will feel comfortable implementing certain procedures to ensure the conflict does not influence research conduct.
For example, a clinician who is an expert in a particular treatment area and has conducted groundbreaking research of that particular type, may also have business ties the company seeking to sponsor a trial.
An institution may decide that if they prohibit this person from involvement in subject recruitment — and restrict him to only the study design and interpretation of data — that the integrity of the data could be protected.
Another IRB, faced with a perceived institutional conflict, might decide that a particular study would best be conducted elsewhere. Different IRBs will be of different minds.
At some institutions, such as Vanderbilt University in Nashville, TN, a separate committee has the responsibility for evaluating COI, says Gerald S. Gotterer, MD, PhD, senior associate dean for faculty and academic administrative affairs at Vanderbilt University School of Medicine.
The university’s faculty COI committee assesses the potential impact of a conflict of interest on the design and implementation of a research project, but then the IRB decides on appropriate measures of disclosure to the patients/subjects.
"Both committees have a role in relationship to COIs," Gotterer says.
The more research conducted at an institution, the more need there is for such "separation of powers" in terms of evaluating the potential impact of COIs and making decisions about how COIs will be managed, Bienkowski notes.
Such extra safeguards help prevent institutional complacency and acceptance COIs, he says.
Screening for conflict
Above all, it is important for institutions to have some way of evaluating investigators for conflicts of interest that is not largely dependent on self-reporting, Bienkowski says.
If the institution has set out definitions of what it considers to be areas of potential conflict, it should still not be left up to investigators to report this to the IRB up front.
"Even those of us with the best of intentions often believe we are incapable of being unduly influenced," Bienkowski notes. "We can feel absolutely certain that we would be able to disregard certain influences in favor of the integrity of the research."
It is up to the IRBs to develop procedures for questioning potential investigators about any potential COIs in such a way that they will feel inclined to report clearly and completely.
"You have a questionnaire that goes over each potential area thoroughly," Bienkowski explains. "For example, Do you have a financial relationship with any of the sponsors of this trial?’ not, Can you report any potential financial conflicts of interest related to the sponsor?’"
Few researchers believe themselves vulnerable to COI, but all are, he continues.
Money, though many will deny it, can be a difficult inducement to resist.
Several years ago, during the course of a clinical trial that had run vastly over budget, Bienkowski and colleagues received an unexpected bonus check for subject recruitment — a practice strictly prohibited by institutional policy.
"We didn’t agree to it or apply for it or anything — it wasn’t mentioned in any contracts," he says. "The check just showed up."
Bienkowski says he is almost embarrassed by how much he wanted to rationalize keeping the money.
"We were way over budget, and the check just appeared. It’s amazing the lengths that your mind will go, thinking up ways that you might be able to cash that check," he notes. "But it clearly said recruitment bonus’ right on it, so we sent it back."
Since then, the experience has served as a lesson that institutions cannot be too careful when considering COI.
"No one wants to admit it, but we are all vulnerable," Bienkowski states.
1. Horton R. A statement by the editors of Lancet. Lancet 2004; 363:820-824.