It might be a new year, but old problems haven’t disappeared. Top on my list of concerns are institutional conflict of interest and research oversight. Although the Duke scandal in which Anil Potti, Joseph Nevins, and others took corrupted data into clinical trials highlighted both issues, the community doesn't seem ready to confront them.
To be fair, the Institute of Medicine's 'omics committee (officially called the Review of Omics-Based Tests for Predicting Patient Outcomes in Clinical Trials Committee) raised both topics during their deliberations. However, based on the proportion of time spent on these issues versus on other topics, such as data reproducibility, I don't expect it will be a key focus of their report. (Moreover, I don't think we can expect them to address every issue raised by this multi-armed fiasco.)
Meanwhile, much of the press attention, driven in large part by the reporting of the Duke Chronicle and The Cancer Letter, has been devoted to uncovering Potti's own whereabouts and ill behavior.
It would be a mistake, though, to let this opportunity slip by without attending to these fundamental issues in science. After all, the problems at Duke, at their heart, are not unique to Duke. Sure, Potti only messed with data at Duke and it was the Duke internal review board (IRB) that rubber-stamped trial approvals even when sections of paperwork were incomplete (e.g., the section where the team should have provided an investigational device exemption number or state that such a regulatory exemption was unnecessary). But when we start to ask why those things might have happened or why they were allowed to continue unattended for so long (despite outside criticism), we come to issues that affect every research institution and our research system as a whole.
Wolf guarding the chicken coop
It is well known that Duke stood to make many millions of dollars if the work by Potti, Nevins, and colleagues succeeded. And while I don't think anyone purposely let that cloud their decisions, I can't imagine how people who are part of an institution could be entirely immune to it –– or indifferent to the laudatory media and scientific attention the work had already garnered.
There is not much concrete data on how IRB members deal with conflicts, according to Helena Hoas, PhD, from the National Rural Bioethics Project at The University of Montana in Missoula.
However, the data that are available, from Hoas and others, suggest that IRB members feel responsible to multiple parties, including their institutions – and not just to patients (see past discussion here). And those insights come from IRB members reviewing typical protocols, not protocols that have unusually high profiles or value within the institution.
Yet, when Duke leadership appeared before the IOM committee on August 22, 2011, Ross McKinney, MD, Institutional Official 2006/2007, and Chair of the Duke Conflict of Interest Committee, and others said that they did not use an external review board to evaluate Potti's clinical trial protocols because the technology under study had not been licensed at the time of the IRB meetings.
It is worth noting that neither McKinney or other Duke representatives argued that the technology was not likely to be valuable, but rather they emphasized that it was simply a matter of timing: Because no current licensing agreement was in place, their standard operating procedure was to let the IRB handle the protocols, regardless of the money-making potential of the work.
When Jesse Gelsinger died in 1999 as a result of treatment he received in a clinical trial, the scientific community tightened individual conflict of interest rules.
It doesn't appear that anyone died directly as a result of the therapy they received during the Duke clinical trials. But what we – and the patients and their families – will never know is whether the 100 or so patients might have done better on a different therapy than the one they were given as a result of participating in the trial.
Shouldn't that concern be enough to trigger some stronger rules about institutional conflict of interest? How hard would it have been for Duke to refer the protocols to an external review board? Based on McKinney's comments to the IOM, that option was available, the institution just chose not to use it.
If the impact of institutional conflict of interest still seems vague, consider the comments made by Sally Kornbluth, PhD, Vice Dean for Research, at the same IOM meeting, regarding the outside review Duke convened in the fall of 2010. When asked why the external reviewers were not given a copy of a letter she received from MD Anderson Cancer Center researchers that detailed specific problems with Potti and Nevins' work, she said she had passed it on to the IRB chair, John Harrelson, who was overseeing he review. He then showed it to Nevins, who said that providing the letter to the external reviewers would bias them. Harrelson followed Nevins instructions and withheld the letter.
In other words, the people being investigated were running the investigation.
"I think one of the chief lessons learned is that there is a balance between trusting investigators who have a very long track record with an institution with thinking about what is necessary to ensure an adequate review," Kornbluth acknowledged to the committee.
Duke has subsequently developed a detailed plan to improve oversight, called the Translational Medicine Quality Framework, but again the issues of conflict of interest and oversight are not issues only at Duke.
"One of the great problems that I worry about –– especially in this world of biotech, where physicians and scientists are encouraged to get patents and, in some instances, to set up businesses that parallel their research at the university –– one of my great concerns is the conflicts of interest are growing, " said Otis Brawley, MD, Chief Medical Officer for the American Cancer Society. "We are being asked to police ourselves, while we are also being asked to do things that will generate income."
If not the wolves, then who?
So if institutions are conflicted and therefore unable to provide adequate oversight for their investigators, who can do it? The Office for Human Research Protection (OHRP) is supposed to ensure that human subjects research is done ethically, but if concerns are raised, the agency asks the home institution to investigate. Similarly, the Office of Research Integrity (ORI) reviews and monitors misconduct investigations done by the home institution. Neither agency is an investigative body.
In fact, when the MD Anderson researchers, Keith Baggerly and Kevin Coombes, learned that clinical trials were underway based on the Duke work, which they deeply believed was wrong, they contacted ORI. The agency asked, according to Baggerly, whether he and Coombes could prove misconduct. As they couldn't prove it, the agency couldn't help them – despite on-going patient trials.
As I understand it, neither agency has the staff or mandate to take on an investigative role, but that could change. Given the vast sums of money the federal government spends on research, wouldn't it make sense to spend a little more of it ensuring it is being spent on honest science that benefits and protects patients?