Last week lawyers filed suit against Duke University and others on behalf of patients who participated in the clinical trials that stemmed from Anil Potti’s work. The civil lawsuits, filed in the Durham County Superior Court in North Carolina, claim that the eight patients (only two of whom are surviving) participated in trials that were fraudulent and occurred under false pretenses, and that the patients were exposed to improper and unnecessary chemotherapy which delayed appropriate care.
The defendants named in the suits are:
- Duke University
- Duke University Health System, Inc.
- Private Diagnostics Clinic PLLC
- Joseph Nevins, PhD
- Anil Potti, MD
- Michael Cuff, MD
- Sally Kornbluth, MD
- John M. Harrelson, MD
- Cancer Diagnostics, Inc.
The lawsuits are not a surprise.
In fact, on August 22nd, when eight members of the Duke faculty and senior administration met with the Institute of Medicine committee charged with sorting out the matter, Rob Califf, MD, Vice Chancellor for Clinical Research, explicitly stated that the possibility of lawsuits would constrain the discussion to some degree. He made a similar caveat regarding an ongoing misconduct investigation.
Despite those limitations, the Duke representatives answered questions at length and seemed to have spent a good deal of their own time tracing over the events.
Several times during the eight and a half hours of discussion with the committee, Dr. Califf referred to the “Swiss cheese effect.” He said Duke’s internal reviews have identified 15 steps that went wrong and thereby allowed Dr. Potti’s work and trials to proceed unchecked. The list includes, for example, problems with clinical trial protocol review, and understanding of the rules and responsibilities of co-authorship and when a consultation with the FDA is needed for investigational devices or drugs.
“I think in general these 15 things do work well at Duke,” Dr. Califf said. “But in this case, they all lined up to produce the results that you saw.”
The institution has made numerous changes, some of which were already underway when the problems occurred, according to the representatives. But, Dr. Califf warned, no system will be perfect.
“We believe that with all of the adjustments going forward in all of these areas, that the odds of [the holes] all lining up again are quite small,” Dr. Califf said. “Not zero. And I don’t think we can produce a zero-error system.”
Too Much Trust in Colleagues
The IOM committee members, led by Tom Flemming, PhD, of the University of Washington and David DeMets, PhD, of the University of Wisconsin, spent a fair amount of time asking about internal review board (IRB) oversight and how the external review of Dr. Potti’s work was handled in the fall of 2009.
The Duke representatives acknowledged that mistakes were made. For example, they decided to have the external review held in private so the reviewers could be completely honest about their view of the work. However, the approach left many people concerned that the review was inadequate.
Additionally, it meant that community members could not share their concerns directly with the reviewers, who were ultimately dependent on the information Drs. Nevins and Potti gave them.
“I think one of the chief lessons learned is that there is a balance between trusting investigators that have a very long track record with an institution with thinking about what is necessary to ensure an adequate review,” Dr. Kornbluth said.
That misplaced collegiality ultimately explained how essential criticism from Keith Baggerly and Kevin Coombes of The University of Texas MD Anderson Cancer Center about data provenance was withheld from the external review committee.
When Dr. Kornbluth received a letter from the Houston researchers detailing their criticism, she forwarded it to John Harrelson, MD, the IRB chair who was overseeing the external review. He, in turn, sent it to Dr. Nevins.
“[Dr. Harrelson] knew that Joe Nevins and Anil Potti were preparing all the materials for the review,” Dr. Kornbluth explained. “So we sent it to them and said ‘Please take this into account as well and address these comments, so that when the reviewers see it, they will see your answers too.’ Dr. Nevins had a very negative reaction. Basically he said that it would bias the review.”
Dr. Kornbluth defended the approach by saying that at the time she still thought they were dealing with a matter of methodological disagreement. Also, she thought everyone – including Drs. Nevins and Potti – wanted to get to the root cause of the problem.
Conflicts of Interest
The Duke representatives acknowledged that individual faculty members had conflicts of interest in this case. However, the way Duke identifies and deals with potential conflicts seemed rather convoluted. On one hand, faculty are supposed to declare their conflicts to the office that handles such issues. On the other hand, it became clear during the discussion that some of the faculty were not even aware they were listed on licensing agreements.
Moreover, the university seemed slow to develop management plans to deal with conflicts of interest. Specifically, the institution had their first management plan for Dr. Potti only in late 2009, which is shocking given that he is named on five patents and his work was behind multiple clinical trials. In one case, he even served as the formal primary investigator for one of the trials, despite those financial conflicts.
The issue of institutional conflict of interest was discussed as well, but only briefly. Ross McKinney, MD, Chair of the Duke’s Conflict of Interest Committee, said the institution had multiple layers in place to protect patients in cases where the institution had a financial interest. As expected, the conflict of interest group first considered whether the trial should be run at Duke. They decided to allow it because the investigators convinced them there was special expertise needed to analyze patient samples.
Therefore, they established a “DSMB-plus.” The DSMB-plus is told explicitly that there is an institutional conflict and asked to monitor the design, conduct, and reporting of the trial.
“We have other experiences with that model that worked better than this one, but it is still an experiment in progress I would say,” Dr. McKinney said. “We are still learning whether you can get adequate oversight.”
The conflict of interest committee also requires that the institutional conflict be stated clearly in the consent documents and any public presentations.
Finally, the university uses an external IRB if the intellectual property involved has been licensed. In this case, there was no licensing agreement at the time of the clinical trial reviews, so the internal IRB handled the decisions.
I have to wonder though -- given that Duke stood to make so much money in this case (some estimates of hundreds of millions of dollars) -- if the institute’s approach to institutional conflict of interest was a bit too formulaic. After all, how hard would it have been to engage an external IRB even in the absence of a licensing agreement? And if the conflicts were serious enough to warrant a discussion about whether the trial should be run at a different institution, why didn’t it automatically warrant an external IRB?
As I mentioned in a previous post, there is evidence that IRB members feel obligated to do more than just protecting patients. When a high profile project (don’t forget the Nature Medicine, New England Journal of Medicine, Journal of Clinical Oncology, and Lancet Oncology publications) lands in their committee, can they really be expected to be immune to peer and institutional pressure?
A Clear Need for FDA Filing
During a discussion of the IRB oversight of the trials, the Duke representatives made clear that they should have required that the investigators obtain an investigational device exemption (IDE) from the FDA before allowing the predictors to be used in clinical trials.
(This view is consistent with statements Alberto Gutierrez, PhD, Director of the FDA’s Office of In Vitro Diagnostics Device Evaluation and Safety, made during an August 19th discussion with the IOM committee.)
Remarkably, though, the Duke representatives, including John Falletta, MD, Senior Chair of the Institutional Review Board, still seemed unclear on the potential harm faced by the patients. The IRB view seems to be that since the patients would receive treatments considered standard in the field, there was no potential harm from treatment assignments based on the predictors. (Dr. Falletta did acknowledge that the extra biopsies required in the trials raised the level of concern for the committee.)
Misconduct Investigation Far Along
With the lawsuits filed, the next expected news event for the Potti scandal will be the conclusion of the misconduct investigation. Though no one from Duke discussed any details of the investigation, Dr. Califf said it was “far along.” He did not say when we might hear the results of the investigation.
In the meantime, Dr. Potti is still practicing oncology. He is now at The Coastal Cancer Center, in South Carolina, an institution that conducts clinical trials.