Important Points Highlighted by Individual Speakers
- A financial relationship does not of itself indicate a conflict of interest, which is why transparency of relationships with context is important.
- Better mechanisms are needed to detect and measure non-financial relationships that may pose a conflict of interest.
- Improving transparency by clearly defining goals, problems, and responsibilities can be helpful when designing precompetitive collaborations.
- Resolving legal conflict of interest issues within the scope of the collaboration would promote innovation.
- Research needs to drive risk–benefit analyses for collaborations, and legal processes need to be devised to enable partnerships within the framework of conflict of interest policies.
- Innovative ways for accessing expertise are needed to provide prompt knowledge and guidance within current policies.
Each member of a collaboration has not only different interests but also a perspective on that collective effort that is different from the other members of the collaboration. These different perspectives can be a spur to creativity, but they can also create conflicts. Individual speakers at the
workshop examined conflicts of interest in relation to transparency, collaboration, and access to expertise.
Conflict of interest policies are designed to reduce incentives that can lead to biased research results and to preserve the public trust, said Pierce. “This is absolutely fundamental to our research enterprise” and to innovation, she said.
Various organizations have conflict of interest policies, including the federal government, state government, research institutions, and professional societies. The policies of these different organizations have overlapping goals—including protecting data and research subjects, accelerating innovation, and achieving better health for all—but they are not identical. The guidance and limits of the policies of a professional society may have different impacts on innovation than disclosure policies reviewed and managed at institutions, Pierce said.
Pierce agreed with other speakers that money is often used as a surrogate in conflict of interest policies for thinking about bias, largely because funding can be measured. But she would not recommend applying the surrogate structure to other types of conflicts. For example, it would be very difficult to write a regulation governing an academician’s objective to publish a paper to improve his or her chances of getting tenure. Instead, other mechanisms may be able to detect and allow for the management of conflicts, such as the third-party analysis of data.
Stakeholders need to know that a process exists for reviewing potential conflicts and determining whether there is a risk of bias, Pierce said. Reporters sometimes ask Pierce questions about a discovery of a financial relationship between a physician and a certain company that is worth, say, $10,000. But such a financial interest is not necessarily a conflict of interest, she said, and the “discovery” of such a financial interest does not necessarily connote a problem. Given the approach that some reporters take, she said, it may seem to investigators that they have an uphill battle if they wish to engage in industry collaborations.
Transparency is important, but transparency without context can be problematic. For example, Pierce described a theoretical example of a simple relationship between an investigator and a pharmaceutical company that could lead to the reporting of different financial relationships depending on the timeframe and the party that is asking for the disclo-
sure. In this scenario, the investigator received a $2,500 consulting payment in August 2013. The investigator also served on a scientific advisory board for which she was paid $1,000 in December 2013 and another $1,000 in February 2014. However, because her institution does not require its investigators to report financial interests that total less than $5,000, she does not have to disclose those relationships to the institution. At a continuing medical education seminar in May 2014, she discloses that she has earned $4,500 in the past 12 months from the pharmaceutical company. At a conference in September 2014, her disclosure is $2,000 over the past 12 months. When the Physician Payments Sunshine Act database appears that same month, it may list payments made to the investigator at $5,100 because she has received $4,500 plus the transfer of value of a plane ticket, hotel, and meals. Without the context, the variation in disclosures could be misleading.
Conflict of interest policies may create transparency and public confidence without reducing bias as they are intended to do, observed Lichter. Physicians may withdraw from partnerships because of their fear of possible negative consequences, such as being listed as a scholar who is supported by industry.
A lesson that could be applied to conflict of interest policies may be learned from the regulations that were put in place to limit the duty hours of medical trainees, Lichter said. The supposition was that medical residents were tired from working long hours and that this was leading to mistakes in patient care (IOM, 2008). In 2003 and 2011 work hours were limited by regulations from the Accreditation Council for Graduate Medical Education with the goal of improving patient safety. However, outcomes data have shown that duty-hour limitations have not only failed to improve safety, but they may have worsened it because of more frequent hand-offs among trainees, Lichter said. A better approach would have been to allow some institutions to try different approaches and then study the outcomes.
Without data, policies are based on what people think is reasonable and a good idea, said Lo. Gathering the necessary data will be a challenge, because it will require characterizing, identifying, and analyzing the good and bad consequences of policies. But, Lo said, “it would be a real shame if 10 years from now we’re in the exact same place.” Evidence and data are needed to form good policies. Sometimes there is not a strong evidence base, Rockey said, and in the case of the NIH financial conflict of interest policy, a proactive strategy for regulations was put in place.
Cohen suggested that in the duty-hour circumstances, patient care problems were caused by lack of supervision rather than sleep deprivation. “We put a solution in place that was for a different problem.” In the case of conflict of interest, the problem that needs to be fixed is preventing relationships that compromise the quality and integrity of the science and the public trust. Some individuals and institutions have compromised the public trust, and even if the numbers are small, the problems that have resulted are significant. Asking how these conflicts occurred and how to prevent them in the future does not imply that everyone has a conflict of interest that is compromising their actions. Rather, addressing the issue is meant to enable reasonable and responsible relationships.
AAMC has a conflict of interest metrics project that has been designed to collect data from academic institutions on these issues, Pierce said. One goal of the project is to aggregate the data from institutions and provide it to NIH to facilitate a retrospective review of the impact of its conflict of interest policies. AAMC originally hoped that 25 institutions would sign up for this project, which requires 4 years of data collection and reporting. Already, 76 institutions from across the country have signed up and have provided extensive historical data from before the new rule was implemented, Pierce said.
Pierce said that the questions being asked by the project include: What did you do to prepare for the new rule?, What was the cost of implementing this rule?, Did it require capital investments or infrastructure changes?, Did you hire new people?, What is the impact on your faculty?, What were you looking at before?, How many financial interests did you collect?, How many were determined to be significant financial conflicts of interest and reported to NIH or another funding entity?, and Are you finding and reporting conflicts of interest related to travel?
AAMC is starting to analyze the background data and intends to provide data updates as the project proceeds, Pierce said. Early analysis of data from before the rule compliance deadline will look at implementation costs, conflict of interest review processes and infrastructure, and non-financial impacts on institutions and faculty.
Finally, Pierce stated, as did other workshop presenters, that the community is good at communicating the risks of conflicts of interest but that it has not done as good a job of identifying and communicating the benefits of collaboration among academia, government, and industry. “If we are not pulling together the best resources from our academic scientists, from our industry-trained scientists, and pulling together the resources that each can bring to the equation,” Pierce said, “we’re missing
out, and not just in small ways but in profound ways that affect discovery in health.”
The mission of NCATS at NIH is to develop innovative methods and “technologies that will enhance the development, testing, and implementation of diagnostics and therapeutics across a wide range of human diseases and conditions.”1 The fact that resources are limited means that achieving that grand mission requires collaboration, which in turn requires the management of conflicts, said Krishna Balakrishnan, senior technology manager in the Office of Strategic Alliances at NCATS.
Collaborations should start at the home institution, said Balakrishnan. All NCATS projects have multiple, cross-disciplinary leaders working collectively as a team, with everyone from the laboratory research personnel to the principal investigators participating in the planning as well as the implementation of the projects. NCATS also works collaboratively with many of the other institutes and centers at NIH. Collaboration is “in the DNA” of those who work at NCATS, Balakrishnan said.
The Therapeutics for Rare and Neglected Diseases program, which supports drug development collaborations between NCATS and extramural partners with expertise in disease areas or drug targets, is an example of a successful NCATS collaboration, Balakrishnan said. Most pharmaceutical and biotech companies would not usually develop these drugs because of the low financial incentives. However, in the past 3 years, the program has undertaken 15 projects, of which 4 have produced investigational drugs taken into human trials. In a program with Genzyme, the University of Cincinnati has explored the repurposing of granulocyte macrophage-colony stimulating factor (GM-CSF) for pulmonary alveolar proteinosis, a very rare disease. In this case, Genzyme supplies the drug at no cost, and the university provides the disease expertise to develop an inhaler-based formulation of GM-CSF and to complete a toxicology study prior to studying its use in clinical trials.
The Discovering New Therapeutic Uses for Existing Molecules program, which Rockey mentioned, seeks to repurpose existing compounds
1National Center for Advancing Translational Sciences. Re-engineering Translational Sciences, http://www.ncats.nih.gov/research/reengineering/reengineering.html (accessed December 31, 2013).
using innovative ideas from researchers. Collaborations are formalized between researchers and industry through collaborative research agreements (CRAs), while memorandums of understanding are used to standardize relationships between industry and NIH. The program also uses confidential disclosure agreements and CRAs between pharmaceutical company partners and applicants, Balakrishnan said. Industry provides the compounds, and they are crowd-sourced to investigators for ideas about how to find new uses for them. The templates are available for the agreements after a relationship is established so that putting new agreements together is not as time consuming. Much of the information, including the CRAs, is available publicly.
Failure of drugs in Phases II and III of clinical trials is often due to drug toxicity and a lack of efficacy, Balakrishnan said. NCATS is trying to address these issues by collaboratively developing a tissue chip, which is an in vitro platform that uses human tissues instead of whole animals to predict drug efficacy, pharmacokinetics, and safety. The intention is to create a modular, reconfigurable platform to produce physiologically relevant, genetically diverse, and pathologically meaningful results. “The problem is massive, because it’s not only a biology problem—it’s an engineering problem, a bioinformatics problem, and a fluidics problem,” Balakrishnan said. The collaboration includes FDA and the Defense Advanced Research Projects Agency and has resulted in requests for applications from academic researchers to develop such chips.
NCATS has developed a set of principles (see Box 3-1) to govern how it approaches collaborations, Balakrishnan said, and these suggested rules could be helpful for other types of collaborations as well. It is important to define problems clearly by setting agreed-upon directions, goals, and responsibilities, he said. In doing so, collaborators are encouraged to look for ways in which there are synergies between complementary
- Complement, rather than compete with, the work of others.
- Revolutionize the process of translation by promoting innovative research.
- Expand the precompetitive space.
- Support and augment regulatory science and its application.
- Galvanize and support new partnerships.
SOURCE: Krishna Balakrishnan IOM workshop presentation June 5 2013.
assets of the partners involved. In collaborations with industry, the science drives the collaboration, Balakrishnan said. Projects focus on new technologies, enabling tools, and dissemination and on de-risking novel therapeutic approaches for industry adoption. Collaborations are typically selected through a solicitation process that is open to anyone. The peer review committees that consider solicitations have members not only from NCATS but also from the venture capital industry, the biotechnology industry, academia, and other sectors. NCATS has also worked to establish itself as an honest broker by being transparent in its work and by designing projects in the precompetitive space so that there is sharing in successes. Lastly, Balakrishnan mentioned the importance of recognizing when a project needs to come to a conclusion and to disengage from the collaboration when it is appropriate. If this is accomplished in a professional manner, then there should be opportunities for the collaborators to work together on future projects. NIH policies, rules, and regulations are strictly enforced and communicated to the collaborators up front. NCATS forms collaborations within the boundaries of conflicts of interest. These policies should serve science rather than penalize it, Balakrishnan said.
Federal advisory committees provide formal recommendations on policy to federal agencies, said McMurry-Heath. The committees are composed of individuals who have expertise in the subject matter, but there are limitations on who can serve based on conflict of interest policies. FDA provides as much transparency as possible, but the agency needs to follow the regulations that were approved by Congress, said Jill Hartzler Warner, acting associate commissioner of special medical programs at FDA.
The decision on whether an expert can serve on an FDA advisory board is made by a team of lawyers who advise FDA, said Henry Brem, chairman of the Department of Neurosurgery at Johns Hopkins University School of Medicine. When volunteering to serve on an FDA advisory panel for neurosurgical devices, he had no financial interests in the companies working on the devices, but he needed to disclose significant amounts of information about himself and his immediate family during the conflict of interest reviews. He was ultimately disqualified, not because of a financial conflict, but for an academic conflict because a
treatment he had developed could potentially compete with a treatment he would be considering in the advisory committee. The decisions are being made through a risk–benefit analysis, Cohen said, but the risks and benefits are being assessed by lawyers who are not involved with the dialogue about the goals of participation. “The lawyers should be providing advice and counsel to the leadership in making the decisions, but often their advice and counsel are taken as dogma because institutions, whether they’re industrial or academic institutions, are very risk-adverse,” Cohen said. Decision makers are needed who take into account the risk but who acknowledge that disclosing the risk and being aware of it could provide the balance that is needed to move forward with something that would benefit patients and society, he said. The reasoning behind these decisions, Lo added, should be transparent to other stakeholders.
Decisions need to be made by balancing risk with what may benefit patients. At NCATS, researchers drive the process of conducting risk–benefit analyses, and lawyers are involved at a later stage of the process once the outlines of a collaboration have been determined, Balakrishnan said. “Once they have come to some sort of an agreement, the lawyers try to figure out how to make it happen.”
In addition to receiving guidance from experts serving as special government employees on review panels or advisory committees, FDA wanted to have a way to access expertise more quickly and flexibly, McMurry-Heath said. FDA achieved this by collaborating with scientific and clinical professional societies to create a Network of Experts to provide “rapid access to scientific, engineering, and medical expertise when it is needed to supplement existing knowledge and expertise.”2 When a specific scientific question arises, FDA can put out a call to the organizations in the network to find experts who can answer the question. The organizations then provide the names of relevant experts. These individuals are further screened for conflict of interest, and they fill out individual conflict of interest forms and, if needed, confidential disclosure agreements. They then can work with FDA either one on one or in group calls.
2Center for Devices and Radiological Health Network of Experts. http://www.fda.gov/aboutFDA/CentersOffices/OfficeofMedicalProductsandTobacco/CDRH/ucm289534.htm (accessed December 9, 2013).
The conflict of interest forms can be filled out rapidly, and the turnaround from the call to completed conflict of interest forms is 7 days. To achieve this so rapidly, the conflict of interest form is streamlined, McMurry-Heath said. The longest such form is one page and should take no more than 1 hour to complete. Which type of form the experts are asked to complete is based on three categories that outline the potential risk of the question being asked. Category A is for questions of limited risk. Category B is for questions of moderate risk, such as those involving a specific product line or a medical indication. Category C is for high-risk products with full conflicts of interest that need vetting. “This allows us to get experts to our scientists and reviewers within 4 to 6 weeks, which in a government timeframe is very quick,” McMurry-Heath said.
FDA has already enrolled 25 scientific and professional societies within its network, and it has invitations out to an additional 75. “We’re trying to cover the landscape so that we have access to any type of expertise that’s needed,” McMurry-Heath said.