Conflict of Interest and Medical Innovation: Ensuring Integrity While Facilitating Innovation in Medical Research: Workshop Summary (2014)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Conflict of Interest Policies: An Overview Important Points Highlighted by Individual Speakers • Achieving a clearer understanding of the phrase conflict of inter- est and of the actual, perceived, and potential conflicts will help stakeholders weigh the benefits and risks of collaborating. • Conflict of interest policies are a mechanism for preventing and managing undue influence as opposed to reacting to conflicts that may impede collaborative efforts. • Effective conflict of interest policies are consistent, enforced, and based on evidence, risk, and benefits so that the policies preserve research integrity and facilitate innovation. • Investment in academic research has been a driver of innovation, but the multiple missions of research institutes provide a chal- lenge to understanding which relationships should be considered a conflict. • The financial disclosures required by the Physician Payments Sunshine Act will provide useful information and will facilitate transparency and communication, but consideration of the context of the conflict of interest is also necessary to accurately represent the value of those. • Conflict of interest policies are intended to enhance academic– industry relationships, support the objectivity of research, and maintain public trust by managing conflicts, not by eliminating or avoiding them. 5 

6 COI AND MEDICAL INNOVATION Conflict of interest policies affect the complex relationships among many stakeholders, including those involved in academic–industry part- nerships. Individual speakers stated that defining conflict of interest and examining the perceptions and goals of these relationships can alleviate some of the challenges encountered when different stakeholders collabo- rate. Effective policies that reflect stakeholder interests can also be used to manage conflicts among collaborators. THE CURRENT LANDSCAPE: DEFINITIONS AND GOALS Conflict of interest policies need to balance countervailing goals, values, and interests, said Bernard Lo, president of the Greenwall Foun- dation and chair of the committee that produced Conflict of Interest in Medical Research, Education, and Practice (IOM, 2009). “Everything we do in clinical practice involves weighing the benefits of an interven- tion for a patient versus the burdens and risks. When we do policy work, we also have to think about the benefits of a policy versus its burdens, risks, and costs.” Because of this, Lo suggested that voluntary standards developed by professional societies may have advantages over regula- tions issued by government agencies. The connotation of the phrase “conflict of interest” itself is a chal- lenge because the phrase seems to suggest that misbehavior has already occurred. It also conflates actual conflicts, potential conflicts, and per- ceived conflicts, Lo said. The IOM committee observed that all conflicts of interest involve perceptions or appearances because they are specified from the perspective of people who do not have sufficient information with which to assess the actual motives of a decision maker and the ef- fects of those motives on the decisions themselves. Bias can only be de- tected after the fact, Lo said, which means that distinctions between perceived and actual conflicts can be misleading. A challenge that exists with describing conflict of interest is that a judgment must be made about what is an “unacceptable risk that primary interests [are] unduly influenced by secondary interests,” Lo said. What risks are unacceptable? What is an undue influence? It is common for most reasonable people to disagree over defining the risks and influ- ences, he said. 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 7 Policy Goals and Impacts One result of effective conflict of interest policies is to enable pro- ductive relationships while maintaining trust in the system. Only when a relationship is disclosed can it be assessed and managed. Conflict of in- terest policies identify collaborative relationships that may raise signifi- cant concerns about undue influence and bias, Lo observed. By doing so, the policies are intended to prevent undue influence ahead of time rather than to respond to bias after a conflicted relationship is revealed. Another goal of conflict of interest policies is to encourage relationships that fos- ter such aims as developing improved therapies for patients, he said. Col- laborations among researchers, physicians, and industry are beneficial for producing new therapies where effective treatments do not now exist to extend and improve the lives of patients. Drug development is a very diffi- cult and expensive process, Lo noted. Developing a single new U.S. Food and Drug Administration (FDA)-approved drug can require screening many thousands of compounds, putting hundreds through preclinical tri- als, and then conducting clinical trials on promising candidates. If con- flict of interest policies hamper the discovery of new drugs that are effective, patients will suffer, he said. Many different groups influence the policy makers responsible for en- acting conflict of interest policies, including the media, professional associa- tions, and advocacy groups. The underlying question, said Gabriela Lavezzari, assistant vice president for scientific affairs at the pharmaceutical industry association PhRMA, is how a patient might be affected by these policies. PhRMA performed a landscape analysis of key stakeholders in the drug development ecosystem, from academia, industry, professional asso- ciations, government agencies, legal experts, health policy analysts, and bio-ethicists, Lavezzari said. The goals of the study were to collect expert views on what constitutes a conflict of interest and to understand how tightened regulations on conflicts of interest have affected stakeholders’ work, with a particular focus on their ability to access and share scientific information. The majority of key stakeholders in the PhRMA study identified fi- nancial conflicts, followed by personal, professional, and institutional conflicts of interest, when asked to describe the types of conflicts they encounter (see Figure 2-1). Lavezzari agreed with other speakers at the workshop that conflicts can arise from non-financial as well as financial relationships and that sometimes the former can be more powerful than the latter. 

8 COI AND MEDICAL INNOVATION Type of Potential Conflict Likelihood of Resulting in True Conflict Tangible financial conflicts of interest could occur if a person has a direct financial tie to the success of a product or research Intangible financial conflicts of interest could occur if a person has received payment for a service tied to the success of a product or research Personal conflicts of interest could occur if someone’s personal beliefs or goals are tied to the research Professional conflicts of interest could occur if a person could gain professionally from the results of a study Institutional conflicts of interest could occur if the interests of an institution or institutional official could affect the institutional processes for the design, conduct, reporting, review, or oversight of research 0 50 100 150 Weighted Score (n=27) FIGURE 2-1 Types of conflict and the likelihood of a true conflict, as reported by industry and academic researchers. SOURCE: Gabriela Lavezzari IOM workshop presentation, June 5, 2013. The analysis also identified five areas where conflict of interest poli- cies and perceptions have the greatest impact, Lavezzari reported (see Box 2-1). More than 75 percent of respondents said that conflict of interest policies had a negative impact on paid speeches, resources invested to monitor policies, time spent on complying with policies, paid consulting with industry, the ability to fill FDA advisory committees, and the time to initiate a collaboration. The survey also revealed that conflict of inter- est policies have had an impact on industry research activities, said Lavezzari (see Figure 2-2). One individual industry researcher who re- sponded to the survey said that conflict of interest policies have made it challenging for industry to collaborate with academic researchers who are also health care professionals and have authority to prescribe drugs and devices. When research collaborations with industry are reduced, that reduces research productivity, said an individual academic research- er. In addition, conflict of interest policies increase the amount of time spent complying with policies and the resources devoted to monitoring policies. A new industry of full-time employees who work on conflict of interest has been created, said an individual government researcher. It has become a costly and time-consuming specialty which also takes time away from those who have other primary duties. 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 9 BOX 2-1 Key Effects of Conflict of Interest Policies The key effects of conflict of interest policies and the perceptions of con- flicts of interest in the PhRMA study were summarized by Lavezzari as • Impaired collaboration. Industry and academic partners both agreed that conflict of interest policies increase the time needed to initiate collaborations. Even more worrisome, they reduce the ability and desire to start a collaboration. • Increased research burden. The resources spent monitoring and complying with policies increase the burden on research. • Limited FDA access to the best experts. Even as FDA increasingly focuses on more targeted therapies and personalized medicine, it has less access to the expertise needed to provide regulators with feedback about a drug. With some diseases, relatively few experts exist, and they tend to have relationships with industry. Experts al- so may be worried about being perceived as having a conflict of in- terest, or they may not have the time to serve on an advisory committee. A waiver process exists so that FDA still can work with experts who might have conflicts, but as therapies become more targeted, this problem will increase. • Decreased physician education. Conflict of interest policies can re- sult in less funding, fewer events, and less expertise for continuing medical education and physician education, with paid company speeches the area most affected by these policies. • Decreased communication. Having fewer company-sponsored talks or physician talks on behalf of companies can reduce overall levels of communication. Extreme instances of true conflicts of interest have generated atten- tion and have led to negative perceptions of industry relationships and to stricter policies. Part of the reason, Lo observed, is that there have been some prominent and well-publicized cases in which biased or inaccurate research findings resulted in harm to patients. These cases have involved drugs such as Vioxx®, Paxil®, and Avandia® as well as Infuse, a recombinant human bone morphogenetic protein, which has been used in spine fusions (DOJ, 2008; Krumholz et al., 2013; Palmer, 2010; Thomas and Schmidt, 2012). As a result, many conflict of interest policies tend to aim to eliminate conflicts by avoiding industry–academic relationships. A better approach, Lavezzari said, is to manage conflicts and promote shared access to scien- tific knowledge. Definitions of conflicts can be standardized and narrowed 

10 COI AND MEDICAL INNOVATION 100% Industry (n = 8) 75% Academic Respondents (n = 18-19) 50% 25% 0% Time to initiate Ability or desire to Ability or desire to Ability or desire to collaboration collaborate with collaborate with collaborate with government academia industry FIGURE 2-2 The impact of conflict of interest policies on industry and academic researchers. SOURCE: Gabriela Lavezzari IOM workshop presentation, June 5, 2013. to remove the perception that all relationships constitute conflicts. In this way, perceptions of conflicts can be addressed, best practices can be institut- ed, and beneficial relationships can be promoted, Lavezzari said. The underlying implicit assumption in most conflict of interest policies today is that more disclosure is equivalent to more transparency. Transpar- ency in turn is assumed to lead to less undue influence and bias, greater accountability and trustworthiness, and enhanced public trust. But well intentioned conflict of interest policies also can have unintended adverse consequences, Lo said. Disclosures made in different contexts—such as writing a grant, submitting an abstract to a professional meeting, submit- ting a manuscript to a journal, or serving on an advisory committee—can be inconsistent. Such variations in reporting relationships can cause peo- ple to infer that misbehavior has occurred, whereas the inconsistencies may arise instead from inconsistent disclosure requirements, different time periods, variations in what must be disclosed, or different thresholds for the amounts that need to be disclosed. The risk of inconsistencies leading to allegations of misconduct also could deter physicians or re- searchers from collaborating on valuable activities. Standardized disclo- sure formats could prevent such inconsistencies, and such formats have been discussed intensively in recent years, Lo said. 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 11 Conflict of interest policies have been criticized as an overreaction to egregious cases of research misconduct or other irresponsible research practices, Lo said. From this perspective, such policies cause potential collaborations to be viewed with suspicion when, in fact, they are trying to move research forward. However, other critics argue that the greatest fault of conflict of interest policies is that they are weak, inconsistent, and inadequately enforced. Other unintended adverse consequences in- clude administrative burdens and opportunity costs. Potential collabora- tors may decline to participate because of the amount of effort required before the collaboration begins. Conflict of Interest in Medical Research, Education, and Practice (IOM, 2009) emphasized the evaluation of conflict of interest policies. It called for empirical research to be conducted and published in peer- reviewed journals so that future policies can be based on evidence, not just on assertions or conjecture. Conflict of interest policies need to be risk-based, with a focus on the greatest risks to the integrity and rigor of research, Lo said. They also need to be benefit-based, so that they reflect the potential benefits of developing new knowledge and treatments, he said. Finally, the policies need to be evidence-based through the identifi- cation of relationships that facilitate innovation or undermine the integri- ty and rigor of research, Lo said. Disclosure then needs to be performed in a way that facilitates the identification of high-benefit/low-risk rela- tionships and high-risk/low-benefit relationships, he said. ENHANCING THE ACADEMIA–INDUSTRY RELATIONSHIP The academia–industry relationship is complex, said Neal Cohen, professor of anesthesia and perioperative care and medicine and vice dean of the School of Medicine at the University of California, San Fran- cisco (UCSF). It encompasses not just academic and industry researchers but also administrators, clinicians, patients, and their families. Each of these groups has its own goals, challenges, and conflicts. “We need to think very carefully about how we build these relationships,” he said. Clinicians care for patients, which can create conflicts involving how they are financially compensated for their expertise, how they manage their patients, how their institutions hold them accountable for providing cost-effective care, and how they relate to their scientific colleagues, graduate students, postdoctoral fellows, and those in industry, said Cohen. In addition, faculty members seek to protect not only their academic freedom but also their autonomy and their control over their research, clin- 

12 COI AND MEDICAL INNOVATION ical, and administrative environments. It is a “challenging and competitive environment” that requires creative solutions beyond the traditional ones put in place in recent years, he said. Academic faculty members get mixed messages about their missions from their institutions, the public, regulators, friends, and family, Cohen observed. Faculty are supposed to create and disseminate new knowledge, secure extramural research supports with indirect cost recov- ery, commercialize advances, create new businesses, and use new tech- nologies and drugs to benefit patients. Despite this multiplicity of goals, faculty are also expected to avoid conflicts of interest and conflicts of commitment, protect intellectual property, ensure academic freedom, and be transparent with their findings. In this last area, they are told to report on all activities for which there is a potential, real, or perceived conflict. But in many cases, Cohen said, discerning whether there is a conflict is difficult because definitions of potential conflicts vary among individuals. “How do I disclose what I don’t recognize as a conflict? This is a big issue.” Lo and Cohen agreed that disclosure requirements can be challeng- ing and confusing. Disclosures can be different for administrators, re- searchers, and clinicians. In most universities, faculty are encouraged to engage in activities outside the university, and these external relation- ships create additional complications for disclosures, Cohen said. Fur- thermore, disclosures or failures to make a disclosure have to be interpreted, which also can be challenging. All stakeholders—including government agencies, commercial pay- ers, and health plans—have their own unique goals and conflicts that go along with them when they build relationships with other groups. Aca- demic institutions have conflicts borne of their multiple missions, even though they are often held less accountable than individuals, Cohen said. They are supposed to be educating students while also conducting re- search, which can cause the teaching mission to get lost. Institutions seek to protect and advance their standing on such measures as funding from the National Institutes of Health (NIH) or the U.S. News & World Report rankings of college and universities. They aim to protect intellectual property and benefit from royalty and licensing payments while also cre- ating and disseminating new knowledge. The collaborative research enterprise for institutions has expanded greatly over the past two decades, according to Todd Sherer, associate vice president for research and executive director of the Office of Tech- nology Transfer at Emory University. Reported research funding has in- 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 13 creased each year, rising from $12 billion in 1991 to $61 billion in the fiscal year (FY) 2011 survey.1 The number of new invention disclosures filed by faculty has closely tracked that increase, rising from about 5,000 to nearly 22,000 during that same time period. In 2011 alone, almost 600 new products were introduced to the market, and more than 600 new startups were formed. Almost 4,000 startups were believed to be opera- tional in 2011, employing an estimated 55,000 people. In FY 2011, $2.5 billion in licensing revenue was reported from an estimated $80 billion to $120 billion in product sales. The relationship between industry and cli- nicians can be extensive and substantial. There are clear financial drivers for institutions to promote innovative research, Sherer said. If the structure of an institution includes a health care delivery com- ponent, it will have commitments to patients, to the families of patients, and to the larger community. Institutions also need to raise funds from philanthropists and others, which may encompass such issues as naming rights. For example, Genentech Hall is the name of one of the buildings on the UCSF campus, but some have objected to the presence of a bio- technology company’s name on an academic campus. However, the company named the building and provided $50 million toward funding the research facility as part of a settlement over a patent dispute for the drug, Protropin (Barinaga, 1999). “Does that create conflicts?” asked Cohen. “Are those [conflicts] real or perceived?” Conflicts of interest also arise from the goals of those in industry. In- dustry would like to control intellectual property, reduce the high cost of product development and regulatory burdens, and balance the needs of its own researchers for autonomy with the need to commercialize products, Cohen said. Some people believe that the only way to manage these conflicts is to avoid developing relationships between industry and academics. But Cohen said that this perspective represents a very narrow view of the complex relationship between physicians and patients. Would this per- spective mean that physicians should discourage their patients from par- ticipating in clinical trials? The fact that many patients seek out physicians who consult with industry for their knowledge and experience should also be considered before calling for the elimination of relation- ships between academia and industry, Cohen said. A workshop partici- 1 Association of University Technology Managers U.S. Licensing Activity Survey Highlights, http://www.autm.net/AM/Template.cfm?Section=FY_2011_Licensing_Activity Survey&Template=/CM/ContentDisplay.cfm&ContentID=8731 (accessed December 6, 2013). 

14 COI AND MEDICAL INNOVATION pant agreed that there should not be a complete separation of industry and academia but suggested that some separation is needed—for exam- ple, that pharmaceutical marketing should be separated from medical education. Cohen suggested several steps that could enhance academic–industry interactions, minimize risks, and create the right balance between collab- oration and competition. Conflicts need to be defined from the perspec- tive of each party, he said, and then they need to be managed so as to protect human subjects and improve medical treatment as much as possi- ble. Transparency is essential, and transparency requires a mechanism to disclose conflicts. While economic relationships are important and easier to measure, other non-financial relationships also need to be understood, he said. Sharon Terry agreed, noting that biases in research could poten- tially lead a field astray. These other types of non-financial biases need to be taken into account when thinking about their impact on outcomes in research, said Heather Pierce, senior director of science policy and regu- latory counsel for the Association of American Medical Colleges (AAMC). Creating a single template for disclosure using standard definitions is important, but it will be difficult, Cohen admitted, because the purposes of disclosures are very different depending on the circumstance of the relationship. Disclosures cover different topics, concerns, time periods, and obligations. Furthermore, conflicts may be unintended or unrecog- nized. In addition, disclosures need to be available to all the parties that need to know, including patients. If a clinician has invented a hip pros- thesis and is collecting profits from its use and then informs his or her patients that this is the best device to use, many patients would trust that the physician was providing expert advice and would thus elect to use the prosthesis. But the key question is whether the prosthesis is the correct one to use for that particular patient; a different surgeon might recom- mend a different device. What would be the non-conflicting recommen- dation? “These are difficult decisions with respect to not only disclosing but interpreting the disclosure in a way that allows patients, families, and the community at large to make the right decisions,” Cohen said. Addressing bias more explicitly and proactively could address the problem of undue influence, Lo said. Having more robust peer review processes before clinical trials start would produce better research, as would having more explicit standards, guidelines, and checklists for pre- clinical and clinical studies. The comparative arm should be a reasonable comparison, not just the weakest result in the control arm so that the ex- 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 15 perimental arm looks better. Flawed comparisons can make a drug under study look better, but they do not help patients. “Some of the attention we’re now spending just on the disclosure of relationships might be bet- ter applied to looking at how to characterize and to remove undue influ- ence and bias,” Lo said. Clarifying and perhaps merging the roles and responsibility of con- flict of interest advisory groups and institutional review boards (IRBs) will provide these groups with the expertise they need to conduct a scien- tific evaluation and to understand where the conflicts are and how those conflicts are managed, Cohen said. “Right now they’re done in parallel. We need to figure out a way to integrate those more effectively.” Indi- viduals who serve on both an IRB and a conflict of interest committee would be helpful for informing policy, said Guy Chisolm, director of the Innovation Management and Conflict of Interest Program at Cleveland Clinic. “This has been a tremendous benefit [for] understanding our- selves as an institution and the possible contradictions or hold-ups [for] investigators,” he said. The reporting requirements of Section 6002 of the Affordable Care Act, often referred to as the Physician Payments Sunshine Act2 that will go into effect in 2014 have been designed to help patients make informed decisions and to discourage financial relationships that inflate health care costs. Beginning August 1, 2013, manufacturers are to collect and track data on their financial relationships with physicians and teaching hospi- tals. These data must be reported annually to the Centers for Medicare & Medicaid Services (CMS) beginning March 31, 2014, and CMS will make the data available to the public on a website that will launch by September 30, 2014. The goal of the Physician Payments Sunshine Act is to promote transparency about relationships by providing consumers with infor- mation to make better informed decisions about their health care. But patients are focused on getting good care rather than reducing health care costs, and Cohen questioned whether disclosing financial relationships to patients could affect costs. On the other hand, he said, it is important to have a disclosure process despite the cost of compliance, which has been estimated to be $1 billion over 5 years (Rosenthal and Mello, 2013). Even though disclosure does not necessarily change behaviors, it provides useful 2 Centers for Medicare & Medicaid Services, Official Website for Open Payments (Physician Payments Sunshine Act): http://www.cms.gov/Regulations-and-Guidance/Legislation/National- Physician-Payment-Transparency-Program/index.html (accessed November 8, 2013). 

16 COI AND MEDICAL INNOVATION information. The information, however, has to be put in context because reporting alone can misrepresent the nature and the value of a relationship. Cohen said that medical schools could do a better job of educating students about the value of relationships with industry and the challenges associated with those relationships. “Until we go back to the basics and talk about what we’re trying to accomplish in thinking about industry– academia relationships, and how to most effectively accomplish them, we won’t have done our duty,” he said. Creative forms of collaboration can break down the silos between industry and academia while avoiding some of the issues associated with traditional collaborations, Cohen continued. One such example is the Medical Device Innovation Consortium developed by FDA, which facili- tates collaborations among academia, industry, and federal partners, said Michelle McMurry-Heath, associate director for science at FDA’s Center for Devices and Radiological Health. The Medical Device Innovation Consortium allows for the pooling of financial resources, intellectual property, and ideas so that members can work together to solve regulato- ry science needs. The idea behind the consortium was to get the hard ne- gotiations out of the way at the institutional level so that the science could be the focus of later work. The “foundation is laid, so now when we have additional research that we do jointly, it should be a smoother path,” McMurry-Heath said. Open innovation research and data networks, screening facilities, compound libraries, and sharing of personnel, can also benefit partners in industry–academia networks, Cohen said (IOM, 2013). Autonomous or- ganizations within academia, such as incubators, can allow for open communication and innovative roles for graduate students, postdoctoral fellows, and faculty members. Such arrangements need to include master agreements, oversight structures, arrangements for distributing credit for discoveries, and agreements for the sharing of financial risks and bene- fits. For example, one model is to have a strategic planning board that defines goals and identifies potential collaborative partnerships and op- portunities; a coordinating committee that identifies and leverages exper- tise and manages databases; and an advisory board that evaluates strategies, provides oversight, and manages conflicts of interest. As an example of such an arrangement, Cohen cited the UCSF Cen- ter for Digital Health Innovation (CDHI), which leverages the institu- tion’s strengths in clinical care, health sciences discovery, education, and innovation by using industry expertise in development, manufacturing, and commercialization. The focus is on developing standards-based, 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 17 open platforms for data exchange and on incubating CDHI startups to the point where they are ready for commercialization. For example, the Health eHeart Study is a clinical trial and therapeutic management plat- form based on social media that allows patients and providers to ex- change information. Patient- and device-recorded data and surveys, genetic and blood analysis with specimen banking, and device integra- tion allow for improved health care delivery and the rapid creation of large cohorts for registries and clinical trials. All of the study data must be secured according to the guidelines of the Health Insurance Portability and Accountability Act of 1996.3 The result is that the study, Cohen said, is similar to the Framingham Heart Study, except that it is able to be car- ried out on a global scale because of the digital nature of the data collec- tion and analysis. The key to enhancing product development and facilitating academia– industry relationships, Cohen said, is to maintain a commitment to core values: “core values of the institution, core values of the faculty, core val- ues of industry, and core values of the community.” Identifying and man- aging conflicts can ensure transparency and allow individuals to determine whether conflicts might influence decisions while maintaining the integrity of science and the public trust. NIH POLICY ON CONFLICTS OF INTEREST Government regulations are not intended to stifle academia–industry relationships, said Sally Rockey, deputy director for extramural research at NIH. Rather, they are meant to enhance those relationships, support the objectivity of research, and maintain public trust through the appro- priate management of conflicts. NIH is “a public agency, and we want the public to trust the research that we’re generating,” she said. Not all financial interests are conflicts, but conflicts are inevitable, so they need to be properly managed, Rockey said, agreeing with other speakers. NIH does not seek to manage the financial interests of investi- gators. Its relationship is with the institution, and it assigns responsibility for dealing with conflicts largely to institutions. It is impractical for NIH to 3 The Health Insurance Portability and Accountability Act (HIPAA) regulates the pri- vacy of individual health information and sets security standards to protect electronic health information. Health and Human Services, Summary of the HIPAA Security Rule: http://www.hhs.gov/ocr/privacy/hipaa/understanding/srsummary.html (accessed January 13, 2014). 

18 COI AND MEDICAL INNOVATION collect information on the financial interests of every single investigator, and it does not perform this role, Rockey said. Conflict of interest policies in the U.S. Public Health Service (PHS), which includes NIH, were promulgated in 1995 and revised in 2011 (HHS, 2011). NIH worked closely with investigators and institutions to arrive at the final, revised regulations. When NIH began talking about revising the conflict of interest regulations, the number of financial con- flict of interest reports submitted to NIH increased dramatically, Rockey said. Relationships were developing, but a large number had not been previously disclosed. The new regulations establish standards that provide a reasonable expectation that the design, conduct, support, and reporting of research funded under PHS as grants or cooperative agreements will be free from bias resulting from investigator financial conflicts of interest, Rockey said. The term “reasonable expectation” signals that conflicts will arise but that they will be managed to produce objective research. Investiga- tors disclose to the university or to the institution all significant financial interests, including any such interests that pertain to their institutional responsibilities. It is a preventive and proactive strategy, Rockey said. Institutions need to have and implement a policy, Rockey stated. They must train their investigators and others involved in the research. They must look at financial interests and determine whether the interests constitute a financial conflict of interest. If conflicts occur, they must be managed. Some institutions have been very strict and restrict relationships with industry, while others lean more toward management. The PHS regula- tion states that institutions must manage conflicts, but it does not specify how this must be done, said Rockey. If a conflict is identified, the institu- tion must report that conflict to NIH and tell the agency how the conflict is being managed. NIH oversees the institutional policies, but it is not an auditing group. It does some targeted and proactive compliance, but its main job is to oversee the process. With the new requirement that institu- tions report how they are managing a conflict, NIH can determine whether specific trigger points have been reached and whether it needs to work with an institution. At times the government’s policies can seem somewhat confusing, Rockey acknowledged. The Bayh-Dole Act4 encourages academia– industry partnerships, while conflict of interest regulations constrain 4 The University and Small Business Patent Procedures (Bayh-Dole) Act of 1980. Pub- lic Law 96-517, 96th Congress. December 12, 1980. 94 Stat. 3015. 

CONFLICT OF INTEREST POLICIES: AN OVERVIEW 19 those relationships. The same balancing process is evident in NIH’s new National Center for Advancing Translational Sciences (NCATS), which was created to catalyze the generation of innovative methods and tech- nologies that will enhance the development, testing, and implementation of diagnostics and therapeutics. For example, a centerpiece of the NCATS program has been the Discovering New Therapeutic Uses for Existing Molecules program, through which eight companies have made available to academic researchers their libraries of molecules and com- pounds that have gone through preclinical trials but have not progressed further. The program had model template agreements for intellectual property, but conflicts of interest have been an issue. Again, Rockey said, NIH’s approach has been not to eliminate conflicts but to manage them. In fact, by managing the conflicts, investigators and companies have become more interested in participating in the program. The con- flict of interest policies “gave people much more peace of mind that we were going to be able to approach this new program and manage whatev- er conflicts of interest might arise in appropriate ways,” said Rockey. International Relationships and Market Forces Relationships with industry are much less regulated in other coun- tries. For example, in the Philippines physicians attend continuing educa- tion programs with shopping bags to fill up with samples from drug companies, Cohen said. They do it for the most part because they want to give their patients access to drugs that they otherwise could not afford, but their prescription behaviors are clearly dependent on those relation- ships. “In many of the countries that I’ve visited, these issues aren’t be- ing addressed at all,” Cohen said. Yet many future clinical advances will require studying groups of patients much larger than the numbers availa- ble in the United States. “We have to look to Asia for those relationships, and then we start dealing with some of the regulatory requirements and restrictions of various companies and the ability to control the intellectu- al property and potentially be able to build markets. It’s incredibly chal- lenging,” Cohen said. International relationships are made even more complicated because many sub-recipients of awards are located in coun- tries without rigorous conflict of interest policies. These U.S. researchers are therefore responsible for managing the conflicts of interest and mak- ing sure those subcontractors or sub-awardees have conflict of interest policies. This creates a need for universities and other institutions to try to figure out what other countries are doing. 

20 COI AND MEDICAL INNOVATION Terry also pointed out that the practice of medicine occurs in a mar- ket economy, which means that treatments have to be sustainable. Drugs are available today because of commercial interests, and the market al- lows access to needed drugs. Individuals need to understand that, and they need to be empowered to engage in their own health care to better understand what decisions might be affected by conflicts of interest. But disclosure policies are not necessarily targeted to the individual patient, Lo said. Most patients will not look at whatever disclosure information is available, but researchers, investigative journalists, policy makers, and others will look for patterns that apply more broadly. Where institutions implement quality improvement measures to improve patient outcomes, “that’s the kind of value we’re most likely to see,” Lo said.