|
FROM THE FIELDImplementation Of Financial Disclosure Policies To Manage Conflicts Of Interest
Amid concerns that clinical and basic research is increasingly vulnerable to pressure from industry sponsors, many federal agencies and professional organizations are recommending new or revised policies regarding financial disclosure and conflicts of interest. Despite pressure for greater external regulation and common professional standards, little is known about how existing policies are implemented. We analyze the implementation of conflict-of-interest policies within the multicampus University of California system. We show that there is variation among campuses in defining problematic relationships and in determining appropriate remedies to mitigate conflicts of interest. Our study suggests the importance of local culture and context for institutional decision making.
Since 1995 the Public Health Service (PHS) has attempted to regulate certain research relationships by requiring academic investigators who receive federal grants to disclose their personal financial relationships with for-profit companies. In particular, recipients of National Institutes of Health (NIH) or National Science Foundation (NSF) funding are required to disclose to their institutions annual income in excess of $10,000 or equity ownership exceeding 5 percent in a company whose "financial interests would reasonably appear to be affected by the research."1 This policy (as well as corresponding state and institutional policies) was designed to protect the integrity of scientific research and participating human research subjects by making possible conflicts of interest among investigators known to their respective institutions and federal funding agencies. As industry sponsorship of medical and scientific research grows, some federal agencies and professional organizations recommend new or revised policies regarding financial disclosure and the management of financial conflicts of interest. For instance, the Association of American Medical Colleges (AAMC) has announced recommendations for oversight of both individual and institutional financial conflicts of interest in research involving human subjects.2 In January 2001 the PHS released for public commentary a Draft Interim Guidance document that clarifies the role of Institutional Review Boards (IRBs) in reviewing financial disclosures for possible conflicts of interest in human-subjects research.3 In 2002 the Institute of Medicine (IOM) released a set of recommendations to institutions, including the need for transparency in conflicts of interest and institutional management of disclosed competing interests.4 Also in 2002, U.S. senators Edward Kennedy and William Frist convened congressional hearings to address the protection of human research subjects and conflicts of interest.5 These recent recommendations would supplement—and strengthen—existing PHS regulations. Despite this pressure for greater oversight, we know very little about how existing policies are implemented, the processes in place to monitor and regulate competing financial interests, or their effectiveness.6 The little empirical evidence that does exist suggests considerable variability in both policy and process.7 This paper addresses these gaps in our understanding.
Existing research. The concerns motivating current conflict-of-interest policy changes are important. More than one billion industry dollars flow into academic institutions each year; it has been estimated that almost one-third of life-sciences faculty receive funding from industry sponsors.8 There is growing concern that industry sponsorship of research may influence the outcomes of research and undermine intellectual freedom, open exchange of ideas, and research for the public good.9 Several studies are germane to these concerns.10 Further, a large number of faculty investigators now report having personal financial ties to their industry sponsors as well.11 Although there is speculation that these ties result in an increased potential for conflicts of interest, there are few empirical data showing the nature and extent of these ties and their consequences for research outcomes.12 In addition, there remains a need for systematic research that examines not only the substance of existing conflict-of-interest policies, but also the processes by which they are implemented by administrators and faculty committees. The UC system. In this paper we examine the implementation of federal, state, and local conflict-of-interest policies within the University of California (UC) system. The UC system offers a unique opportunity to examine the full range and nature of academic–industry relationships (for both clinical and basic research), faculty financial interests, and institutional strategies for managing conflicts of interest. These factors are available for study because of state legislation requiring faculty researchers at public universities to disclose to their institutions personal financial ties with private entities, in accordance with the California Political Reform Act of 1974.13 As state employees, faculty investigators must disclose annual income over $250, $1,000 in equity holdings, or a management or decision-making position within a company sponsoring a research project.14 The result is that UC investigators not only must adhere to federal guidelines, but also must disclose their personal financial interests when private companies sponsor their research. In addition, some campuses have local disclosure requirements and conflict-of-interest policies. Thus, UC investigators are obligated to disclose virtually all of their research-related financial interests above $250. None of the policies in effect defines conflict of interest explicitly or prescribes specific strategies for managing conflicts. All specify a financial threshold for disclosure. Some policies suggest general management practices, but committees at each institution decide whether a conflict exists and, if so, how it should be managed.
Study sites. We used a combination of data sources and qualitative and quantitative analyses to develop a set of comparative case studies of seven UC campuses. Of the nine UC campuses in the system at the time of the study (Berkeley, Davis, Irvine, Los Angeles, Riverside, San Diego, San Francisco, Santa Barbara, and Santa Cruz), one campus did not maintain a standing conflict-of-interest committee or archival records, and one campus declined to participate. The UC system represents a range of possible variations in campus size (ranging from 13,150 students and 672 faculty at Santa Cruz to 37,500 students and 3,238 faculty at Los Angeles), location, research orientation, and reputed institutional climate.15 Overall, UC faculty conduct a large amount of research for which financial disclosures are required. In fiscal year 2001, for example, the UC campuses received $1.01 billion in NIH funding and $578 million in private research funding, and five campuses rank among the top twenty-five institutions nationwide in NSF funding.16 Since 1997 each UC campus has been required by the Office of the President to appoint a committee to review new financial disclosures and to advise the vice chancellor of research of potential conflicts of interest.17 Although the committee’s role at each campus is explicitly advisory, in almost all instances, the vice chancellor implements the recommendations of the committee. Each conflict-of-interest committee comprises faculty members and administrators and may include a member of the Contracts and Grants, Legal Affairs, or Technology Transfer Offices, or the IRB for human or animal subjects research. One campus includes two members chosen from outside the campus community. Data. The research activity of faculty investigators produces a large number of annual financial disclosures. Whenever a faculty member reports a personal financial relationship with a sponsoring or related company, that disclosure is said to be "positive." Our data sources included the archived records of positive financial disclosures maintained by the Office of Research Administration (or comparable unit) at each of the campuses for January 1996–June 2001; in-depth interviews with conflict-of-interest administrators and committee members at each campus; and the publicly available Web sites maintained for faculty investigators at each campus. Archived records were reviewed in paper form; interviews were conducted in person and by telephone. From the archived records, we abstracted a set of variables for each disclosure into a database. The Committee on Human Research at UC San Francisco and the IRB at each individual campus approved the study.
Process of disclosure and review. Across the seven campuses we studied, the process of disclosure is identical: Investigators must disclose related financial interests upon submission of a sponsored-project grant application. Also consistently across campuses, the conflict-of-interest administrative responsibilities are housed within the Office of Research Administration and, frequently, within the Sponsored Projects Office. The practical consequence of this is—at least formally (and increasingly in practice)—interaction and cooperation between internal units of the Office of Research Administration. Efforts to coordinate these different units are ongoing at all seven campuses. Although the layout of the 730-U (state) and federal disclosure forms varies across campuses, the content is identical at all campuses. However, there is considerable variation in the information solicited by supplemental forms. These "addenda" may include detailed questions related to consulting agreements, licensing agreements, and how research and academic work will be kept separate. So across campuses, committees have variable information available to them as they make their evaluations. For example, Campuses 4, 5, and 7 require investigators to complete a lengthy addendum that solicits detailed information regarding the research project, the nature of the financial relationship, the specifics of the compensated activities, the use of university facilities, company prospectuses, lists of boards of directors, and so forth. By contrast, Campuses 1, 3, and 6 collect much less detailed information. We consider the possible implications of this for the management of potential conflicts of interest below. As mentioned above, committee structure varies across campuses, primarily in size and individual membership, but all committees go through a similar process of reviewing disclosure forms, addenda, and other materials collected by the individual campus. The campuses have implemented nearly identical procedures for disclosure and review. However, we found considerable variation in the nature and extent of faculty relationships, the types of relationships deemed problematic, the consequences of local policies, and the management strategies recommended. Extent of faculty relationships. Between January 1996 and June 2001, 1,991 positive financial disclosures were reviewed at the seven campuses. The number of positive disclosures at each campus varied widely, from a low of 15 to a high of 560. Because of record-keeping inconsistencies for negative disclosures, we were unable to calculate the proportion of positive disclosures for all seven campuses. For the four medical campuses, the proportion of positive disclosures related to clinical trials (versus other types of research) ranged from 3.4 percent at Campus 6 (7 out of 207 disclosures) to 29.6 percent at Campus 5 (98 out of 331). Nature of financial relationships. On a single disclosure form, a faculty member may disclose one or several types of relationships with the sponsoring or related company. Nearly one-third of companies were pharmaceutical companies, 27.5 percent were biotechnology companies, 8.3 percent were educational or nonprofit organizations, 7.9 percent were electronics companies, and 5 percent were computer companies. Between 5.8 and 43.9 percent of disclosures involved the investigator reporting multiple activities with the sponsor or related company; these ranged from 5.8 percent to 36.1 percent of all disclosures. More than half of these multiple relationships (68 percent) involved income below the federal disclosure limit of $10,000 a year. Across the seven campuses, payment for consulting activities accounted for 54.4 percent of the financial disclosures; equity holdings, 37.7 percent; payment for talks, 14.4 percent; Scientific Advisory Board membership, 12.7 percent; membership on a company’s board of directors, 11.9 percent; and company founder, 7.3 percent. Management of conflicts of interest. After receiving a positive financial disclosure, the conflict-of-interest committee at each campus decides whether the financial relationship represents a conflict of interest and, if so, how to mitigate or eliminate the conflict. State (and UC) guidelines suggest that committees determine whether an existing financial relationships would interfere with an open academic environment, free dissemination of ideas, intellectual freedom for students, appropriate use of university facilities and resources, and the fair licensing of new technologies. Federal policy requires the committee to determine whether the disclosed financial relationship would have a "direct and significant" effect on the design, conduct, or reporting of the research. Overall, the committees determined that 26 percent of the reviewed cases involved a conflict of interest in need of management. The percentage of cases identified as conflicts of interest and managed varied across campuses. Campus 3 recommended some form of management in less than 10 percent of its cases; Campus 4 recommended management in almost half of its cases. In less than 2 percent of the cases, a conflict-of-interest committee would decide that a conflict was completely unmanageable and would recommend that the funding not be accepted. The written records, including letters to investigators and to the vice chancellor of research, indicate that committees typically examined the nature of the proposed scientific work (that is, basic or applied), the overlap between the paid activities and sponsored research, the length of relationship between the company and the investigator, the presence of graduate students or postdoctoral fellows on the project, the details of the consulting agreement, and the degree to which the investigator could be seen to be independent of the company’s interests (that is, whether there was the "appearance" of a conflict). The management strategies selected by the committees ranged from disclosure of the financial relationship in publications and public presentations, a reduction in equity holdings, eliminating consulting activities, resigning as principal investigator or agreeing to third-party oversight and annual review. The three most commonly applied management strategies were requiring disclosure in publications and presentations (40 percent of managed cases recommended this strategy), appointing an oversight committee to protect the interests of graduate students and postdocs involved in the project (21 percent), and eliminating the existing relationship during the project (22 percent). While aggregate figures provide an overall sense of committees’ tendencies, there was, in fact, much variation among the campuses in the identification of conflicts of interest and the management strategies deemed most appropriate. For this comparison, we selected cases from each campus that shared certain characteristics (for instance, a disclosure involving a single consulting relationship for which the investigator earned less than $10,000 annually).
Across each of the examples, it is clear that the campuses employ different definitions of conflict of interest (Exhibit 1
As the examples in Exhibit 1  suggest, even among cases that are superficially similar, no two cases are identical, and committees consider each case individually and base management decisions on a range of factors. Even within a single campus, a case with the "same" characteristics may be defined and managed differently, as the relative weights of different characteristics may vary for any given case. Interviews with administrators and committee members support this assessment. Committee members and administrators recognize the ad hoc nature of their work and consistently offer this as a critique of legislative or other efforts to standardize the decision-making process. Instead, a case-by-case approach is advocated as the best way to protect against arbitrary decisions. Nevertheless, each committee has developed a coherent framework within which it evaluates cases of financial disclosure and makes management recommendations, even in the absence of clear definitions or standardized decision protocols.
All seven campuses in our study have nearly identical processes for collecting and reviewing faculty financial disclosures, yet each campus employs different standards for determining the existence of conflicts of interest and recommendations to mitigate or eliminate them. A complex range of factors account for these differences; we discuss several below. Supplemental forms. First, the campuses use different supplemental forms to gather additional information when investigators disclose a financial interest. When the committees met to discuss individual cases, then, important details might not have been discovered at some campuses. For instance, knowing the scientific direction of a company (from its prospectus or Web page) may allow the committee to detect more subtle overlap between an investigator’s research and his or her compensated activities. This is particularly important in determining the degree of separation between consulting and research activities. Thus, the relatively low rates of management at Campuses 1, 3, and 6 could, in part, be attributable to the relative lack of information that the committee receives.18 Campus policies. A second factor that obviously contributes to committees’ interpretation and decision making is local (campus-level) policies. For instance, one campus established a formal policy prohibiting any compensated activities or management position in a company sponsoring a clinical trial during the course of that trial. A second campus adopted a de facto policy prohibiting all but the most minimal compensation (less than $250) during a clinical trial. A third adopted a formal policy requiring clinical investigators to provide written justification of the use of an open-label design (nonrandomized, not blinded) for a clinical trial. These policies immediately affect the interpretation of disclosures that come before the respective committees and provide a partial explanation for the high rate of management at campuses where these policies exist. Relationships among faculty investigators. In addition to these procedural differences, the committees also see different relationships among their faculty investigators. For instance, nearly half of the disclosures at Campuses 4 and 7 involve investigators with multiple relationships to industry. These committees, then, consider cases in which investigators may consult, hold equity, sit on a scientific advisory board, or be a founder of a sponsoring or related company. The relatively high rates of management may be partly responsive to the complexity of the cases seen. Less than 10 percent of the cases seen at Campuses 3 and 6 involve multiple relationships, and they manage a much smaller percentage of their cases. In general, the less complex the disclosed relationship, the less likely any campus committee is to see it as problematic. To speculate, if we take complexity of the relationships as a proxy for entrepreneurial activity of faculty investigators, the more entrepreneurial the investigators are, the higher degree of control the committees appear to exhibit (that is, the percentage of managed cases), which suggests that greater control of research activities may not necessarily inhibit the entrepreneurial activities of faculty. Institutional culture. Finally, regardless of the type of disclosed relationships or their complexity, the management recommendations made by each campus reflect a local institutional culture as well as a shared orientation to certain broader concerns. Although campuses vary in their application of management strategies, there is consistency in the rationales used to explain the committees’ choices. Thus, Campuses 2, 4, and 5 consistently appoint oversight committees to protect the interests of graduate students and postdocs working on projects in which the principal investigator has a financial interest in the sponsor. Campuses 4, 5, and 7 consistently treat clinical trials with extra care, requiring higher standards of conduct for clinical investigators. Campuses 4 and 7 frequently use public disclosure to mitigate conflicts of interest, suggesting sensitivity to transparency and public perception. All of the campuses are concerned with protecting the financial interests of the university, requiring investigators to modify any consulting or other agreement with a sponsor that inappropriately forfeits university rights to patents, licensing, and technology or restricts publication of results. Little is known about the effects of different management strategies on the research enterprise or public opinions about research integrity. Study limitations. One limitation to our study is its focus on the UC campuses. However, our findings are unlikely to be unique to these particular institutions, except insofar as UC campuses are subject to California disclosure laws. We have no reason to suspect that UC campus committees are fundamentally different from any other conflict-of-interest committees, and we anticipate that these results will prove generalizable to other institutions. As noted above, similar variation in policy implementation across divisions or offices has been observed in a range of institutional settings.
This study raises several important questions for proposed recommendations to strengthen existing conflict-of-interest policies or implement consistent standards. How effective are current policies in detecting conflicts of interest? Our study suggests that the current federal policy, when combined with the state policy, is adequate for detecting investigators’ financial relationships with industry sponsors of their research. In California the addition of the state requirement means that any interest more than $250 must be disclosed. Researchers sign state and federal disclosures under penalty of law, so we have no a priori reason to suspect serious underreporting of financial interests. Furthermore, increasing coordination between offices in research administration (such as Human Subjects and Contracts and Grants) reduces the chance for underreporting. Assuming, then, that investigators do disclose their relevant financial interests, the California policy ensures that all but the most minimal relationships will be disclosed. In contrast, under the federal policy alone, at least two-thirds of the disclosures included in this study would not have been reviewed by the committee—68 percent of the disclosures involved income under $10,000 per year, and very few involved equity ownership over the federal 5 percent threshold. While neither the state nor the federal policy actually provides a definition of a conflict of interest, the federal policy sets a much higher threshold for disclosure, which, by inference, suggests that income under $10,000 or 5 percent equity does not constitute a conflict.19 The committees we studied often enforced a stricter standard in practice. Lesser income amounts, particularly when coupled with other relationships, were viewed as contributing to the possibility of a conflict of interest and in need of management. Thus, the practical definition of conflict of interest employed by the different committees did not rely solely on threshold amounts; rather, relationships were assessed in context for their potential for a conflict of interest. Is variation in policy implementation a concern? In the absence of a clear and consistent definition of conflict of interest, individual committees have developed their own sets of standards in evaluating financial disclosures. Those standards appear to be based on specific institutional values that the committees felt were important to protect. Thus, as noted above, some campuses put the protection of students at the forefront of their standards; others evaluated clinical trials according to a different standard. Similar cases were interpreted and managed differently, depending in part on the orientation of the committee to these values or standards as well as the specific details of each case. It is important to note that all of these values are considered commonly held within academic communities.20 Whether variation in implementation is viewed by some policymakers as problematic may depend, in part, on the goals of the policies themselves. One obvious goal is to protect human research subjects and to encourage scientific integrity by making investigators’ personal financial interests known to the university community. If this is the only goal, then variation in management across campuses may be the expected outcome of a case-by-case decision-making process. However, if another goal is to define and implement consistent national, statewide, or professional standards of research behavior or, ultimately, to develop standardized decision-making protocols (or other procedures for identifying and managing conflicts of interest), then this variation might be of concern. State and federal policies now leave individual campuses to devise their own standards. In our study, each campus committee was committed to upholding a set of standards that addressed the potential for conflict while at the same time remained sensitive to the context of individual case details. As legislators and policymakers revisit conflict of interest, the important role that local context and institutional values play in the determination of appropriate research relationships must be considered with care.
This research was supported by the Research Integrity Program, Office of Research Integrity/National Institutes of Health collaboration, Grant no. RO1NS42398-01. The authors thank Mildred Cho, Bonnie Glaser, and Andrew Roth for their comments on an earlier draft and the conflict-of-interest committees and administrators at each participating campus for their time and encouragement of this project. Elizabeth Boyd is an assistant adjunct professor at the University of California, San Francisco (UCSF), Department of Clinical Pharmacy, and an assistant professor at Keck Graduate Institute of Applied Life Sciences in Claremont. Shira Lipton is a medical student at the University of California, Los Angeles. Lisa Bero (bero{at}medicine.ucsf.edu) is a professor in the UCSF Department of Clinical Pharmacy and Institute for Health Policy Studies.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Background
Study Methods
