Conflict of Interest in Medical Research, Education, and Practice (2009)

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.

F Model for Broader Disclosure This appendix has two parts. The first is a proposal by three committee members for a model for broader disclosure of financial relationships and conflicts of interest than is presented in the committee report. The second is a response by the other committee members. I. A PROPOSED MODEL FOR BROADER DISCLOSURE Lisa A. Bero, Robert M. Krughoff, and George Loewenstein We believe that the recommendations in Chapter 3 regarding disclosure of financial relationships or conflicts of interest would be greatly improved if they explicitly called for more extensive and standardized public dis- closure by researchers, physicians, and senior officials of institutions. We believe that—with the help of interpretation by the press, public-interest groups, researchers, health care consultants, patient representatives, and other information intermediaries—expanded disclosure would provide im- portant information for physicians, patients, researchers, health plans, regu- lators, policy makers, financial donors, and others who rely on research, practice guidelines, educational programs, and the quality and efficiency of medical care. We believe that the recommendations should be extended to a “broader- disclosure model” in which • The consensus-development process described in Recommenda- tion 3.3 would not only set out the standardized content, formats, and procedures for disclosure to institutions but also design a secure national 384 

APPENDIX F 385 online database system that could be used to report the same information to appropriate institutions and to the public. • The consensus-development process would set out minimum stan- dards as to the data elements that must be reported by people in specific roles (such as physician researchers and hospital administrators) to their institutions and would allow institutions to have higher standards if they so choose. • Each institution would require that any information on financial relationships or conflicts of interest that must be reported to it—or at least all information included in the consensus-defined minimum standards—also be made available to the public through the online system. We envision the broader-disclosure model working as follows: • The first time that a person was required to report information on financial relationships or conflicts of interest to an institution, he or she would register on the secure national online database system; create a profile with name, location, and other nonconfidential information; select a permanent ID; and get a confidential password. The system would have procedures for verifying the person’s identity. The person’s profile associ- ated with the ID would be sufficient to identify the person to public users of the database. (In the case of physicians, the profile might include a field for the National Provider Identifier number.) • The person would enter, in a standardized format dictated by the database system, at least the minimum standard information on all financial relationships or conflicts of interest that he or she was required to report to an institution. Depending on the person’s role in each institution with which he or she had a relationship, more or less information might be required. • When the person instructed the database to make any information available to any institution, the same information would automatically become available to the public. (Provisions might be made to protect some details of intellectual property, for example, of a drug formula until patent registration.) • The person-reported information in the database designed by the consensus-development process would ideally be linked to industry-reported information called for in Recommendation 3.4. One objective of the con- sensus-development process would be to make it convenient to find—in one place, in one format, for any person—any information reported by the person (to an institution) or by industry. Each database could be used to check the completeness of reporting in the other. 

386 CONFLICT OF INTEREST We stress the following elements of the model: • The model would require public reporting only by persons al- ready required to report to institutions and would not add any reporting burden—just “one more press of the button.” In fact, the model would probably reduce the overall burden on people by eliminating the need to re-enter information for reporting to multiple institutions. • The model would make it easier to report correctly by providing explicit standards and instructions. • The vast majority of practicing physicians do not have to report relationships or conflicts to institutions, so they would not be required to report to the public. • The information that people would have to report would be lim- ited to financial relationships and conflicts of interests related to drug, medical-device, or biotechnology companies, not other financial or personal information. • The person-reported information available to the public would add to what would be reported by industry according to Recommendation 3.4 in that it would include information on equity ownership in companies and testing facilities, patent rights, and other types of interests (see list in Table 3-3), not just payments from industry. • The system of person-reported information would allow people to incorporate more explanatory material about payments received (for example, reasons for payments for consulting) than would probably be reported by industry. • The centralized nature of the system would make it easier to up- date reporting requirements for everyone involved if future consensus- development processes deemed it important to include different types of information in standard reports. • The model would allow persons who might rely on information on financial relationships or conflicts of interest to obtain it when they want it—for example, before enrolling in a continuing medical education program or long after participating in one, or when meeting with family and friends before or after meeting with a surgeon rather than in the brief time with the surgeon. We believe this model would be a strong and flexible tool for manag- ing conflicts of interest. In key areas of health care, including those in the conflict of interest charge as the committee has defined it for purposes of this report, we are troubled by the possible harms that might arise from conflicts between commercial interests and patient and public interests. This is true in research, in education, and in the development of practice guidelines. But we believe that in each of these areas, totally eliminating all 

APPENDIX F 387 conflicts—for instance, removing all industry roles in translational research or barring all educational organizations from having any direct or indirect support from industry (even for research or for an endowed chair)—might involve more change than could be justified in light of how research, edu- cation, and medical-care systems have evolved. We conclude that greatly expanded requirements for public disclosure would create incentives and monitoring tools that would reduce the risk posed by some of the conflicts that it might not be practical to eliminate. As documented throughout this report, there are serious limitations in the accuracy, completeness, comparability, and timeliness of conflict of ­ interest information reported to institutions and to the public—for e ­ xample, as conflicts are shown in National Guideline Clearinghouse docu- mentation of practice guidelines or as conflicts are reported by speakers in continuing medical education programs. These limitations make it dif- ficult for patients, students, clinicians, and others who might be affected by conflicts to make timely assessments of their presence or severity. These limitations also make it difficult for researchers, the press, policy makers, and others to assess the extent of conflicts and the effectiveness of efforts to manage them. We believe that the broader-disclosure model would help to overcome the limitations of currently available information and that the information made available by the model would encourage and facilitate expanded efforts by researchers, the press, public-interest groups, and other information intermediaries to assess and compare conflict of interest poli- cies and practices of all relevant parties. Even if information on financial relationships or conflicts of interest were rarely used by patients, physicians, or others to make decisions, the fact of public reporting would probably motivate some researchers, physicians, and senior officials to eliminate unproductive conflicts. The model would also create incentives for people to report to institutions completely and accurately to avoid the risk of being identified as having failed to do so. We recognize the challenges of reaching broad agreement on standard content, formats, and procedures for reporting in an online system—even if the information would be reported only to institutions and not the public. But we believe, on the basis of academic research and experiences in our own organizations, that the cost of maintaining such a system would be minor. Knowing that the information would be public would encourage organizations to participate in planning and designing the system. We are aware that proposals for public disclosure often elicit concerns about compromising personal privacy. But most people would not have information on financial relationships or conflicts of interest to report and so would have nothing to report publicly. Most mutual-fund shares, stocks, bonds, bank accounts, salaries from institutions, income from medical 

388 CONFLICT OF INTEREST practice, and other forms of assets and income would ordinarily not be reported. We assume that even among people who would have relevant financial interests or conflicts of interest to report, the financial interests involved would usually constitute a relatively minor part of their financial affairs and not be a meaningful indicator of individual or family income or wealth; if this assumption is not accurate, public reporting of the informa- tion would be all the more important. There are numerous examples of public reporting of financial infor- mation currently in effect that have not been shown to have substantial adverse consequences or to discourage people from participating in the institutions or programs that require reporting—for example, the required public disclosure of salaries of government employees, the public disclosure of individual contributions to political candidates, the public disclosure (on Internal Revenue Service Form 990) of salaries of higher-paid employees of most tax-exempt nonprofit organizations, and, most pertinent, the currently required public accessibility, under state freedom of information laws, of financial relationships or conflicts of interest reported to state universities and health care systems. We are not persuaded by arguments that the model would create an unfair imbalance in reporting requirements between physicians who work for institutions and physicians who work only in private practice. We note that physicians who have relationships with universities and other institu- tions already have reporting requirements (to the institutions and subject to public release in the case of public institutions) that other physicians do not have. And we believe that the distinction between institution-affiliated physicians and other physicians is logical: physicians affiliated with in- stitutions are more likely than other physicians to have equity interests, intellectual-property interests, and other interests that may represent con- flicts, whereas reporting by every practicing physician would create a large and burdensome system that would not contribute much public informa- tion beyond that expected to be included in the industry disclosures under Recommendation 3.4. We are aware that there might be concerns about misinterpretation of the disclosed information. In a society with freedom of speech and press, any type of information can be misinterpreted or overemphasized. But we believe that the very discipline of free speech, armed with widely available information, would lead generally to better decisions than would result from less complete information. 

APPENDIX F 389 II. THE RESPONSE OF THE COMMITTEE MAJORITY TO THE PROPOSED MODEL FOR BROADER DISCLOSURE Wendy Baldwin, Lisa Bellini, Eric G. Campbell, James F. Childress, Peter B. Corr, Todd Dorman, Deborah Grady, Timothy S. Jost, Robert P. Kelch, Bernard Lo, Joel Perlmutter, Neil R. Powe, Dennis F. Thompson, and David A. Williams As described in Chapter 3 of the report, the full committee supports the development of a public database for company reporting of payments and generally favors making more information on financial relationships and conflicts of interest public. We do not, however, endorse the proposed broader-disclosure model, which calls for institutions that require disclosure ����������������������������������������������� from physicians and researchers to require that those individuals also make their disclosures public each time that they report a financial relationship or conflict of interest to those or any other institutions. (That is, the institu- tion would impose the requirement, but the individual would transmit the information.) We do not endorse the proposed broader-disclosure model for several ������������������������������������� ������������ reasons. First, most members were not convinced of the value that would be added by the suggested expansion of institutional requirements if the other recommendations made in this report were adopted. According to Recommendation 3.4, pharmaceutical, medical-device, and biotechnology companies would be required to report their payments to various individu- als and institutions, and that information would be available in a public searchable database. Depending on how many institutions adopted the additional public-disclosure requirements, the proposed expansion might yield some additional information about relationships or interests, such as holdings in publicly traded stock and possibly some expert-witness fees. Such relationships might already be public in specific contexts, for example, in connection with a journal article or educational presentation. In con- trast, congressional action on the Medicare Payment Advisory Commission (MedPAC) proposal for the disclosure of physician-ownership interests in health care facilities would provide information about conflicts of interest that are more likely to influence physician decisions about patient care. A second concern of the committee majority involved intrusions on privacy if physicians and researchers were required to make public the additional information that they disclose to academic medical centers and other institutions. It is likely that many people will not want further expo- sure to the risks of identify theft, mischaracterization by the mass media, or other kinds of harm, particularly if the database of expanded disclosures is privately managed. The privacy of family members is also at stake because some institutions require the disclosure of the financial relationships of 

390 CONFLICT OF INTEREST spouses or domestic partners and children for some purposes. Managing a secure and up-to-date website with personal information requires resources and expertise to protect against errors in disclosure, to offer ways to correct errors, and to clarify disclosures with supplementary information. If the information becomes public without such safeguards, there could be allega- tions of intentional deception when honest mistakes occur or when a person discloses information to other institutions that have different requirements or formats for disclosure. A system would also need to protect against the malicious entry of erroneous information. We believe the committee did not have the expertise to investigate many matters like these. A third concern was the additional cost of expanded public disclosure. For example, the proposed unified database would require that the ad- ditional disclosures be approved for integration into a federally mandated and overseen database of company-reported payments, or, alternatively, some party would have to create and manage an integrated, secure private database. Either would involve additional costs for creating, maintaining, updating, and correcting the integrated database and maintaining security. In an era of increasing cost pressures on medical institutions and govern- ments, the committee is not convinced that spending for additional, mar- ginal public disclosures can be justified over such alternatives as spending for electronic medical records. In addition, in the committee’s experience, estimates of costs for information systems, even seemingly straightforward ones, often fall short of actual costs. Fourth, we were concerned about setting up a disparity, in particular, between university faculty and private practitioners and between medical institutions that require additional disclosure and ones that do not. Al- though it is not clear how many institutions would choose to require physi- cians and researchers to make public their disclosures to all institutions, the institutions that did so would place an extra burden on people who, for the most part, are faculty members whose relationships and conflicts of interest are already overseen by their academic institutions. In contrast, many physi- cians in private practice have no reporting requirements and no oversight. Thus, the expansion of disclosure is not targeted to higher-risk situations. Furthermore, unless the additional public reporting of institutional disclo- sure was mandated by the U.S. Congress, there could be perverse conse- quences for academic or other institutions that required people to make public the information that they disclose both to those institutions and to other institutions. Some physicians and researchers might be attracted by such transparency; but we believe that others would prefer to work at institutions that kept their disclosures confidential, except when disclosure is required for specific purposes, such as publication of a journal article or participation in the development of a clinical-practice guideline. Finally, we are concerned about other risks and unintended adverse 

APPENDIX F 391 consequences of requiring additional public disclosure beyond company- reported payments. For example, the requirement would add to the risk that information from different sources might fail to match exactly because of technical errors or differences in reporting requirements, procedures, or periods. Some might seize on the lack of an exact match as evidence of mis- behavior—that is, a deliberately incomplete or inaccurate disclosure—on the part of institutions or individuals, who then might have to respond to public accusations; this would distract from their primary responsibilities for research, education, or clinical care. Misinterpretation already may oc- cur with the reporting of payments by companies to physicians; for exam- ple, reporters may treat scientific and promotional consulting as equivalent and deserving of the same criticism. Overall, the majority of committee members thought that making public the information that physicians and researchers report to institutions was not supported by the principle of proportionality and that responses to conflicts of interest should be based on assessment of their severity. The likely burdens on individuals and institutions of an expanded public- disclosure system beyond that proposed in Recommendation 3.4 or already in place in accordance with other public or private policies are dispropor- tionate to any benefits from the marginal amount of additional information that would be provided.   According to Recommendation 3.3, consistency in institutional disclosure requirements and formats would increase and reporting burdens would decrease for people who must make disclosures to multiple institutions.