High-profile incidents, such as the 1999 death of Jesse Gelsinger in a gene-transfer study of ornithine transcarbamylase deficiency (OTC), have drawn attention to the threat that competing interests pose to the integrity of the research enterprise and the safety of study volunteers. The enactment of conflict management policies by governments, universities, and professional societies has helped to clarify the ethics of conflict of interest in clinical trials. However, much of the analysis and many of the policies intended for the ethical management of conflicts have centered on investigators’ relationships with financial sponsors of research. As such, financial interests not related to sponsorship have tended to receive less attention.
Consider, for instance, the conflict of interest policy devised by the American Society of Gene Therapy in the wake of Gelsinger’s death. This policy’s only prescription states that “all investigators and team members directly responsible for patient selection, the informed consent process and/or clinical management in a trial must not have equity, stock options or comparable arrangements in companies sponsoring the trial.”1 The policy seems poorly suited to addressing conflicts surrounding the OTC gene-transfer trial. First, investigator James Wilson was reportedly not involved in any aspects of patient care; rather, these were left to other clinical investigators who were never alleged to harbor conflicting financial interests.2 Second, conflicts in the trial were not directly related to sponsorship. Though the company that James Wilson had founded (Genovo) provided funding to the institute where the study was conducted, the trial was, in fact, sponsored by the National Institutes of Health. Conflicting interests in the study stemmed instead from a web of patents, licensing options, and institutional relationships. Specifically, Genovo was entitled to negotiate licenses for products developed by Wilson, who, along with the University of Pennsylvania, held major equity interest in Genovo.3 Also, Wilson was named as an inventor on key patents related to the trial, including one describing the use of adenovirus to treat liver disorders, such as OTC deficiency.4 The patents were owned by the University of Pennsylvania.
The following study and analysis are concerned with the ethics of trials conducted by investigators who are named as inventors on any patent included in their investigation. Current human protection policies of the U.S. and Canadian governments, nine major medical and research societies, 13 leading institutions, and high-impact journals were canvassed to identify key differences in these policies. These differences have been used as a basis for an ethical analysis and a series of recommendations.
I sought federal policies on conflict of interest in clinical trials from guidelines published by the Office of Human Research Protections and the Food and Drug Administration. I obtained Canadian guidelines from the Tricouncil Policy Statement.
I sought a representative sample of major professional society policies by searching PubMed using the terms “(trial OR research) AND conflict AND interest,” restricting the query to the guideline publication type. When policies were identified, I conducted a search for other policies using the Related Articles function on PubMed, or by checking citations in the guidelines. Only society policies relating to clinical trials, and policies of societies likely to oversee early phase clinical research on a regular basis, were retained for extraction. Thus, for example, I did not extract policy information from the American Dental Association or the American College of Emergency Physicians, though I acknowledge that members of such groups might conduct studies on products they have invented. I supplemented the collection of policies by searching Web sites of several organizations that I felt would be obvious candidates for conflict policies relating to trials, including the American Association of Universities, Pharmaceutical Research and Manufacturers of America and Biotechnology Industry Organization. Searches were conducted in Spring 2005. Two coders independently extracted features relating to intellectual property.
The 10 medical schools in the United States and the three in Anglophone Canada receiving the most federal research funding for 2002 were identified from National Institutes of Health5 and Canadian Institutes of Health Research6 sources. Our reasoning for restricting our search to these 13 universities was based on the premise that institutions receiving high public funding for biomedical research would be more likely to seek patents for translational innovations; therefore, these institutions would be the most likely to have human-protection policies that address intellectual property. After selecting these 13 universities, two research assistants and I retrieved conflict policies from the universities’ Web pages and extracted features relating to patents. In March and April of 2005, I contacted senior administrators of ethics and/or conflict of interest committees at each institution to confirm the accuracy and completeness of our extractions. All but one responded to three or fewer queries; for the one institution that declined contact, a research assistant independently extracted their policy; agreement between the two assessors was complete.
Finally, I selected 17 journals for analysis on the basis of their reputations for publishing high-impact translational studies, a large volume of early-phase trials, or coverage of a research area where intellectual property issues are prominent, such as gene transfer. Using the ISI Web of Knowledge Journal Citation Reports for 2003, I identified the top three journals within the categories of general and internal medicine, research and experimental medicine, and oncology. I excluded journals if they primarily published review articles or tended to publish research originating from outside of North America. I searched the ISI journals database using the search words “gene transfer,” “gene therapy,” or “gene medicine” to identify gene-transfer journals.7 I also included Science and Nature, the two most highly cited nonreview journals for 2003 not already contained in the above categories. Two coders extracted policies relating to patents; in extracting these policies, coders disagreed on two items (94% agreement); these differences were resolved by discussion.
Federal guidance in North America on managing inventor–investigator conflicts is limited. The U.S. Department of Health and Human Services’ draft guidance on conflict of interest, issued in early 2001, granted broad discretion to institutions for defining financial interests, but it identified holding a patent on a product being tested as a conflict that IRBs should consider in their reviews. It further urged institutions to disallow clinical investigators with financial conflicts from conducting informed consent or data analysis, and it encouraged investigators (or institutions) to disclose financial arrangements to research subjects when conflicting interests cannot be eliminated.8 In ensuing years, policy makers have removed the latter recommendations.9 The Food and Drug Administration’s policy, which is intended for assessing scientific validity rather than protecting human subjects, demands that clinical investigators report relevant patent holdings in any agency submissions.10 However, the policy excludes phase 1 submissions,11 which probably account for the bulk of studies involving inventor–investigators. Canada’s Tricouncil Policy Statement makes no mention of patent holding in trials, but instructs investigators to disclose any “actual, perceived or potential” conflicting financial interests to ethics review bodies.12
Policies of professional associations, research institutions, and journals
One obvious reason why policies might vary among societies, institutions, and journals is that inventor–investigator conflicts of interest might be very rare in some settings. We therefore biased our sample of societies, institutions, and journals towards those that would be expected to confront such conflicts on a somewhat regular basis.
Several professional societies have issued conflict policies that address clinical trials (Table 1). Among the 13 universities examined, all had conflict of interest policies, and most referenced patents (Table 2). However, professional society and institutional policies showed disagreement along three lines. The first concerns thresholds: several stipulate licensing of a patent (rather than its issuance) as a trigger for conflict of interest management. Second, policies varied with respect to requiring disclosure to prospective subjects; of the 10 institutions that reference patents in their conflict policies, six formally require that inventors disclose patent holding to research subjects. Third, several policies presumptively exclude inventors from participation in clinical trials of their invention.
The survey of journals showed similar variation. Competitive journals generally mentioned patent inventorship, and several use a stricter trigger for management—patent filing—than did any association or institution studied. Only one uses licensure as a trigger for disclosure to editors, and one uses a policy that unambiguously requires disclosure of related patent holdings to readers. Despite the adverse publicity concerning financial conflict of interest, no gene-transfer specialty journals (all of which occasionally publish clinical trials) mention patents in their instructions to authors or policies.
The heterogeneity of policies with respect to presumptive prohibitions, disclosure requirements, and policy triggers suggests the need for an ethical examination of patent inventorship in clinical trials. Below, I present my own analysis of this issue.
Two aspects of patents—their nature and the process of obtaining them—present ethical challenges to clinical trials. By their nature, patents exist to promote commercial activities by converting knowledge into property. In the context of university research, patents therefore merge financial interests into knowledge-seeking activities. Patents originating from research conducted at a university typically are obtained by technology transfer offices, which pay application fees and negotiate licensing. Though institutions, as applicants, are generally assigned ownership of these patents, universities encourage faculty to seek patents by sharing patent income. Indeed, the Bayh–Dole Act requires patent-holding institutions to share royalties.13 Policies on how royalties are shared vary. Harvard University, for example, entitles faculty inventors to 35% of the first $50,000 of royalties, and 25% of royalties exceeding this amount.14 The University of California system awards 35% of all royalties to its inventors.15 Thus, although academic inventors generally cede ownership of a patent to their institution, they have potential financial interests tied up in any patents they help file.
Whether such financial interests warrant management in the context of clinical trials depends in part on whether they might cause investigators to act in a manner that jeopardizes the safety, informed consent, or scientific value of a study (several policies canvassed above refer to this threshold as “significance”). One factor favoring significance is that a patent’s commercial value depends in part on a trial’s outcome. Agents that show safety and efficacy are more likely to attract licensees (and hence, income) than those that fail testing. In general, conflict of interest policies regard financial arrangements that are contingent on a study’s outcome with particular concern.
A second factor favoring significance derives from the time-sensitive nature of a patent’s value. Because patents confer on holders a monopoly for a limited period, their earning potential diminishes as expiration approaches. Investigators who hold patents thus have financial incentives to conduct trials expeditiously. These pressures can interfere with patient safety, as when recruitment difficulties tempt investigators to relax eligibility criteria. Haste to promote a product through the various trial phases can also erode an experiment’s scientific value; patent holders may be less inclined to collect information that contributes generalizable insights but does not necessarily advance the invention’s commercial prospects.
A second set of reasons why patents can conflict with research ethics relates to the secrecy that is integral to obtaining patent rights. Information about new inventions must be kept secret before patent filing (or within the 12 months before filing, depending on the jurisdiction) and will continue to be secret until the patent application is published (normally a period of 18 months after filing in most jurisdictions). Secrecy is crucial because, to qualify for patent rights, inventions must be deemed novel, which means that pursuant to U.S. law, the invention has not been “described in a printed publication … more than one year prior to the date of the application for patent.”16 Most jurisdictions include similar provisions in patent legislation. Recent U.S. case law has increased the stringency with which inventors enforce prefiling secrecy.17
Secrecy surrounding patent filing is obviously a concern when considering clinical trials if patents have not been filed, or filed and not published, before the start of a trial (as is sometimes the case in European Union countries) or one year before the start of a trial (as mandated in Canada and the United States). Moreover, even when details about the invention are available after disclosure in a published patent application, it is not uncommon for investigators to withhold specific details about their inventions from the public. The details that are not disclosed in a patent application may include information about the optimum function of the invention; this nondisclosure helps to preserve the inventor’s competitive advantage and to ensure that the patent encompasses a broad scope. Such details are usually protected as trade secrets.
Studies involving agents for which the inventor has filed for or received patent protection raise concerns about safety when investigators withhold invention details from their patents or publications. In such circumstances, investigators might also withhold information from coinvestigators, review bodies, or colleagues about an investigational agent’s properties. Such secrecy can be particularly problematic for the types of inventions likely to emerge from translational research, where little is known about a product’s behavior and safety. For example, the Recombinant DNA Advisory Committee (RAC) normally conducts public review of novel gene-transfer trials. Although investigators who hold or intend to pursue patents may prefer to avoid this public forum, such nondisclosure thwarts peer review, which is an important mechanism for maximizing safety.
Any argument that patent holding by investigators in clinical trials is “significant” must overcome several objections. The most obvious rests on the indeterminate nature of a patent’s financial value: many university patents are never licensed and therefore never accrue returns for inventors. The Massachusetts Institute of Technology, for example, failed to license 49% of its patents assigned between 1980 and 1996; the figure may be higher at less competitive institutions.18 Even when patents are licensed, earnings remain speculative and might not influence the behavior of investigator–inventors. For example, people tend to prefer present and certain gains to those that are delayed or uncertain (a phenomenon known as “discounting”).19 Because royalty earnings involve greater delay and uncertainty than many other financial interests, the potential financial returns of patents might not significantly influence investigator conduct.
These objections are undermined by several considerations. First, patent holding can be remunerative in ways other than collecting royalties. When universities license a patent, they typically charge the licensee an up-front fee (generally, $10,000 to $50,000, but occasionally as high as $250,000) in addition to negotiating a share of a product’s royalties.20 Because fees are immediately paid, an inventor’s financial interest in a patent is sometimes less delayed than if income were entirely royalty based. Second, universities are increasingly favoring equity-based licensing, in which institutions cede rights to royalties in exchange for equity in the licensee. Equity-based licenses are considered less risky than royalty licenses because returns on the former are spread across the value of an entire company rather than a single product. This would tend to diminish the uncertainty and delay involved in patent-related financial interests. Third, discounting is a continuous (rather than categorical) phenomenon: that persons discount future earnings doesn’t logically imply that they don’t value them at all. Finally, patent filing requires that universities commit resources to prosecuting a patent and paying application fees. Universities primarily (though not exclusively) pursue patents because they believe an invention stands a reasonable chance of earning fees or royalties.21 If universities anticipate that a product will be licensed and produce returns, they should anticipate that a patent related to a human study will create a significant conflicting interest. To do otherwise would, at best, create the appearance of a double standard.
Without data showing how investigators behave when they hold patents, determinations about whether inventorship influences the conduct of a trial can only be speculative. Nevertheless, the harms from not treating patents as significant in clinical trials may exceed those of treating them as such. In most circumstances, patents probably do not return significant earnings to inventors. However, conflict committees or IRBs cannot know in advance whether a lucrative license will be negotiated on a patent after initiation of a trial. Devising a conflict policy that covers inventorship becomes a matter of weighing the harms of type 1 errors (declaring significance when it does not or will not exist) against those of type 2 errors (failing to declare significance where it should have been). A precautionary policy on protecting human subjects would favor the former.
If patent inventorship warrants management in clinical trials, what sorts of measures might be appropriate? Any such management policy should promote three major principles that underlie research ethics: consent, safety, and scientific value.
First, investigators’ patent holdings should always be disclosed to IRBs, and presumptively to research subjects. Disclosure to IRBs is necessary because patent conflicts are potentially significant; IRBs should have sufficient information to decide whether a conflict requires additional management strategies. Disclosure to subjects is predicated on the notion embodied in legal doctrine that the consent process should disclose any information a reasonable potential participant might consider material to his or her enrollment decision. Although data are scarce on how volunteers regard patents, one survey indicates that most potential participants consider patent holding relevant to consent: 59% to 64% of potential participants deemed it “extremely important” to know about a researcher’s patents, and 23% to 31% would be less inclined to enroll if an investigator held a patent on the study agent.22 The previous sections argued for the “reasonableness” of such preferences.
Second, investigator–inventors should be presumptively barred from certain responsibilities. Categorical bans on patent holding would be undesirable, because patented inventions are novel (though legal and scientific definitions vary) and often scientifically intricate; exiling the inventor’s expertise could undermine a study’s safety. Nevertheless, certain circumstances (for example, once patents are returning royalties) demand excluding patent holders from any role in a trial. A preferable approach would curtail patent holders’ responsibilities in trials. Patent holders should not function as principal investigators, because their responsibilities permeate every aspect of a trial. Even if the principal investigator is insulated from certain activities, his or her leadership threatens the disinterestedness of cooperating investigators because of team identification: in the process of serving the inventor, collaborators often find their independence and objectivity compromised.23 Activities like recruitment, obtaining consent, assessing and reporting adverse events, evaluating the eligibility of volunteers, and data analysis should presumptively exclude patent-holding investigators, because each may be influenced adversely by conflicting interests.
Third, patents authorized by the Bayh–Dole Act also create institutional conflicting interests; these should be managed in a manner that parallels those for investigators. Namely, a categorical disallowance of institutions from hosting trials on agents for which they hold patents would be unwise, because such institutions may have the most suitable facilities and expertise for conducting these trials. However, institutional, patent-related conflicting interests should be subjected to independent review and presumptively disclosed to research subjects. To avoid concerns that an IRB might internalize its institution’s aspirations for prestige and financial flourishing,24 a committee external to the institution should conduct review. (There is insufficient space in this report to address the ethical complexities of external review; the interested reader is directed to proposals outlined by others.25,26)
Fourth, investigators should always disclose to subjects and IRBs that they have filed for a patent, even if the patent has not yet been received. Although this extends conflict of interest management into realms where financial interests are incipient, oversight of commercial intention is not unprecedented. Concerned about ensuring “as full communication as possible among investigators and clinicians concerning research methods and results”—RAC, for example—invites investigators to disclose patenting intentions.27 Similarly, as shown in Table 3, many journals ask authors about patent intentions.
Fifth, “presumptive” policies described above should be adjusted according to the trial and patent. Conflicts caused by newer, broader, and pioneering patents warrant more cautious management than those deriving from older, narrower, or more incremental patents, and conflicts are of greater concern when trials involve newer and riskier technologies. IRBs and/or conflict committees might use technology transfer officers on an ad hoc basis to gauge a patent’s broadness and significance and to adjust regulations on inventor–investigators accordingly.
Lastly, the foregoing analysis highlights numerous questions concerning the extent and effects of patent holding in clinical trials. At present, little information is available about the frequency with which inventor–investigators conduct clinical trials, whether patent holding influences the decision making of clinical investigators, whether it impinges on consent discussions, and whether patent filing or holding often results in withholding information from oversight bodies or other investigators. Having reviewed hundreds of gene-transfer protocols and consent documents, my own anecdotal observations are that it is not uncommon for inventors to be principal investigators without disclosing their status as patent holders in informed consent documents. Obtaining quantitative data on such issues would help to anchor conflict policies on evidence rather than inference and anecdote.
The wide variation in governmental, institutional, and professional society policies reported here underscores the degree to which the ethics of patents and human research ethics remain unsettled. It also likely reflects the fact that institutions have tended to prioritize the management of sponsor-related conflicts of interest.
The conflict of interest policies I have presented in the results are subject to three important provisos. First, I did not set out to provide a comprehensive survey of policies, and instead I focused my analysis on societies, institutions, and journals that I considered more likely to be involved in inventor-run trials. Accordingly, the sample of policies may not have captured some variations. Second, institutional guidelines are evolving rapidly, and it is possible that by the time this report is published, some policies may have been modified. Last, this analysis only captured formalized policies. It is conceivable that some society and institutional practices depart somewhat from formal requirements.
Having reviewed conflict policies and offered approaches to the ethical management of trials involving inventor–investigators, it is worth reflecting, once again, on how well policies described above would manage conflicts like those encountered in the University of Pennsylvania OTC gene-transfer trial. First, policies like that of Johns Hopkins University, requiring patent licensure to trigger protections, would not have managed some of the important financial interests maintained by James Wilson. The University of Pittsburgh’s policy would not have restricted Wilson’s responsibilities because, as already noted, he was not a principal investigator. The University of Washington’s policy would not have required disclosing to subjects that an investigator on the trial held patents on the study agent.
At least one observer has stated that the suggestion that Wilson was influenced by financial interests is “irresponsible,”2 and this is not an appropriate forum to retry James Wilson. It’s also not clear that a better conflict of interest policy would have prevented Gelsinger’s death, though it might have prevented some of the public outrage in the study’s aftermath. The response of journalists, ethicists, and Gelsinger’s father to Gelsinger’s death underscores that the types of conflicting interests surrounding the OTC study, which included held patents, are often perceived as significant. Not surprisingly, the University of Pennsylvania has since enacted one of the most restrictive policies surveyed.
For understandable reasons, institutions have tended to prioritize the management of sponsor-related conflicts of interest. Nevertheless, renewed emphasis on translational research28 makes it likely that clinical investigators and IRBs will need to navigate the ethical challenges of patent-related conflicts of interest. Research policies should steer a course that safeguards the trust that human volunteers vest in clinical investigators and their institutions.
This work was funded by the Canadian Institutes of Health Research. Without implying their endorsements, the author thanks Richard Gold and members of the CTRG for helpful suggestions on a previous draft, and David McLaughlin and Lindsey Miller for research assistance. Particular thanks go to Karen Durell for invaluable input. The author also wishes to thank institutional administrators who responded to his queries about their conflict policies. The author declares he is an employee of McGill University, and serves in a non-remunerative capacity on the ethics committee of the American Society of Gene Therapy. Conflict policies for both were reported in this manuscript. The author declares no other competing interests.
2 Marshall E. Gene therapy’s web of corporate connections. Science. 2000;288:954–955.
3 Gelsinger v. Trustees of the University of Pennsylvania. (Phila Cnty Ct of CP filed September 18, 2000), Sherman, Silverstein, Kohl, Rose and Podolsky Law Offices. Available at: (http://www.sskrplaw.com/links/healthcare2.html
). Accessed September 18, 2006.
4 Wilson JM, inventor. Method of treating liver disorders. WIPO patent WO9730167. August 21, 1997.
8 U.S. Department of Health and Human Services. Draft Interim Guidance: Financial Relationships in Clinical Research: Issues for Institutions, Clinical Investigators, and IRBs to Consider When Dealing with Issues of Financial Interests and Human Subject Protection. Available at: (http://www.hhs.gov/ohrp/humansubjects/finreltn/finguid.htm
). Accessed September 18, 2006.
9 U.S. Department of Health and Human Services. Final Guidance Document: Financial Relationships and Interests in Research Involving Human Subjects: Guidance for Human Subject Protection. Available at: (http://www.hhs.gov/ohrp/humansubjects/finreltn/fguid.pdf
). Accessed September 18, 2006.
10 21 C.F.R. Sect. 54.4(a)(3)(iii).
13 35 U.S.C. § 202.(c)(7)(B).
14 Bloomberg CA. Developing an intellectual property portfolio for the academic or not-for-profit institution. Nat Biotechnol. 2005;23:119–121.
15 Slaughter S, Rhoades G. Academic Capitalism and the New Economy: Markets, State and Higher Education. Baltimore: Johns Hopkins University Press; 2004.
17 Teitelbaum R, Cohen MS. Publish and perish: what constitutes a bar under the patent laws. Nat Biotechnol. 2004;22:1449–1450.
18 Shane S. Selling university technology: patterns from MIT. Manage Sci. 2002;48:122–137.
19 Green L, Myerson J. A discounting framework for choice with delayed and probabilistic rewards. Psychol Bull. 2004;130:769–792.
20 Bray MJ, Lee JN. University revenues from technology transfer: licensing fees vs. equity positions. J Business Venturing. 2000;15:385–392.
21 Thursby JG, Jensen R, Thursby MC. Objectives, characteristics and outcomes of university licensing: a survey of major U.S. universities. J Technol Transf. 2001;26:59–72.
22 Kim SY, Millard RW, Nisbet P, Cox C, Caine ED. Potential research participants’ views regarding researcher and institutional financial conflicts of interest. J Med Ethics. 2004;30:73–79.
23 Sharpe VA. Science, bioethics and the public interest: on the need for transparency. Hastings Cent Rep. 2002;32:23–26.
24 Barnes M, Florencio PS. Financial conflicts of interest in human subjects research: the problem of institutional conflicts. J Law Med Ethics. 2002;30:390–402.
25 Lemmens T, Freedman B. Ethics review for sale? Conflict of interest and commercial research review boards. Milbank Q. 2000;78:547–584.
26 Wood A, Grady C, Emanuel EJ. Regional ethics organizations for protection of human research participants. Nat Med. 2004;10:1283–1288.
27 National Institutes of Health. Guidelines for Research Involving Recombinant DNA Molecules. Bethesda, MD: National Institutes of Health; 2002.
28 National Institutes of Health. NIH Roadmap. Available at: (www.nihroadmap.nih.gov
). Accessed September 18, 2006.
29 AAMC Task Force on Financial Conflicts of Interest in Clinical Research. Protecting subjects, preserving trust, promoting progress I: policy and guidelines for the oversight of individual financial interests in human subjects research. Acad Med. 2003;78:225–236.
30 Alpert JS, Shine KI, Adams RJ, et al. Task force 1: The ACCF and AHA codes of conduct in human subjects research. J Am Coll Cardiol. 2004;44:1724–1728.
31 Califf RM, Nissen SE, DeMaria AN, et al. Task force 2: investigator participation in clinical research. J Am Coll Cardiol. 2004;44:1729–1736.
32 Jacobs AK, Lindsay BD, Bellande BJ, et al. Task force 3: disclosure of relationships with commercial interests: policy for educational activities and publications. J Am Coll Cardiol. 2004;44:1736–1740.
33 American Society of Clinical Oncology. American Society of Clinical Oncology policy statement: oversight of clinical research. J Clin Oncol. 2003;21:2377–2386.
34 American Society of Clinical Oncology. Revisions of and clarifications to the ASCO conflict of interest policy. J Clin Oncol. 2006;24:517–518.
35 Redmond DF, Freeman T. American Society for Neural Transplantation and Repair. The American Society for Neural Transplantation and Repair Considerations and guidelines for studies of human subjects. The practice committee of the society. Approved by council. Cell Transplant. 2001;10:661–664.
36 Task Force on Research Accountability, American Association of Universities. Report on Individual and Institutional Financial Conflict of Interest. Washington, DC: American Association of Universities; 2001.
37 Luce JM, Cook DJ, Martin TR, et al. The ethical conduct of clinical research involving critically ill patients in the United States and Canada: principles and recommendations. Am J Respir Crit Care Med. 2004;170:1375–1384.
38 Komesaroff PA, Bach MA, Danoff A, et al. The Endocrine Society Ethics Advisory Committee: ethical aspects of conflicts of interests, October 2003. Endocrinology. 2004;145:3032–3041.
39 Pharmaceutical Research and Manufacturers of America. Principles on Conduct of Clinical Trials and Communication of Clinical Trial Results. Washington, DC: Pharmaceutical Research and Manufacturers of America; 2004.