GEOFFREY LEVITT
WHAT IS the purpose of communicating information about the safety of a drug? From a public health perspective, that purpose can be defined as supporting informed choices about the use of medicines that lead to optimal patient outcomes—i.e., that provide the greatest possible patient benefit with the least possible patient risk. Indeed, the U.S. governmental body charged with analyzing and providing advice on the effective communication of drug safety information—the Food and Drug Administration (FDA) Risk Communication Advisory Committee (2009)—has described its own mission very much along those lines.
So how then do we ensure that drug safety communication actually accomplishes this purpose? Two main ingredients are necessary: (1) there must be valid and adequately substantiated safety information to communicate, and (2) that information must be communicated accurately and meaningfully.
While those two ingredients may seem straightforward enough, in practice each of them has given rise to vast bodies of theory and practice that are still very much in a state of evolution. The process of generating adequately substantiated safety information about a drug in a timely and actionable form—the realm of pharmacovigilance—is the subject of intense medical, statistical, and epidemiological analysis, with a superstructure of comprehensive government regulation and a substructure of human and electronic information processing resources that must cope with thousands of pieces of data flowing in every day. And the process of determining how best to communicate that information—whether in the context of the product’s approved labeling or in less formal messaging to patients and health care providers—is a matter of endless revision and active debate among regulators, academics, legislators, and the public.
Alongside all this complexity, however, is a key element without which the drug safety communication system cannot be properly understood, namely, the element of governance. Who has the right, and who has the responsibility, to engage in communications about the safety of a drug? In colloquial terms, who “owns” drug safety communication? Regulators? Industry? Third-party experts? Health care providers? Patients, consumers, or caregivers?
The thesis of this chapter is that we are in a time of unprecedented upheaval in the governance of drug safety communication that has resulted in a diffusion of control away from the parties who traditionally held primacy within this system—in particular, regulators and industry. This upheaval has been driven by three overlapping trends: (1) increasing doubts on the part of a number of vocal observers as to the ability and willingness of regulators and industry to communicate promptly and accurately about drug safety; (2) a related movement toward greater disclosure of certain kinds of health-related information, in particular the results of clinical research, which goes hand in hand with the ever-increasing ease of access to such information through electronic channels; and (3) a broader trend toward increased transparency of scientific and technological information overall (sometimes referred to as the “open-source” or “open-science” approach). Regulators and drug companies, while still central players in the system, now must contend with a constant flow of safety analyses and communications from a wide variety of independent third parties, often putting them in a reactive or even defensive position. This changing environment will require the formulation of new “rules of the road” in order to ensure the quality and integrity of both drug safety information and the manner in which it is communicated to the various stakeholders who rely on it.
I. GENERATING VALID, SUBSTANTIATED DRUG SAFETY INFORMATION
Traditionally, the postmarketing drug safety system has relied on two main sources of information: spontaneous adverse event reporting, and data from organized research studies (21 CFR § 314.80).1 The limitations of spontaneous reporting are well known. Among other challenges, it is inherently haphazard, arbitrary, and prone to a variety of reporting and interpretation biases. Yet as a generator of possible safety signals and an indicator of broad patterns in the patient population at large following approval of a drug, spontaneous reporting has an undeniable role. Over the years, various additions and modifications to spontaneous reporting processes have been implemented in the United States, Europe, and other major jurisdictions. The current system in the United States is based on MedWatch and dates back roughly twenty years (Kessler 1993). The European Union has only recently overhauled its safety-reporting system to provide more transparency, better communication, and greater engagement of patients and health care professionals, including direct consumer reporting of suspected adverse drug experiences to regulators (European Medicines Agency (EMA) 2014).
Yet even with these reinforcements, spontaneous adverse event reporting remains at best an imperfect source of safety information. Indeed, its primary value often lies not in the information it furnishes directly but in its ability to provide signals that can then be addressed through more systematic research—the paradigmatic process of signal detection followed by hypothesis generation and testing. The testing phase of this process can take a number of different forms depending on the circumstances, ranging from case reviews through epidemiological studies to full-blown randomized safety trials that can last many years and cost hundreds of millions of dollars.
Systematic testing of this kind clearly can provide valuable information on drug safety issues. In recognition of this, regulators have not shied away from requiring drug sponsors to undertake such studies when they perceive safety issues needing to be addressed, whether through postmarketing commitments or the outright imposition of study requirements under legal authorities such as section 505(o) of the U.S. Food, Drug, and Cosmetic Act (FDCA) (21 USC § 355(o)). Yet such studies have their limitations as well. Exponential increases in clinical study costs mean that only a fraction of the safety issues that may be present with a drug can feasibly be addressed in this fashion. Even when they are, the study in question may take a long time to yield results, during which patients are still being exposed to the drug. Even with the best possible intentions, design and implementation challenges may ultimately prevent a study from providing fully useful data.
For these reasons, there has been longstanding interest in devising a postmarketing safety surveillance system that could supplement the traditional model by offering timely, accurate, and actionable safety information at a reasonable cost. In the United States, this interest led to the creation of the Sentinel System in 2007 as part of legislation reauthorizing FDA to collect drug application user fees (FDA Amendments Act of 2007 (FDAAA), Pub. L. No. 110-85, § 905(a), 121 Stat. 823, 944). Still in its implementation phases, the Sentinel System aims to establish a network of health care data sources spanning government agencies, academic medical centers, and managed-care organizations that will provide targeted real-time safety information and analysis (Behrman et al. 2011; FDA 2014a). One of the innovative features of the Sentinel System is that the information relied upon to support safety analyses is not submitted, gathered, and processed at a central location but remains with its original owners, who become active participants in the network (FDA 2010). This “distributed” model plainly has significant potential time and cost advantages over the traditional centralized approach where the regulator or the sponsor has the burden of gathering, collating, and analyzing masses of safety data. At the same time, for better or worse, it reduces the sponsor’s ability to control both the analysis and the reporting of emerging safety issues that may affect one of its products.
II. COMMUNICATING DRUG SAFETY INFORMATION
The effective communication of drug safety information, once such information is generated by the data gathering and analysis processes summarized above, faces a number of challenges. Such information is inherently complex and may be in a form or at a stage that is still subject to further verification, requiring careful calibration and qualification to accurately convey the associated level of uncertainty. In addition, it often must be communicated to multiple audiences with varying levels of sophistication (lay people, caregivers, health care professionals) and with disparate roles in relation to the drug itself (patients, prescribers, payers). For instance, in March 2013, FDA issued a public drug safety communication about new findings indicating an increased risk of pancreatitis and precancerous cellular changes in diabetic patients treated with a commonly used class of antidiabetics known as incretin mimetics. The unpublished findings, which were provided by a group of academic researchers, were based on examination of a small number of pancreatic tissue specimens taken from patients after they died from unspecified causes. The FDA communication noted that the agency had not concluded that these drugs may cause or contribute to the development of pancreatic cancer and emphasized that there should be no change in prescribing or use of the drugs pending further review (FDA 2013). Plainly a message of this nature requires an extremely careful approach in order to appropriately balance the intent to provide meaningful new safety information on a current basis with the need to avoid unnecessarily deterring patients or health care professionals from using a potentially lifesaving treatment.
It is therefore no wonder that an entire discipline has grown up around the process of formulating and transmitting drug safety information. While a detailed look at that discipline is beyond the scope of this chapter, for current purposes we can identify a few major categories of drug safety communication. The most fundamental, of course, is the approved drug label itself, which contains detailed official information about a drug’s side effects, precautions, warnings, and contraindications (21 CFR § 201.57). The contents of the label are primarily directed to health care professionals and are accordingly written in technical language, which, at least in principle, requires medical training to fully understand and apply. The primary intent of the drug label is to provide the prescribing physician with sufficient information about the risks and the benefits of the drug to make informed prescribing decisions tailored to the individual patient. In some cases the label may also contain a section specifically directed to the patient to help that patient better understand the use, risks, and benefits of the drug that has been prescribed—the Patient Package Insert (21 CFR § 310.501, 21 CFR § 310.515; FDA 2012).
Many other types of drug safety communication can be considered a supplement, update, or gloss on the approved label. For instance, the approved MedGuide is intended to offer the patient further details about drugs from a user’s, as distinct from a prescriber’s, perspective (21 CFR Part 208). Similarly, the sheaf of information a patient receives at the drugstore when a prescription is filled is also intended to “translate” available information about drugs into lay terms for the convenience of the patient. Incidentally, this information packet—usually prepared at the behest of the pharmacy and not necessarily possessing official approved status—also poses some risk of inconsistency and/or confusion absent validated quality standards.
At another level entirely are professionally oriented messages conveying newly acquired information about a drug’s risks. That new information may emerge from any of the sources discussed above, such as postmarketing surveillance, new analyses, or new studies. If disseminated by the sponsor (whether on its own initiative or as directed by a regulator), such messages typically take the form of a “Dear Healthcare Professional” letter (FDA 2012). Over the past few years, FDA has put in place mechanisms to allow the agency to disseminate drug safety information directly to health care professionals and the public without the involvement of the sponsor. These so-called “Drug Safety Communications” may be used to communicate emerging information about the risks of a drug relatively early in the analytical process, often before the risks in question have been fully analyzed or understood. Because these communications are public, they may also be picked up and rebroadcast by the lay media, Internet services, or other third parties, with or without additional commentary or interpretation.
Finally, there are the much broader and more amorphous kinds of messages about a drug’s risks that may be contained in reports put forward by entirely independent actors based upon their own reviews or analyses conducted without any direct input from either regulators or drug sponsors. This category includes everything from summaries in third-party drug databases to entries in drug compendia to comprehensive meta-analyses published in academic medical journals. Again, any of these, but particularly the latter, may be picked up, amplified, and/or modified by other communication channels so they reach a much wider audience than their initial dissemination.
In principle, however, all such messages—if their content proves valid—are ultimately supposed to be incorporated into the authoritative source of information about a drug’s risks and benefits, namely the drug label. This incorporation occurs through an ongoing process of dialogue, negotiation, and sometimes debate between sponsor and regulator. While under current law in the United States FDA has considerable direct power over the drug label—including the authority to require a sponsor to include new safety information, subject to specific triggers and criteria—no less an authority than the U.S. Supreme Court has determined that, as a legal matter, ultimate responsibility for the content of the label, as it relates to drug safety, rests with the sponsor, at least for branded drugs (Wyeth v. Levine 2009:570–71), as is described in greater depth in chapter 20 in this volume. This view of the world in effect bestows primary control of the communication of authoritative drug safety information upon the drug sponsor. In practice, however, the picture is considerably more complex.
III. THE STRUGGLE FOR CONTROL OF DRUG SAFETY COMMUNICATION
At the center of the traditional model of drug safety communication was a dyad: drug sponsor and health agency. In this model, the drug sponsor was primarily responsible for collecting and analyzing safety information and proposing related labeling changes or other safety updates, while the health agency played the role of neutral arbiter and gatekeeper to ensure that the presentation of safety information in these channels was accurate and timely.2 Of course, third parties (compendia, researchers, formulary managers) also engaged in drug safety communication, but their roles were largely subordinate to those of the two main players.
Today, while the health agency and the drug sponsor have more or less maintained their traditional roles—with a few interesting adjustments to be discussed following—the world around them has changed dramatically. A variety of third parties have emerged as powerful, at times even dominant, forces in the drug safety communication system. Academic and government medical researchers conduct increasingly ambitious analyses, reanalyses, and meta-analyses of drug safety data, or even clinical studies, and proactively communicate the results to the public and to prescribers, whether through publications, counter-detailing activities,3 or outright public health advocacy efforts (Loke et al. 2011; Nissen and Wolski 2010; Nissen and Wolski 2007). Independent drug information centers devote substantial resources to the production and dissemination of drug safety information completely outside the influence of sponsors and regulators.4 Major drug payers, such as government agencies and managed-care organizations, in effect build their own picture of a drug’s safety and risk–benefit balance on the basis of a wide array of such sources, among which the information offered by the drug sponsor itself is only one voice among several (Academy of Managed Care Pharmacy 2012).
This shift in the balance of power is in large part the product of a trend of increasingly vocal skepticism among critics of the traditional model, who may express doubts about the ability and willingness of both sides of the traditional dyad to act as effective stewards of drug safety information. These critics may portray the drug sponsor as commercially motivated to downplay, conceal, and delay the communication of accurate and up-to-date safety information (Doshi 2013). At the same time, these critics may accuse the health agency of effectively colluding with the drug sponsor in this behavior, partly out of the institutional imperative to defend its drug approval decisions and partly out of simple lack of resources (Goldacre 2013).
Though these kinds of allegations had surfaced before, the withdrawal of Vioxx (rofecoxib) in 2004 was interpreted in some quarters as reinforcing concerns about the traditional model’s ability to handle emerging drug safety issues. The catalyst for the withdrawal was the announcement of study results by both FDA and the sponsor confirming that the users of the drug experienced an increase in cardiovascular risk, including significant numbers of fatal heart attacks (Martinez et al. 2004; FDA 2004). Shortly after the withdrawal, the medical journal Lancet published a meta-analysis of rofecoxib studies on the basis of which the authors asserted that the drug should have been withdrawn years earlier (Jüni et al. 2004). Critics charged that the drug sponsor had hidden essential safety information and that FDA, instead of forcing the sponsor to face the truth, had dithered until the weight of the safety evidence became too heavy to ignore—thus needlessly exposing thousands of patients to unacceptable risk (Martinez et al. 2004; Senate Committee on Finance 2004; Berenson et al. 2004).
The Vioxx experience galvanized two main currents of action. The first was a push to reform FDA’s own structures for overseeing drug safety and to enhance the agency’s authority to force sponsors to address emerging safety issues more promptly. This initiative was encapsulated in a comprehensive 2006 Institute of Medicine (IOM) report on drug safety that called for a number of systemic reforms, including providing FDA with direct legal authority to order safety-related labeling changes and to require sponsors to conduct studies to address emerging safety signals (IOM 2007). The timing was auspicious, as FDA’s drug approval user-fee authority was due for legislative renewal the following year, providing a convenient legislative vehicle that allowed the IOM report’s major recommendations to be enacted into law relatively quickly (FDAAA, Title IX). In the wake of these legal changes, the balance of power within the traditional dyad of health agency and sponsor shifted distinctly toward the health agency.
This shift was bolstered by operational changes undertaken by FDA itself. Most prominently, the agency decided to move away from its historical caution about publicly communicating emerging drug safety information. If the traditional approach had been to wait for confirmation or near confirmation of a safety concern before communicating it, the new paradigm reversed that presumption. Henceforth, FDA would proactively communicate information about emerging safety issues at a markedly earlier stage—even at the risk of prematurely announcing a potential safety issue that might later prove unfounded. New structures, most notably the Drug Safety Oversight Board, were set up within the agency to provide a forum for diverse internal perspectives on safety issues and to promote effective oversight of drug safety communication (FDCA § 505-1(j); 21 USC § 355-1(j)).
The second major focus of action sparked by the Vioxx experience and other safety issues with marketed drugs—as well as a number of law enforcement proceedings in which major drug companies were accused of improperly concealing negative clinical trial results (Bowe and Dyer 2004; Attorney General of NY 2004)—stemmed from the view that it was necessary not merely to further empower the regulator side of the regulator–sponsor relationship but to actually open up the entire universe of clinical data to public scrutiny. In this view, only by allowing more or less unfettered access to those data would the entire health care community be able to properly judge for itself the risks, as well as the benefits, of the medicines being offered to patients.
Once again, as this perspective emerged and intensified in the United States in the mid-2000s, the 2007 reauthorization of the FDA user-fee law provided a convenient vehicle for a rapid legislative response. An entire title of that reauthorization bill was devoted to requiring drug sponsors to post clinical trial results on a publicly available governmental website, clinicaltrials.gov (FDAAA, Title VIII). Several U.S. states, unwilling to wait for the federal requirements to be implemented, put forward their own (and not always entirely consistent) versions of clinical data disclosure requirements (Me. Rev. Stat. Ann. tit. 22. § 2700-A).5
Almost overnight, it seemed, transparency became the new watchword (Zarin and Tse 2008; IOM 2013; Rabesandratana 2013). The U.S. trade association of innovative pharmaceutical companies, the Pharmaceutical Research and Manufacturers of America (PhRMA), adopted a set of “Principles on Conduct of Clinical Trials and Communication of Clinical Trial Results” that committed member companies to transparency through the timely submission and registration of all clinical trials involving patients on the database specified in the federal disclosure legislation, regardless of outcome (PhRMA 2009). The association noted in a joint position statement that it “recognize[s] that there are important public health benefits associated with making clinical trial information more widely available to healthcare practitioners, patients, and others” (PhRMA et al. 2009). Posting of clinical trial results quickly became the norm.
But relying on drug sponsors to post results of their own trials still did not go far enough for some transparency advocates. More recently, calls have emerged not just for study results but for actual patient-level data and full clinical study reports to be released to third parties so that they may perform their own analysis of the data. The EMA has announced a policy of providing patient-level data submitted in drug marketing applications to third parties upon request and has indicated that it will affirmatively publish full clinical study reports submitted by drug sponsors.6 PhRMA and the European Federation of Pharmaceutical Industries and Associations (EFPIA) (2013) outlined their commitment to provide clinical trial information to the public, patients, and private researchers when they released joint “Principles for Responsible Clinical Trial Data Sharing.” This document supplements PhRMA’s “Principles on Conduct of Clinical Trials and Communication of Clinical Trial Results” (PhRMA 2013). In addition, a number of major pharmaceutical companies, including GlaxoSmithKline, Roche, Pfizer, and Johnson & Johnson, announced their own specific data-sharing platforms (Kmietowicz 2014). From the perspective of drug safety communication, however, the most pressing issue posed by the public or semipublic release of these kinds of information is the specter of unrestricted reanalyses of clinical data by third parties that may lead to disparate results and potentially conflicting messages to the health care community about the risks and benefits of a drug product (EMA 2012): “[I]f inappropriate data analyses were picked up by the media early in a product’s lifecycle, it could confuse doctors and prevent patients from receiving effective medicines” (statement of Susan Forda, Eli Lilly& Co.). To address this concern, some observers have called for health agencies to issue quality guidelines or standards to govern third-party reanalysis of clinical data.
Once again, it seems, the issue of trust lies at the heart of the matter—pitting those who believe public health is best served by leaving the formulation and communication of drug safety information primarily in the hands of those responsible for drug development and approval, i.e., sponsors and regulators, against those critics who believe that those parties cannot be relied upon to consistently provide accurate and timely drug safety information. Viewed through this lens, the drug safety communication debate also reflects broader social currents around the control and dissemination of information. Advocates of an open-source or open-science model call for a more transparent approach to all kinds of scientific data and information (Ross and Krumholz 2013). In their view, claims to proprietary ownership of scientific data must generally give way to the rights of the broader community to share in the understanding, analysis, and dissemination of those data. Assertions of superior expertise on the part of drug sponsors or regulators, in this perspective, are trumped by the alleged bias and self-interest of those parties, which supposedly render them incapable of acting as full fiduciaries of the public trust with respect to scientific data of interest to the public. Implicit in this outlook is a view of scientific data as a public or common good, in contrast to the view that those who devote significant resources and expertise to the production of those data—i.e., drug sponsors—or those who are legally entrusted with reviewing and overseeing the use of those data—i.e., regulators—have a legitimate role as a primary (though not necessarily the sole) source for analysis and dissemination of the data.
We are rapidly entering a world in which more and more drug safety information, from more and more sources, will be available to more and more audiences more and more quickly—as some have put it, a world in which we will be “awash in data.” As the sources traditionally relied upon for an authoritative interpretation of that information come under increasing challenge, how are practitioners, patients, and other concerned audiences to determine which drug safety messages are validated and which are unconfirmed or even questionable—to separate fact from fiction? The public debate thus far has been dominated by criticism of drug sponsors and regulators. In fact, however, other potential sources of information in the system are not necessarily free of their own self-interest and potential bias. Government and private payers and their agents may have an inherent interest in lowering short-term out-of pocket drug costs—potentially at the expense of other types of health care costs that might have been saved or avoided by using the drugs in question. Plaintiffs’ lawyers and their allies may have an inherent interest in undermining confidence in drug sponsors’ safety messages or in advancing safety theories that support their litigation interests. In contrast to drug sponsors, these potential sources of bias may not be as apparent.
The goal of drug safety communication—supporting informed choices about the use of medicines that lead to optimal patient outcomes—is not an easy one to attain. As we have seen, choices about how best to promote effective drug safety communication are tied to much broader perspectives about the proper roles of government, industry, prescribers, patients, and independent researchers in creating, communicating, understanding, and acting upon drug safety information. Driven by these perspectives, a new model of drug safety communication is taking shape—one that elevates transparency over certainty and that opens the door for multiple and at times conflicting messages about matters that may be critically important to prescribing decisions and patient care. If this model is to achieve its basic purposes, there is an urgent need for a comprehensive and balanced consensus around system governance, data quality standards, and effective communication processes to support the fundamental public health goals that are at stake. Achieving such a consensus will require not just hard work but also a forthright acknowledgment of the differing perspectives and institutional imperatives that all involved parties bring to the process.
NOTES
The author gratefully acknowledges the assistance of Mark Odynocki and Rachael Martins of the Pfizer Legal Division in the research for this chapter. The views expressed in this chapter are those of the author only and do not reflect the positions of Pfizer Inc or any of its affiliates.
1. A third traditional source of safety information, literature reports, is more in the nature of an adjunct (though at times an important one) to these two main categories.
2. The health agency of course also played the “meta” role of regulating the overall process of pharmacovigilance, but that role is less relevant to this discussion.
3. Counter-detailing refers to the practice of representatives of academic institutions or government agencies affirmatively calling on physicians or payers to provide ostensibly independent information about a drug so as to “counter” traditional sales or “detailing” activities of drug company sales representatives.
5. Clinical trial disclosure requirement repealed by Maine’s legislature.
6. In a December 2013 press release, EMA to Push Ahead in 2014 Towards Publication and Access to Clinical Trial Data, the agency noted that its policy and implementation plan for access to clinical data would be discussed at a March 2014 management board meeting.
REFERENCES
Academy of Managed Care Pharmacy. 2012. AMCP Format for Formulary Submissions: A Format for Submission of Clinical and Economic Evidence of Pharmaceuticals in Support of Formulary Consideration, Version 3.1.
Attorney General of New York. 2004. In re GlaxoSmithKline, Assurance of Discontinuance.
Behrman R. E., J. S. Benner, J. S. Brown, M. McClellan, J. Woodcock, and R. Platt. 2011. “Developing the Sentinel System—a National Resource for Evidence Development.” New England Journal of Medicine 364:498–99.
Berenson, A., G. Harris, B. Meier, and A. Pollack. 2004. “Despite Warnings, Drug Giant Took Long Path to Vioxx Recall.” New York Times (November 14).
Bowe, C. and G. Dyer. 2004. “GSK Settles Paxil Lawsuit with Spitzer.” Financial Times, August 26.
Doshi, P. 2013. “Putting GlaxoSmithKline to the Test Over Paroxetine.” British Medical Journal 347:f6754.
Goldacre, B. 2013. Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients. London: Faber and Faber.
Institute of Medicine (IOM), Board on Health Sciences Policy, Board on Health Care Services. 2013. Sharing Clinical Research Data: Workshop Summary, Rapporteurs. Steve Olson and Autumn S. Downey. Washington, D.C.: The National Academies Press.
Institute of Medicine (IOM), Committee on the Assessment of the U.S. Drug Safety System, Board on Population Health and Public Health Practice. 2007. The Future of Drug Safety: Promoting and Protecting the Health of the Public, ed. Alina Baciu, Kathleen Stratton, and Sheila P. Burke. Washington, D.C.: The National Academies Press.
Jüni, P., L. Nartey, S. Reichenbach, R. Sterch, P. A. Dieppe, and M. Egger. 2004. “Risk of Cardiovascular Events and Rofecoxib: Cumulative Meta-analysis.” Lancet 364:2021–29.
Kessler, D. A. 1993. “Introducing MEDWatch: A New Approach to Reporting Medication and Device Adverse Effects and Product Problems.” Journal of the American Medical Association 269:2765–68.
Kmietowicz, Z. 2014. “Johnson & Johnson Appoints Yale Project Team to Run Data Sharing Scheme.” British Medical Journal 348:g1361.
Loke, Y. K., C. S. Kwok, and S. Singh. 2011. “Comparative Cardiovascular Effects of Thiazolidinediones: Systematic Review and Meta-Analysis of Observational Studies.” British Medical Journal 342:d1309.
Martinez, B., A. W. Matthews, J. S. Lublin, and R. Winslow. 2004. “Merck Pulls Vioxx from Market After Link to Heart Problems.” Wall Street Journal (October 1).
Nissen. S. E. and K. Wolski. 2010. “Rosiglitazone Revisited: An Updated Meta-Analysis of Risk for Myocardial Infarction and Cardiovascular Mortality.” Archive of Internal Medicine 170:1191–1201.
——. “Effect of Rosiglitazone on the Risk of Myocardial Infarction and Death from Cardiovascular Causes.” New England Journal of Medicine 356:2457–71.
Pharmaceutical Research and Manufacturers of America (PhRMA). 2013. EFPIA and PhRMA Release Joint Principles for Responsible Clinical Trial Data Sharing to Benefit Patients.
PhRMA and European Federation of Pharmaceutical Industries and Associations (EFPIA). 2013. Principles for Responsible Clinical Trial Data Sharing.
PhRMA, EFPIA, Japan Pharmaceuticals Manufacturers Association and the International Federation of Pharmaceutical Manufacturers and Associations. 2009. Joint Position on the Disclosure of Clinical Trial Information via Clinical Trial Registries and Databases.
Rabesandratana, T. 2013. “Drug Watchdog Ponders How to Open Clinical Trial Data Vault.” Science 339:1369–70.
Ross, J. S. and H. M. Krumholz. 2013. “Ushering in a New Era of Open Science Through Data Sharing: The Wall Must Come Down.” Journal of the American Medical Association. 309:1355–56.
Senate Committee on Finance. 2004. FDA, Merck, and Vioxx: Putting Patient Safety First? Hearing Before the Senate Committee on Finance, United States Senate, 108th Congress, 2d sess. (November 18).
——. 2013. Drug Safety Communication: FDA Investigating Reports of Possible Increased Risk of Pancreatitis and Pre-Cancerous Findings of the Pancreas From Incretin Mimetic Drugs For Type 2 Diabetes. (March 14).
——. 2012. Draft Guidance, Drug Safety Information—FDA’s Communication to the Public.
——. 2010. The Sentinel Initiative: Access to Electronic Healthcare Data for More Than 25 Million Lives.
——. 2004. FDA Issues Public Health Advisory on Vioxx as Its Manufacturer Voluntarily Withdraws the Product.
U.S. Food and Drug Administration, Risk Communication Advisory Committee. 2009. Charter of the Risk Communication Advisory Committee to the FDA.
Wyeth v. Levine, 555 U.S. 555 (2009).
Zarin, D. A. and T. Tse. 2008. “Moving Towards Transparency of Clinical Trials.” Science 319:1340–42.
