At the time that a drug is approved by the US Food and Drug Administration (FDA) for sale in the marketplace, uncertainties necessarily remain about the drug’s benefits and risks. The research that is conducted before a drug’s approval is limited in the numbers and types of patients who are involved and in the length of time that patients’ experiences with the drug can be monitored (Borer et al., 2007; Hiatt, 2006; IOM, 2007; Ray and Stein, 2006). Ensuring that drugs continue to have an acceptable benefit–risk profile after they are approved for sale on the US market is as important to FDA’s public health mission as ensuring the acceptability of the benefit–risk profile before it is permitted to enter the market. In support of the equal public health importance of regulatory oversight of drugs before and after approval, the authorities granted to FDA by the Food and Drug Administration Amendments Act (FDAAA) of 20071 provide FDA with the tools it needs to adopt a comprehensive lifecycle approach to the assessment of the benefits and risks associated with marketed drugs.
In the lifecycle approach, responding in a timely and responsible way to safety signals that emerge after a drug is on the market is among the most important and challenging public health responsibilities of FDA. Permitting a drug that is on balance harmful to stay on the market threatens public well-being, but so does limiting access to a drug whose benefits outweigh its harms. FDAAA provides FDA with greater statutory authority in the postmarketing setting than it had before, including the authority to require manufacturers to conduct studies of drugs in the postmarketing setting. That authority, however, presents a number of new challenges to the agency, including determining when it is appropriate for
1Food and Drug Administration Amendments Act of 2007, PL 110-85, 121 Stat. 823 (2007).
FDA to require a postmarketing study and what types of studies to require when that is the case, how best to protect the rights and interests of patients who serve as participants in the research that it requires, and how it should use the information from the required postmarketing studies and from other available research (for example, studies initiated by academic researchers) in making regulatory decisions. The present committee’s charge to evaluate the scientific and ethical issues involved in conducting studies of the safety of approved drugs reflects those challenges.
How should FDA factor in different kinds of safety evidence in considering different kinds of regulatory actions?
In response to this question, the committee notes that no single algorithm can determine how to factor different kinds of safety evidence into regulatory decision-making, but does specify processes and principles to guide how this should occur. The committee identifies five actions, discussed below, that FDA can take to improve its decision-making processes in response to different kinds of safety evidence: (1) adopt a specified decision-making framework; (2) create a Benefit and Risk Assessment and Management Plan (BRAMP) document for each drug that is maintained across the drug’s lifecycle; (3) characterize the nature of any disagreements about the evidence of benefits or risks; (4) create effective multidisciplinary teams with wide-ranging expertise, including in observational study design and interpretation, outcomes research and pharmacoepidemiology, Bayesian methods and modern causal inference approaches, and (5) adhere to the principles of reproducible research.
The committee proposes that FDA use a three-stage framework—adapted from 2009 Science and Decisions: Advancing Risk Assessment (NRC, 2009) and consistent with a framework recommended by A Risk-Characterization Framework for Decision-Making at the Food and Drug Administration (NRC, 2011)—any time in the lifecycle of a drug that FDA needs to make a regulatory decision, and for planned reviews of regulatory decisions. Given its charge, the committee focuses on the use of the framework in the postmarketing setting where it could be employed, for example, when the emergence of a serious safety signal may
2The committee presents the questions in the order they are discussed in the previous chapters, not the order they are presented in the charge.
precipitate or require a regulatory decision, including the reaffirming of the drug’s current regulatory status. The three stages of the adapted framework include (1) define the public health questions of importance, (2) assess the drug’s benefits and risks, and (3) make, communicate, and implement the regulatory decision. The three-stage framework is designed to be broadly applicable to regulatory decisions, to support decision-makers’ judgment, and to facilitate the resolution of disagreements about the scientific evidence and the best regulatory actions to protect public health. For FDA’s regulatory decisions about approved drugs to be ethical and appropriate, FDA needs to consider the perspectives of patients; and the concerns of consumers, health care providers, and industry; securing this input is an important element of the proposed framework.
Establishing and maintaining a BRAMP document for each drug throughout its lifecycle would also enhance FDA’s ability to respond appropriately to safety evidence. The document would summarize the benefits and risks of the drug, the rationale for FDA’s decisions in light of those benefits and risks, and how any risks will be managed throughout the drug’s lifecycle. The BRAMP document, as proposed by the committee, is designed to support the systematic implementation of the lifecycle approach to regulatory oversight of drugs, to foster collaboration between FDA and drug sponsors in that oversight, and to increase the transparency of FDA’s decisions. Because the benefit–risk profile of a drug can change over time, the BRAMP document would become a living document that is updated when there is new information that warrants re-evaluation of the drug’s benefit–risk profile. Each update would include summaries of the three stages of the decision-making framework discussed above and any plans for identifying or managing risks (such as a risk evaluation and mitigation strategy). In the premarketing setting, the drug sponsor would provide initial information about the benefits and risks of a given drug, uncertainties in the information relevant to the public’s health, and detailed plans to decrease those uncertainties if they exist. FDA should review and finalize the BRAMP document. In the postmarketing setting, FDA staff who did not play a primary role in the drug’s approval process and who have expertise in surveillance, epidemiology, and the evaluation of safety data collected from different observational and clinical trial designs, would review and modify the BRAMP document at pre-specified intervals throughout the lifecycle of the drug and when new information warrants re-evaluation of the drug’s benefit–risk profile.
Disagreements among experts about scientific evidence lead to some of the more challenging regulatory decisions. When such disagreements occur, it is important for FDA to characterize the nature of the disagreements. These can occur because experts have different prior beliefs about the plausibility of a given benefit or risk in light of prior evidence, different views about the quality of the studies supplying the evidence or about the relevance of the new evidence to the public health question that calls for a regulatory decision, or different ideas about
how to synthesize all the available evidence relevant to the public health question or about the threshold of certainty needed to justify concern or regulatory action.
Bayesian approaches to measuring the strength of evidence and to characterizing the uncertainty of scientific conclusions about the presence or absence of a drug benefit or risk can be enormously useful in decision-making, which should incorporate the chances of being wrong and the attendant consequences in the choice of regulatory option. Standard approaches to statistical analysis cannot provide those inputs.
Outside researchers can be key partners with FDA in identifying safety concerns, and FDA can greatly augment its own efforts in the safety arena by allowing the research community to be more fully engaged. FDA should explore, seek support for and implement practices that enhance the ability of the external community of scientists to both identify drug-safety issues and to assess the validity of FDA’s attempts to do the same. These include policies and practices that contribute to transparency, reproducible research, and sharing of data from both the premarketing and postmarketing contexts. Few studies currently follow FDAAA requirements to publish even summary results in ClinicalTrials.gov within a year of drug approval; enhanced compliance with these requirements can facilitate that engagement.
What are the strengths and weaknesses of various approaches, includingobservational studies, including patient registries, meta-analyses, including patient-level data meta-analyses, and randomized controlled trials, to generate evidence about safety questions?
The strengths and weaknesses of the many ways to explore drug-safety questions depend critically on context-specific facts, priorities, data sources and the nature of the benefits and risks being considered. Whether an adverse event is rare or common, mild or serious, and known or unknown, and whether an anticipated drug effect is small or large could dramatically change the relative advantages of various designs. For example, the value of an observational study of a harm based on existing data depends on whether the harm was reliably recorded in the dataset being used. A clinical trial too short to find a delayed effect is going to provide less relevant safety evidence than a design based on patient registry data with long follow-up. The invocation of broad principles that are inapplicable to a specific case (for example, that randomized controlled trials [RCTs] always provide the best evidence) can sometimes impair the investigation of drug harms. In any specific case, regulators need to have the input of a wide variety of experts who can help to make context-specific judgments.
The committee does, however, outline some general considerations that are important for evaluating the value of various designs for decision-making purposes. The initial set of considerations is how strong the safety signal is that motivates the design, and whether it primarily involves an elevation in risk, a
decrease in benefit, or both, for either the general population or a definable subgroup. Second is how time-urgent is the need for a regulatory response, based on the nature of the safety signal. The third involves how large the change in risks or benefits must be, on both relative and absolute scales, to justify a regulatory response. Fourth is what the other causes of a given adverse event (or failure of benefit) might be, and how strongly they are predictive. Fifth is the quality of data likely to be gathered as part of any given design on drug exposure, outcomes, confounders and other relevant patient, disease or contextual characteristics. Sixth is a judgment of how study design, conduct or context is likely to affect the transportability of the study results. Seventh is what the logistical requirements of a design will be, including data access, cost and feasibility. Finally, there are considerations of ethical burden, consent, confidentiality, and study oversight. These factors can lead to the choice of either a single design type or a combination of studies with counterbalancing strengths and weaknesses.
With the above considerations in mind, the committee made some general observations about the strengths and weaknesses of specific designs. The RCT is considered the gold standard for studies of a drug’s benefits because of the ability of randomization to control for potential biases and confounders, both known and unknown. Although the committee agrees that a well-conducted, high-quality RCT has many theoretical advantages over other study designs, it also recognizes that what can be achieved in practice in assessing safety endpoints can fall short of the ideal. Noncompliance, cross-over and dropout, limitations in study size or duration, failure of the study population or procedures to adequately represent circumstances in the general population of users, and the realization that safety endpoints are sometimes unforeseeable and cannot always be specified in advance can decrease the advantages of RCTs over observational studies for evaluating the risks posed by approved drugs. In many cases, the latter may provide estimates closer to the actual risks in the target population if one considers the combination of bias, precision, and transportability of results.
In addition, because RCTs alter a patient’s clinical experience, they may entail more ethical complications than observational studies. (That said, as part of the consent process, the information patients receive about benefits and risks of study treatment options, as well as alternative treatments, may be more complete than a typical health provider supplies.) Other disadvantages of RCTs are the cost and time required to conduct such studies; the duration of studies is particularly problematic when an urgent public health question needs to be answered. An advantage of an RCT, however, is the ability to ascertain moderate relative risk elevations of common outcomes with confidence. A small relative risk (for example, RR <1.5) increase in a common outcome (for example, MI) may represent a very large absolute increase in risk with great public health importance. The adequacy of confounding control in many observational designs may not be sufficient to estimate such risk elevations with high confidence. Additionally, RCTs have the potential ability to assess both benefit and risk in the same group
of patients at the same time. While this is not frequently done, it sometimes can be necessary to find subgroups in which the benefit–risk balance may be unacceptable, to make finely-grained assessments of the benefit–risk balances when new risks arise, or to make fair benefit–risk comparisons with active comparators.
Observational studies can provide data on a large number of people under real-world conditions. They also typically have greater heterogeneity of participants and may be more likely to detect drug–drug interactions and adverse effects in populations that might not have been included, or specifically excluded from premarketing RCTs. Observational studies are more prone to confounding than RCTs but often have better transportability of results (that is, external validity or generalizability) to those populations that might not be included in RCTs. They are generally less prone to confounding for safety endpoints than they are for effectiveness endpoints, particularly when the harms were unintended or unsuspected at the time the drug was prescribed and don’t share a common mechanism with benefit (Psaty and Vandenbroucke, 2008). Most important is the magnitude of the relative elevation in risk in relation to the potential for confounding; if the anticipated relative elevation in risk is quite large, beyond plausible degrees of confounding, observational designs with weak confounding control can be sufficient. As previously noted, modest relative risks, particularly those less than 1.5, can require substantial control of confounding that might only be achievable in a clinical trial. One way to ameliorate this problem is to conduct multiple observational studies with a variety of designs and data sources unlikely to share similar biases.
If observational studies can be based on existing data or can use data systems that are already in place, they typically are less expensive and, unless a drug is new to the market and is not in widespread use, can be conducted more quickly than RCTs. If the availability and quality of electronic medical records and other electronic data sources increase, the quality of information and the ability to identify and control for potential confounders will improve, and the cost and time needed to complete a study might decrease. In addition, observational studies, which by their nature do not interfere with the treatments that people would receive in the course of regular care, generally have fewer ethical complications than RCTs.
In meta-analysis, data from a number of studies—either RCTs or observational studies—are combined, in aggregate or at the individual patient level. Meta-analyses are observational studies that use other studies as the unit of analysis. Their advantages include the speed with which they can be conducted, the use of existing data with few ethical issues, increased statistical power, and the ability, because a large number of participants can be included from the pooling of data, to detect adverse events or groups at risk. Other than the larger sample size, however, the same limitations and biases of the underlying observational and randomized trials persist in meta-analyses, and publication and reporting bias may jeopardize the validity of meta-analyses that use only published studies. Finally, biases are potentially incurred by the criteria for study selection. FDA
can improve the validity of later meta-analyses by providing, early in the postmarketing phase, guidance on common data definitions and other design features that will make subsequent safety research conducted by others more likely to be mutually informative and combinable.
The committee looked specifically at noninferiority and superiority studies; the former are increasingly used in analyses of safety. Noninferiority studies evaluate whether a new treatment is “no worse” than a previous, accepted treatment by a specified margin or, in the case of safety studies, poses no more than an “acceptable” excess risk of adverse effects compared with the accepted treatment. The definition of acceptable often implies a tradeoff against a known benefit, but the comparison must be made explicit in interpreting such studies. Superiority studies evaluate whether a new treatment performs better than a previous, accepted treatment or, in the case of safety studies, poses less risk of adverse effects than the accepted treatment. One concern with noninferiority studies is the consequence of poor study conduct. Poor study conduct that leads to data of poor quality may introduce bias toward no effect, that is, lead to an erroneous conclusion that there is no difference between the two treatments. In a noninferiority study, that erroneous conclusion may be incorrectly interpreted as supporting the claim that the risk of adverse effects is the same for both treatments. When interpreting noninferiority and superiority studies, it is important that FDA evaluate the magnitude of the differences between the drugs and not rely on the study’s preset designation of what constitutes acceptable inferiority or sufficient evidence of superiority. Perhaps more important is for FDA to develop and implement performance standards for the conduct, analysis and interpretation of noninferiority studies for safety.
Finally, the committee found that it is critical to recognize that the analytic approach, not just design, is an important contributor to the strength of evidence provided by any study. For example, the use of causal inference and Bayesian methods—with sensitivity analyses and proper treatment of missing data—can produce estimates of benefit, risk, and the uncertainty associated with those estimates, that differ from estimates derived with standard frequentist approaches. Intention-to-treat approaches that are appropriate for the assessment of relative efficacy, may not be appropriate for the assessment of risk. Given the importance of using the optimal analytical technique to reap the advantages of various designs, bringing together teams that have broad and deep technical expertise in both the design and analysis of drug-safety studies is integral to having the best evidence to help answer the public health question.
Considering the speed, cost, and value of studies, what types of follow-up studies are appropriate to investigate different kinds of signals (detected pre-approval or post-marketing) and in what temporal order?
The optimal follow-up studies to investigate different safety signals, and the order of those studies, will depend on the specific circumstances of the safety
signal. The committee does provide general guidance to FDA in making these determinations, beyond the statutory provision in FDAAA,3 that permits FDA to require a clinical trial only if sufficient information cannot be obtained with an observational study, a presumption that the committee finds consistent with FDA’s ethical obligations to research participants in the postmarketing context and to the public’s health.
First, all research strategies will work best if anticipated and planned for early. As outlined elsewhere in this report, there are a number of characteristics that should signal heightened concern about the possibility that harm will outweigh benefit in the postmarketing context. Those characteristics might appear in the case of drugs that were approved on the basis of surrogate endpoints when different surrogate endpoints yield conflicting evidence about clinical effect or safety; drugs that are first-in-class and were validated on the basis of surrogate endpoints with drugs in a different class; drugs about which safety signals appear in premarketing data or postmarketing surveillance when there is a substantial public health concern, drugs where a severe adverse event is seen, or there is a strong biologic rationale for a particular adverse effect; drugs that are expected to have a different benefit–risk profile in a particular group or under real-world conditions; drugs in a class about which a substantial safety signal has previously been identified; and drugs of which evidence of a lack of benefit emerges in the postmarketing setting.
The earliest and easiest step that FDA can take is to ensure that it is making maximum use, perhaps through meta-analysis, of the data already in its possession, which often would have been submitted as part of a New Drug Application, or may pre-exist because of studies performed for a different indication. Recognizing that much postmarketing safety information will come from studies not specifically commissioned by the agency, through the BRAMP FDA can define how exposures, covariates, and outcomes are to be assessed in future safety studies conducted by industry or by independent researchers. That information is also available to other investigators. The agency can be yet more effective by bringing together researchers at the time of or shortly after approval to standardize various design dimensions to make studies appropriate for future data pooling, a process known as prospective or collaborative meta-analysis.
If at the time of approval it is judged that FDA must require new studies, the postmarketing study strategy should be incorporated into the BRAMP. A variety of study designs can be used early in the introduction of a drug that are much more difficult or impossible to implement once it is in wide use, and this underscores the need to plan and start such studies as early as possible. Early initiation of such studies can also allow longer followup, so both prospective observational studies and RCTs are more likely to provide necessary evidence close to the time
321 USC § 355(o)(3)(B) (2010).
when public health decisions have to be made, if postmarketing safety signals are indeed found.
This last point is critical for RCTs in that one of their main disadvantages cited is that they take too long. An alternative perspective is that they are started too late. If an RCT is considered only when very strong suspicion of a safety signal arises, it may be too late to initiate one, particularly if the adverse events occur long after exposure. In addition, some of the ethical difficulties that arise in trying to conduct an RCT of a widely used drug are minimized if the RCT is initiated soon after market introduction. In 1999, when rosiglitazone and pioglitazone had just been approved, the TIDE trial would have been an important, timely, and well-designed study. Because at that time there would have been populations of patients who were not on the drug, individual or cluster randomization poses fewer logistical and ethical difficulties than would exist later. However, because of FDAAA restrictions, a study could only be required when there were premarketing signals of a potential for serious risk (for example, adverse lipid alterations, or low frequency serious adverse events) and observational studies were deemed inadequate. Early initiation of postmarketing investigations can dramatically change both the scientific value and the ethical calculus of such studies, making the optimal sequence dependent on when in the lifecycle of a drug the studies are being contemplated.
One key determinant of the kinds of designs that might be considered is whether the safety signal is a harm to be offset by a known benefit or the harm is a failure to provide expected benefit, either overall or in specific populations. FDAAA defines the latter as a safety concern, or more specifically “any failure of expected pharmacological action of the drug”.4 A failure to provide expected clinical benefit is most likely to be observed if that benefit was not directly tested in the approval process, for example, if surrogate endpoints were used, or if non-responsive subpopulations were not well represented in premarketing trials. If the safety concern focuses on the issue of no or reduced benefit, an RCT is more likely needed because confounding by indication in observational studies of drug benefit can be difficult to overcome. Conversely, if the harm is distinct from the mechanism of benefit, is unforeseen and not strongly linked to patient or disease characteristics, or is strongly linked to conditions of general practice (such as cotreatments or inconsistent monitoring), observational studies can often provide sufficiently reliable evidence related to risk. As previously noted, that degree of sufficiency also depends on the relative degree of increase in the risk that is deemed important to detect, weighed against the likely magnitude of confounding.
The dimensions of quality that must be judged for each combination of study design, data source, and analytic approach are the precision, bias, and transportability of the result. Each of those contributes to the observed effect’s potentially
421 USC § 355-1(b) (2010).
differing from the true effect in the general population of patients taking the drug. The intrinsic design qualities must be weighed with extrinsic issues, such as the time that it will take for a given design to deliver a result, the cost and complexity of various designs, and the ethical dimensions of the study. Some of the ethical dimensions depend on how uncertain the benefit–risk profile of the drug in question is, so properly assessing the uncertainty before the studies begin is quite important.
At every step in the process, FDA is faced with the choice between making a decision on the basis of evidence already gathered, and waiting for more or higher-quality evidence. If there is strong evidence of a safety problem, or new evidence about the benefit–risk balance for the population or a definable subset of it, FDA’s decision can be extraordinarily difficult. Methodologies such as Bayesian analyses or other approaches to incorporate prior relevant information with newly emerging information could provide decision-makers with better quantitative assessments of evidence. An example would be through sensitivity analyses of clinical trials data that illustrate the influence of prior probabilities on estimates of probabilities that the intervention induces unacceptable safety risks. These insights can help enlighten judgments, allowing for more rational decision-making, and permitting input from multiple stakeholders and experts.
What are the ethical and informed consent issues that must be considered when designing randomized clinical trials to evaluate potential safety risks?
An assessment of the ethics of FDA’s requiring a postmarketing RCT is inextricably intertwined with an assessment of the science related to the underlying public health question and regulatory decision. There are circumstances in which FDA is ethically justified in requiring a postmarketing clinical trial, and there are circumstances in which a clinical trial is required by statute.
When a trial is not required by statute, a decision to require a trial to resolve a postmarketing benefit–risk profile question should be based on the determination that (1) uncertainty about the benefit–risk balance is such that a responsible decision about the future regulatory status of the drug cannot be made on the basis of existing evidence, or evidence that could be obtained from new observational studies; (2) an RCT can be properly designed and implemented to reduce uncertainty about the benefit–risk balance sufficiently to inform a responsible regulatory decision; (3) FDA will use trial results in making a regulatory decision in a timely fashion; and (4) the RCT can be carried out in a manner that provides sufficient protection of and respect for research participants.
In making the fourth determination—that the RCT can be carried out in a manner that provides sufficient protection of and respect for research participants—FDA must attend to multiple considerations, including whether the trial should be designed to include an active medical intervention as the comparator
(in contrast, for example, with a placebo) and issues of consent. The ethics of selecting an appropriate comparator for an FDA-required RCT are discussed under Question 4, below.
Informed consent obligations may be especially salient in the context of required postmarketing trials because patients may be asked to submit to a drug regimen about which a safety signal has prompted concerns about risk, and potentially about the acceptability of the drug’s benefit–risk profile. FDA should work with manufacturers, investigators, and institutional review boards (IRBs) to ensure that the following occur as parts of the informed consent process:
• Information is provided about why a new study is required, particularly to persons already taking the drug who might have to undergo a change in regimen as a result of study participation. Prospective research participants need to understand why additional research is important even though the drug they are currently taking was found by FDA to have a favorable benefit–risk profile on the basis of existing evidence and why it is reasonable to ask them to consider participating in the study.
• Special care is taken to ensure that prospective participants understand the potential risks of study participation in the postmarketing context. When a substantial amount of information indicating that a drug to be studied may involve serious risks has already accumulated, there is a heightened obligation to ensure that potential participants understand the risks posed by study enrollment. At a minimum, the disclosure of risks should include any boxed warnings, the “major statement” currently listed in direct-to-consumer advertisements, any formal conclusions about adverse effects made by FDA staff or an FDA advisory committee, and a summary of evidence from published peer-reviewed studies or relevant, quality studies submitted to FDA. Special efforts should be made to ensure that people who have low health literacy or educational attainment, who have shown poorer understanding of disclosed information on consent forms (Flory and Emanuel, 2004; Kass and Taylor, 2008; Lindegger et al., 2006), understand this and other study information.
• In addition to considerations of benefits and risks, people who are considering participation in research need to know how the care that they will receive in a protocol may differ from the care that they would ordinarily receive. Thus, information about “alternatives to participation” should convey the current standard of care for the health condition that the study drug targets. That is particularly crucial in cases in which medical practice has shifted away from prescribing the study drug because accumulating evidence from passive surveillance, observational studies, and small trials or meta-analyses suggests that another therapy is as effective and has a more favorable benefit–risk profile. It should be communicated in this situation that a potential participant who does not enroll in the trial
is more likely to have a different drug prescribed. If clinical practice continues to shift during the trial period, the latter statement should be strengthened; researchers have an ethical obligation to disclose all new developments that may affect a person’s willingness to continue to participate in a research study.
Comprehensive informed consent processes can help to ensure that trial participants understand the potential consequences for them of study participation, in addition to what they are contributing to the advancement of public health in the regulatory arena. These processes cannot, however, serve as exclusive or sufficient ethical justification for conducting a postmarketing trial. The other conditions for initiating a trial should be independently satisfied. People should not be asked to assume risks that are not justified by the potential benefits of the trial to participants or society. Particularly in research settings in which participants have low literacy, low income, and poor access to modern health care and medicines, even a robust consent process may do little to countervail the pressures that lead people to participate in research.
Informed consent and other ethical considerations become more complex as the clinical risks to participants increase and the clinical benefits decrease. In making the determination that an RCT that FDA is considering requiring can be carried out in a manner that provides sufficient protection of and respect for research participants, FDA must always balance its ethical obligations to protect the public from unsafe drugs with its ethical obligations to safeguard the rights and interests of people who participate in research supporting the agency’s decisions about drug benefits and risks. Difficult choices must be confronted when the study design that seems to offer the greatest potential for obtaining knowledge relevant to the public health question also involves the greatest burden on and risks to research participants.
If uncertainty about the benefit–risk profile of a marketed drug exists, there may be circumstances in which it is ethically acceptable to ask patients to participate in an RCT that exposes them to risks that are not likely to be outweighed by any prospect of clinical benefit to them and that are readily avoidable with treatment options available to patients outside research participation. These circumstances may be a satisfied when a question of pressing public health importance cannot be properly answered without the conduct of the study, the study may be appropriately designed to provide high-quality evidence that is needed to answer the question, and other conditions that are intended to safeguard the rights and interests of participants can be satisfied. Those safeguards should include but are not limited to:
• Determination by an appropriately constituted review committee that the additional net risk is small enough for it to be ethical to ask people
whether they are willing to accept the risk solely to contribute to the public good.
• Minimization of additional net risk by careful study design and implementation of a robust monitoring plan throughout the study.
• Inclusion of special measures in the process of soliciting informed consent to confirm that patients understand and willingly accept that they are assuming an additional net risk—beyond what they are likely to face in clinical practice—solely in the interest of the public good.
• Implementation of processes to ensure that over the course of the trial participants are regularly informed of any changes in clinical practice or the medical literature that are relevant to assessments of the comparative benefits and risks of trial participation and (non-research) clinical management.
External IRBs and data monitoring committees (DMCs) overseeing FDA-required postmarketing RCTs should have all the information necessary to ensure that the trials they oversee are ethically acceptable and adequately monitored. To that end, FDA should provide all relevant IRBs (centralized and multiple IRBs) and DMCs with sufficient information to permit appropriate continuing oversight of the RCT in accordance with their roles. That should include information about the public health question at issue, the specifics of the study design that it has deemed suitable to address the question—including any design features that it views as necessary for the ethical justification of the study, and any changes in clinical practice or professional standards that arise over the course of the RCT that might affect the benefit–risk profile of the drug and influence a person’s decision about whether to continue to participate.
Under what circumstances should head-to-head randomized clinical trials for safety be required?
The committee’s answer to this question assumes that it has already been determined, according to the criteria and processes outlined elsewhere in this report, that it is appropriate for FDA to require a postmarketing study and that this study should be an RCT.
A head-to-head trial involves a comparison of two active treatments that are both indicated for the same patients who have the same condition. The committee considered study designs in the context of a public health question about the benefits or risks associated with a drug. The public health question is most likely to be addressed by comparing the drug at issue with the therapies likely to be used if the drug were removed from the market or its use were restricted; that is the decision-relevant public health question. However, for such a study to be scientifically valid and ethical, the active comparator must have a well-defined benefit–risk profile and be a clinically acceptable alternative. The dose of the comparator
needs to be carefully defined so neither the benefits nor risks differ appreciably from what would be expected in common use. Unless precluded by toxicity or tolerability, it would be expected that the dose of the comparator should be at least equal in effectiveness to the target agent. If no comparator treatment exists or no comparator has a well-defined benefit–risk profile, then typically at least one arm of the study should be some form of “usual care” or a placebo if usual care is not a proven or active treatment. If there are ethical reasons for not having a usual-care or placebo arm in the study—for example, if the treatment in question is for an irreversible and fatal disease—a treatment that does not have a well-defined benefit–risk profile might be the only ethically acceptable comparator. In such cases, FDA should take the questionable benefit–risk profiles of the drug and its comparator into account when interpreting the results of the study.
Borer, J. S., H. Pouleur, E. Abadie, F. Follath, J. Wittes, M. A. Pfeffer, B. Pitt, and F. Zannad. 2007. Cardiovascular safety of drugs not intended for cardiovascular use: Need for a new conceptual basis for assessment and approval. European Heart Journal 28(15):1904-1909.
Flory, J., and E. Emanuel. 2004. Interventions to improve research participants’ understanding in informed consent for research. JAMA 292(13):1593-1601.
Hiatt, W. R. 2006. Observational studies of drug safety—aprotinin and the absence of transparency. New England Journal of Medicine 355(21):2171-2173.
IOM (Institute of Medicine). 2007. The future of drug safety: Promoting and protecting the health of the public. Washington, DC: The National Academies Press.
Kass, N. E., and H. A. Taylor. 2008. Empirical research on informed consent: A review of the literature prepared for the National Academies Workshop on Collecting, Storing, Accessing, and Protecting Data Containing Biological Measures (unpublished).
Lindegger, G., C. Milford, C. Slack, M. Quayle, X. Xaba, and E. Vardas. 2006. Beyond the checklist: Assessing understanding for HIV vaccine trial participation in South Africa. Journal of Acquired Immune Deficiency Syndromes 43(5):560-566.
NRC (National Research Council). 2009. Science and decisions: Advancing risk assessment. Washington, DC: The National Academies Press.
NRC. 2011. A risk-characterization framework for decision-making at the Food and Drug Administration. Washington, DC: The National Academies Press.
Psaty, B. M., and J. P. Vandenbroucke. 2008. Opportunities for enhancing the FDA guidance on pharmacovigilance. JAMA 300(8):952-954.
Ray, W. A., and C. M. Stein. 2006. Reform of drug regulation—beyond an independent drug-safety board. New England Journal of Medicine 354(2):194-201.