Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 231
B
Committee’s Letter Report
ETHICAL ISSUES IN STUDYING THE SAFETY
OF APPROVED DRUGS: A LETTER REPORT
Margaret Hamburg, MD
Commissioner
US Food and Drug Administration
10903 New Hampshire Avenue
Silver Spring, MD 20993-0002
Dear Dr. Hamburg,
In April 2010, the US Food and Drug Administration (FDA) asked the Institute
of Medicine (IOM) to respond to five questions about ethical and scientific issues
in studying the safety of approved drugs. FDA requested a final report on the five
questions in 2011. In light of the scheduling of a joint meeting of FDA’s Endo-
crinologic and Metabolic Drugs Advisory Committee and Drug Safety and Risk
Management Advisory Committee on July 13–14, 2010, FDA further requested
a letter report addressing question 1 of the charge—“What are the ethical and
informed consent issues that must be considered when designing randomized
clinical trials to evaluate potential safety risks?”—by July 2010. The attached let-
ter report, which has been reviewed in accordance with IOM review procedures,
addresses that question.
Sincerely,
Ruth R. Faden
Steven N. Goodman
Cochairs, Committee on Ethical and Scientific Issues in Studying the Safety of
Approved Drugs
231
OCR for page 232
232 STUDYING THE SAFETY OF APPROVED DRUGS
CONTEXT OF THE INSTITUTE OF MEDICINE
STUDY AND CHARGE TO THE COMMITTEE
Public Law 110-85, the Food and Drug Administration Amendments Act of
2007 (FDAAA 2007; PL 110-85) expanded the US Food and Drug Administra -
tion (FDA) authorities and responsibilities over drugs1 during the postmarketing
period (that is, after a drug is approved to enter the US market). The expanded
authorities, many of which were recommended in The Future of Drug Safety:
Promoting and Protecting the Health of the Public (IOM, 2007), provide FDA
with additional regulatory tools, such as requiring clinical trials or other stud -
ies after a drug has been approved, to protect the health of the public. With the
expanded postmarketing authorities comes the recognition that critical decisions
regarding the study of drugs after approval raise new challenges and questions,
both ethical and scientific, for the agency to consider. FDA therefore asked the
Institute of Medicine (IOM) to “convene a committee to evaluate the scientific
and ethical issues involved in conducting studies of the safety of approved
drugs.” The specific questions that the committee was asked to evaluate are
presented in Box 1. In light of the scheduling of a joint meeting of FDA’s Endo-
crinologic and Metabolic Drugs Advisory Committee and Drug Safety and Risk
Management Advisory Committee on July 13–14, 2010, FDA requested a letter
report addressing question 1 of the charge—“What are the ethical and informed
consent issues that must be considered when designing randomized clinical trials
to evaluate potential safety risks?”—by July 2010.
COMMITTEE’S APPROACH TO ITS CHARGE
In response to FDA’s request, IOM convened a committee of persons who
had expertise in clinical trials, epidemiology, pharmacoepidemiology, bioethics,
law, patient safety, biostatistics, public health, and health policy. Those experts
agreed to prepare both this letter report, which focuses on question 1 of the
charge, by July 2010 and a final report that addresses all the questions in the
charge by 2011.
For the present letter report, the committee held one meeting, which included
an open session in which it heard from representatives of FDA and representatives
of the Agency for Healthcare Research and Quality (AHRQ) and the National
Institutes of Health (NIH), which funded this report with FDA. The committee
provided an opportunity for other stakeholders to present their perspectives and
concerns at the meeting. The committee conducted searches of the literature on
the ethics of clinical trials and informed consent relevant to postmarketing clini -
1 For simplicity, the committee uses the term drugs throughout this report, but similar considerations
would apply to biologics.
OCR for page 233
233
APPENDIX B
BOX 1
Charge to the Committee
The Food and Drug Administration (FDA) has requested that the
Institute of Medicine convene a committee to evaluate the scientific and
ethical issues involved in conducting studies of the safety of approved
drugs. Questions to be explored by a committee include:
1. What are the ethical and informed consent issues that must be con-
sidered when designing randomized clinical trials to evaluate potential
safety risks?
2. What are the strengths and weaknesses of various approaches,
including observational studies, including patient registries, meta-
analyses, including patient-level data meta-analyses, and randomized
controlled trials, to generate evidence about safety questions?
3. Considering the speed, cost, and value of studies, what types of
follow-up studies are appropriate to investigate different kinds of sig-
nals (detected pre-approval or post-marketing) and in what temporal
order?
4. Under what circumstances should head-to-head randomized clinical
trials for safety be required?
5. How should FDA factor in different kinds of safety evidence in consid-
ering different kinds of regulatory actions?
cal trials. This letter report does not, however, present a comprehensive literature
review of the subject.
Given the short period available for preparing this letter report, the com -
mittee focused on identifying a conceptual framework to guide its analysis of
the ethics of the design and conduct of postmarketing safety research required
by FDA, including key issues that need to be taken into account in assessing
ethics and informed consent in randomized controlled trials. In developing this
framework, and in its explication in this letter report, the committee relied on
the extensive body of codes, regulations and guidance on the ethics of research
involving human participants, much of which is built around a commitment to
several basic moral principles, including beneficence, respect for persons and
their autonomy, and justice. The committee did not enumerate all the ways in
which the issues raised in this letter report can affect the ethics of a study, did
not detail how the various issues should be weighed against one another, and did
not explore in depth issues related to the ethical and scientific justifications of
randomized controlled trials. A more detailed analysis of those issues and their
implications and effects will be included in the committee’s final report.
OCR for page 234
234 STUDYING THE SAFETY OF APPROVED DRUGS
BOX 2
Conceptual Framework for Analyzing the Ethics of
Postmarketing Randomized Clinical Trials Required
by the Food and Drug Administration: Four Central
Classes of Considerations and Recommendations
I. The Public Health Context. The Food and Drug Administration
(FDA) should determine that there is a substantial public health
question about the nature or acceptability of the risks, or the risk–
benefit profile, of a marketed drug—a question that requires a policy
decision from FDA.
II. Regulatory Science and Public Accountability. FDA should use
regulatory-science principles and practices that include processes
of public accountability and transparency to determine the need for
a policy decision, the need for new knowledge to support a policy
decision, and the policy decision based on the new knowledge.
III. Design Considerations. It is appropriate for FDA to require that
a randomized controlled trial be conducted to provide additional
evidence about an approved drug’s efficacy and safety only when
(i) uncertainty about the risk-benefit balance is such that a respon-
sible policy decision cannot be made based either on the existing
evidence or on evidence from new observational studies, and (ii)
the trial is properly designed and implemented to reduce uncertainty
about the risk-benefit balance sufficiently for a responsible policy
decision to be made.
IV. Additional Ethical Obligations to Trial Participants. FDA should
ensure that the trial will answer the public health question with a
design that minimizes risks to trial participants and involves ongo-
ing monitoring of risks. The risks should be judged to be acceptable
by appropriate oversight bodies before and during the trial and by
trial participants at enrollment and as appropriate during the trial.
Specifically, FDA and appropriate oversight bodies should ensure
that the trial includes a comprehensive and meaningful informed
consent process that continues during the trial and that takes into
account any substantial changes in clinical practice and professional
standards and any new research findings relevant to a participant’s
willingness to accept the risks associated with the trial. The FDA and
appropriate oversight bodies should ensure that those conducting
the trial convey such changes to participants in a timely and under-
standable fashion.
OCR for page 235
235
APPENDIX B
The committee’s conceptual framework consists of four classes of consider-
ations, as shown in Box 2. In accordance with the framework, the remainder of
this letter report is organized in four major sections: the public health context of
drug safety, regulatory science and public accountability, design considerations,
and additional ethical obligations to research participants.
THE PUBLIC HEALTH CONTEXT OF DRUG SAFETY
The ethics of any postmarketing study required by FDA, including random -
ized controlled trials, should be assessed in the context of FDA’s mission to
promote and protect public health. The safety of the US drug supply contributes
to the nation’s health, and FDA is the agency responsible for ensuring this safety.
As stated by the FDA commissioner and deputy commissioner, “to be healthy,
people need access to . . . innovative, safe, and effective medical products” and
“FDA’s job is to support this access and, in doing so, to promote health, prevent
illness, and prolong life” (Hamburg and Sharfstein, 2009). With specific reference
to drugs, FDA’s job includes (FDA, 2010a)
• “Protecting the public health by assuring the safety, effectiveness, and secu-
rity of human and veterinary drugs, vaccines and other biological products”
• “Advancing the public health by helping to speed product innovations”
• “Helping the public get the accurate, science-based information they need
to use medicines and foods to improve their health”
The committee believes that FDA, to fulfill its public health mission, should
allow a drug to enter and remain on the market only if the balance of the risk to
the benefit is appropriate for its intended use. The committee also believes that
it is critical to FDA’s public health mission that the agency: provide information
needed by clinicians to prescribe a drug responsibly and needed by patients to
take it appropriately; foster innovation and drug development by using decision-
making processes that are predictable, clear, and timely; and conduct its respon -
sibilities in a way that fosters public trust in the drug oversight system.
Ensuring the acceptability of the risk–benefit profile of a drug after it is
approved for the US market is no less central to FDA’s public health mission than
ensuring the acceptability of the profile before it is permitted to enter the market.
As discussed later in this letter report, because of the infrequency and delayed
occurrence of some adverse events, there is often more uncertainty about the
risks posed by a new drug at the time of approval than there is about its efficacy.
In addition, when an agent is approved on the basis of surrogate end points, the
estimated degree of benefit may change when the effect on clinical end points
is studied. Postmarketing research may be important for examining such clinical
end points. Therefore, the committee agrees with a previous IOM committee that
a drug-safety system “has at its core a lifecycle approach to drug risk and ben -
OCR for page 236
236 STUDYING THE SAFETY OF APPROVED DRUGS
efit” and that such a system “would require continuous availability of new data
and ongoing, active reassessment of risk and benefit to drive regulatory action
(responsive to the accumulating information about a given drug), and regulatory
authority that is strong both before and after approval” (IOM, 2007).
The new authorities and regulatory tools provided in FDAAA 2007 (PL
110-85) expanded the possibilities for FDA to adopt a comprehensive life-cycle
approach to the assessment of the risks and benefits associated with marketed
drugs. FDAAA 2007 mandated that FDA establish an active surveillance system
for monitoring drugs by using electronic data from health-care information hold -
ers and gave FDA new authorities that include the ability to require revisions to a
product label, to require further study of a drug, to restrict the use of a treatment
to specified populations, and to require a formal Risk Evaluation and Mitigation
Strategy (REMS). Those authorities provide new regulatory opportunities that
are short of the pre-existing option of drug withdrawal. Under FDAAA 2007,
FDA can require postmarketing studies and clinical trials under the following
circumstances (PL 110-85):
“To assess a known serious risk related to the use of the drug involved.”
“To assess signals of serious risk related to the use of the drug.”
“To identify an unexpected serious risk when available data indicates the poten -
tial for a serious risk.”
The ability to require further study of a drug is a powerful tool for FDA to
use in acquiring additional information to make informed, science-based deci -
sions as part of its public health mission. In making a decision whether to require
a postmarketing study, however, FDA not only should consider the ethical issues
that arise in obtaining information to clarify a policy decision2 but should bear
in mind that such issues vary among types of studies.
The committee concludes that for FDA-required postmarketing research to
be ethical, a critical first step is the determination by FDA that it is facing a policy
decision of importance to public health that cannot satisfactorily be resolved with
existing evidence.3,4
2 When referring to a policy decision the committee means choosing among the range of responses
available to the FDA when safety signals emerge—including the decision to continue a drug’s
monitoring plan without modification, the decision to add a warning to a drug’s label, the decision
to require a postmarketing trial, and the decision to remove a drug from the market—some of which
are not mutually exclusive.
3 The committee’s conclusion is consistent with that of a previous committee of the National
Research Council that was related to Environmental Protection Agency consideration of research
involving human subjects (NRC, 2004).
4 This conclusion, and this entire letter report, is specific to research on postmarketed products
required by FDA. In this regulatory and public health context, it is critical from an ethics standpoint
that existing evidence be insufficient to make an appropriate policy determination. Scientific studies
OCR for page 237
237
APPENDIX B
REGULATORY SCIENCE AND PUBLIC ACCOUNTABILITY
As noted above, FDA can require a postmarketing trial “to assess signals of
serious risk” (PL 110-85). The key to the ethics of a postmarketing safety trial
is a determination that a safety signal, if it represents a true risk, would warrant
a policy decision and that new knowledge is needed to determine the existence
and magnitude of the risk and thereby inform the nature of the decision. If, for
example, the existing information about a safety risk is sufficient to warrant
the removal of a drug from the market, then it would be unethical to conduct a
trial. On the other hand, existing evidence about a new safety signal may be suf-
ficient to warrant a change in labeling but not sufficient to warrant removal from
the market, a policy decision that may be appropriate once the risks, or risks in
relation to potential benefits, are better characterized. In such a context, it may
be possible to design and implement an ethically acceptable trial. The same
reasoning applies to judgments about whether a current trial should be stopped.
If new evidence from any source, including the trial itself, is determined to be
sufficiently compelling to ground a policy decision without waiting for additional
new information, allowing the trial to continue would be unethical.
The ethics of postmarketing studies requires that the kinds of determinations
outlined above be based on the best principles and practices for making policy
decisions under conditions of uncertainty, including appropriate processes for
transparency in decision making and public accountability. Those principles and
practices, sometimes referred to as the emerging field of regulatory science, 5
require that policy decisions reflect the best available scientific evidence and ana-
lytic techniques drawn from a wide array of disciplines and technical expertise,
including decision sciences, behavioral economics, and cognitive psychology.
Public accountability and transparency increase the likelihood that the perspec -
tives of stakeholders,6 who have kinds of knowledge different from those of
technical experts, are included in the making of policy decisions. Transparency
and other public accountability processes also may increase the likelihood that the
public will view regulatory and policy decisions, including the conduct of a trial
and a decision to continue or discontinue a clinical trial, as fair and acceptable. 7
of approved and marketed medical products outside this FDA context are an increasing component
of biomedical and health services research and also can contribute significantly to population health.
5 FDA defines regulatory science as “the development and use of new tools, standards and approaches
to more efficiently develop products and to more effectively evaluate product safety, efficacy and
quality” (FDA, 2010b).
6 Relevant government stakeholders include FDA, NIH, AHRQ, and the Centers for Disease
Control and Prevention. Relevant nongovernment stakeholders include industry, academe, health-care
providers, payers, patients, and other members of the public.
7 As the committee has already noted, FDA and those advising FDA therefore should have access to
all information relevant to a given public health question, whether or not the information is deemed
proprietary information or trade secrets. One tension in meeting acceptable standards of transparency
with stakeholders is managing public access to such information.
OCR for page 238
238 STUDYING THE SAFETY OF APPROVED DRUGS
Accurately assessing the risks posed by and the potential benefits of a drug
requires the use of a wide variety of scientific data, including findings from
animal studies of toxicology, basic research (for example, mechanistic studies
and structure–activity relationships), clinical trials, high-quality epidemiologic
and health-services research (such as observational studies and meta-analyses),
and postmarketing surveillance systems that detect and analyze adverse events.
FDA and those advising FDA therefore should be able to consider all data, and
the design and analyses that led to those data, that are relevant to a given public
health question, whether or not they are deemed proprietary information or trade
secrets.
Judgments about the adequacy of available evidence for FDA decisions
require input from a multidisciplinary team acting through a process that can
integrate and take advantage of the different kinds of knowledge and perspectives
that reside in clinical practice, biologic science, ethics, biostatistics, epidemiol -
ogy, and research design. The decision-making process should also minimize
and correct for potential cognitive and intellectual biases that arise from previous
policy decisions or strongly held opinions—for example, the human tendency to
focus on evidence that confirms a pre-existing belief or decision and to discount
evidence that contradicts it.
Modern tools for risk communication and public engagement should be used
to ensure that all stakeholders—including physicians, other health professionals,
interested patients and their families, and members of the public—understand the
decision problem facing the agency, including what is known about the benefits
and risks associated with the therapy in question and the pertinent uncertainties.
Uncertainties could pertain to the quantity and quality of evidence, the risk–ben -
efit profile, or the effect of policy decisions on future risks. Engagement with
stakeholders is required to explain the types of uncertainties at issue and how
the agency is dealing with the uncertainties in making its policy decision and to
permit the agency to understand how those affected by its actions weigh risks
and benefits.8
In using best practices to determine whether additional research is required,
the agency should also keep in mind that one aspect of its mission to advance
public health involves accelerating the evolution of effective new therapies from
bench to bedside by encouraging product innovation (FDA, 2010a). That is most
likely to occur when FDA’s regulatory regime facilitates industry’s ability to
make informed research-and-development decisions—for example, by applying
consistent requirements and criteria for assessing risks and possible benefits, by
making decisions in accord with a clear and understandable framework, and by
8 The committee acknowledges that there are significant challenges to implementing policy making
and regulatory processes that appropriately balance scientific evidence and stakeholder input (Lomas
et al., 2005).
OCR for page 239
239
APPENDIX B
responding in predictable ways to different kinds of information, including new
information about risk.
DESIGN CONSIDERATIONS
It is never ethical to involve research participants in an inappropriately
designed or inappropriately conducted study or any study that does not have a
reasonable prospect of answering the research question under study. Without
a reasonable prospect of contributing to scientific knowledge, the exposure of
research participants to even minimal risk or inconvenience can never be justi -
fied. In the case of postmarketing clinical trials required by FDA, that ethical
precept requires further specification and strengthening. In particular, before a
clinical trial is selected as the design of choice, it should be determined that no
other research or information gathering effort—including a new observational
study—can reduce the uncertainty about a drug’s risk–benefit profile sufficiently
to support a responsible policy decision.9 It is also critical that the clinical trial
be designed to provide precisely the data needed to facilitate the policy decision
that needs to be made. Finally, there should be sufficient continuing monitoring
of the trial to ensure that the associated risks (if any) continue to be acceptable. 10
A comprehensive assessment of risks associated with a drug is often impos-
sible in the premarketing-study phases because of small samples, short followup,
and the selected nature of the populations included in preapproval trials. In
addition, across the lifespan of a drug, the benefit profile and consequently the
acceptability of risks in relation to benefits can change with the development
of alternative treatment or prevention methods or even with the evolution of
the disease or causative agent, such as the development of resistance to a given
antibiotic. The assessment of benefits and risks is a dynamic process that requires
continual revisiting and monitoring, and changes in evidence about risks should
be considered against evidence about benefits at the time of the reassessment.
Postmarketing safety studies constitute an important part of understanding the
dynamics of the risk–benefit balance.
The most important features of any research are that the research question
is properly conceived and that the proposed study is designed appropriately to
address the question that has been specified. In the postmarketing context all
such questions pertain to the risk–benefit profile of an approved treatment. How -
ever, not all changes in the risk–benefit balance are policy concerns, although
they might merit alterations in medical practice. For example, the introduction
9 This observation is again specific to the FDA context under consideration in this letter report.
10 Continuing monitoring of a trial is essential to ensure that risks (if any) to participants continue to
be acceptable. How monitoring should be conducted is also an essential feature of a properly designed
trial. In this letter report, we have elected to discuss monitoring in the section on design considerations.
It would fit equally well in the section on other obligations to trial participants.
OCR for page 240
240 STUDYING THE SAFETY OF APPROVED DRUGS
of lower-risk therapies of similar or greater efficacy would justify changes in
medical practice; without a new safety concern about the old agent, however, this
situation might not require action by FDA.
A number of questions of policy relevance can remain or arise after approval.
They include:
• Whether treatments approved on the basis of surrogate end points or
biomarkers—such as lipid concentrations, blood pressure, or glycated
hemoglobin—show improvement in clinical end points.
• Whether benefits seen in preapproval studies are not experienced by iden-
tifiable patient groups, in which case the acceptability of risks in these
groups might be altered.
• Whether additional safety concerns that affect the risk–benefit profile
arise from
Newly identified serious adverse events.
More serious or more frequent harms than expected in the intended
population or in identifiable patient groups that may be defined by
co-treatments, patient characteristics or co-morbidities, or disease or
treatment-delivery characteristics.
New safety signals may arise from various sources: spontaneous reports of
adverse events, safety-surveillance systems, observational studies, meta-analyses,
and randomized trials. FDA can require new research to address key safety ques -
tions if the existing evidence is insufficient to infer causality or to characterize
the frequency and severity of observed harms with adequate confidence or if such
evidence is not complete enough to judge the acceptability of the risk–benefit
profile for a drug’s intended use.
The first step in deciding whether new research is needed is to assess the
strength of the existing evidence related to new safety or risk–benefit concerns.
The traditional hierarchies of evidence based on study design, which are regularly
used in determinations of efficacy (Barton et al., 2007; Owens et al., 2010), might
not apply in a straightforward manner to safety evidence. Randomized controlled
trials are optimal for efficacy determinations because the randomization of large
numbers of patients creates groups that have similar average risks of the outcome
of interest. Observational studies designed to evaluate the efficacy or anticipated
effects of treatment, either intended benefits or expected harms, are often liable to
confounding by indication (Vandenbroucke and Psaty, 2008). That is, the reasons
that physicians treat patients differently or that patients prefer particular treatment
options are often related to factors that themselves affect outcomes. For example,
if sicker patients choose medical care more often and avoid surgery, observational
studies of surgical vs medical care could provide false evidence that surgery has
more favorable results than medical care. Similarly, if an adverse effect of a drug
is known, physicians may avoid prescribing it for patients who are at higher risk
OCR for page 241
241
APPENDIX B
for the effect. Thus, in observational studies of the anticipated effects of treat -
ment, it may be difficult to determine whether differences in outcomes are due to
the treatments themselves or to the other factors that led to the treatment choices.
Although such differences due to other factors can often be minimized through
design and analysis, they cannot be eliminated with the same confidence as one
would attach to a high-quality randomized trial.
In the evaluation of unintended or previously unsuspected effects of drugs,
however, observational safety studies are less likely than studies of known effects
to be influenced by confounding by indication. Under specific circumstances,
observational studies may be adequate not only to identify the presence of an
important safety issue but, if the findings are replicated, to provide convincing
evidence that an association is causal. For instance, a well-designed and well-
conducted observational comparison of two similar drugs that came onto the
market at the same time, that are used for the same condition at the same stage,
and that have similar side-effect profiles could provide useful and valid estimates
of the risk associated with a safety signal (Vandenbroucke and Psaty, 2008). In
addition, some observational studies of safety may have distinct advantages over
trials. They can often be much larger than randomized controlled trials, involve
longer patient followup, include a broader diversity of patients and care settings,
and be completed more quickly. Because of those features, observational stud -
ies evaluating infrequent outcomes that occur long after exposure and in which
confounding by indication is unlikely can sometimes provide higher-quality
safety evidence than randomized controlled trials, if the trials were not optimally
designed to capture such safety outcomes.
The relative strength of other research designs may be different between
safety and efficacy determinations. Meta-analysis of randomized controlled trials
can increase the ability to detect rare events, but if the trials encompassed by the
meta-analysis were not well designed or well conducted to capture safety out -
comes or reported them inconsistently (Ioannidis and Lau, 2001), the meta-analy -
sis may produce misleading results. An unexpectedly low incidence in the control
group of a randomized trial may signal a problem with the conduct of the study.
All observational studies and meta-analyses of randomized trials may be
affected by confounding or bias. If the estimated relative risks are small, selection
bias, confounding, and measurement error may be alternative explanations for
associations found in an observational study. But small relative risks of serious
outcomes associated with widely used agents can have substantial public health
consequences. Under such circumstances, if there is substantial uncertainty about
a safety signal, a well-designed and well-conducted postmarketing randomized
clinical trial is the best approach for characterizing the risk–benefit profile. The
opportunity to evaluate both risks and benefits in the same study is an important
advantage of randomized trials.
In evaluating or proposing a postmarketing randomized trial, the design
and conduct should be closely scrutinized for quality and relevance to the US
OCR for page 242
242 STUDYING THE SAFETY OF APPROVED DRUGS
context. Findings from trials conducted in countries where medical care differs
substantially from that in the United States may be less relevant to US popula -
tions (HHS, 2010).
Non-inferiority studies—designed with the one-sided intent to show that a
therapy is not worse than another by some predetermined margin—pose some
special challenges compared with the superiority trials traditionally used to evalu-
ate efficacy (Fleming, 2008; Kaul and Diamond, 2006, 2007). The implications
of poor quality in the design or conduct of a non-inferiority study are often the
opposite of those in a superiority trial (Temple and Ellenberg, 2000). Low-quality
study conduct, such as poor compliance with treatment regimens, usually biases
a superiority trial toward a finding of no difference between treatments but often
biases a non-inferiority trial toward a finding of “equivalence” or “non-inferior-
ity” between treatments. Thus, the findings of a poorly conducted non-inferiority
trial may inappropriately support a conclusion that the treatments under study
are “equally” efficacious or “equally” safe. Non-inferiority trials may therefore
require special oversight and scrutiny by FDA, as well as appropriate adjustment
for poor compliance, to ensure valid inferences.
Another critical, and perhaps underappreciated, aspect of non-inferiority
designs that makes them problematic for safety assessments is the rationale for
the choice of the non-inferiority margin. The selection of a margin that is too
large can result in a finding that the two treatments are “equally safe” even if their
risks are substantially different. Regardless of the hypothesis-test verdict in such
a trial, FDA should look carefully at the estimated difference and its confidence
interval in deciding whether meaningful differences in safety have truly been
ruled out (Kaul and Diamond, 2006).
All those considerations also apply to the assessment of existing evidence
and to a determination of what kind of research design is needed to generate new
evidence. Because observational designs usually generate fewer ethical concerns
than randomized controlled trials, a decision to require a randomized controlled
trial to resolve safety questions should be based on the determination that neither
the existing evidence nor new, prospectively conducted observational studies
can provide safety evidence sufficiently reliable for FDA to make a sound policy
decision.
If a randomized controlled trial is deemed necessary for an FDA-policy deci-
sion, its characteristics should include the following:
1. The evidence gap should be clearly present and specifically identified,
and the research question and study design should be precisely crafted
to address the gap.
This effort involves not only the review of the quantity, quality, and consistency
of the existing evidence but careful selection of a study population, end points,
treatments, comparators, and setting.
OCR for page 243
243
APPENDIX B
2. The trial should be adequately powered, and the trial procedures
and the pre-specified analytic plans should be appropriate to provide
answers to the study questions.
If a study addresses more than one question or end point, it should be powered
so that all major outcomes of interest can be adequately studied. If the proposed
trial uses a non-inferiority design, the non-inferiority margin should confidently
exclude small risks of serious events, especially for widely used drugs. The ana -
lytic plan should be laid out in detail at the time the study protocol is approved
by the sponsor and institutional review boards (IRBs). The data-management
and quality-assurance plans should be fully described and adequate both for the
protection of research participants and for the trial to achieve its aims.
3. The inclusion and exclusion criteria should reflect the best available
knowledge about risks and potential benefits in the population.
From a public health perspective, it is desirable to test the effectiveness and safety
of a drug for its intended use in a sample that is representative of the population
receiving the drug. However, the ethical obligation to minimize risks to research
participants may require excluding some who are at a high risk of adverse events.
It is never ethically justified to include in a postmarketing trial participants for
whom the drug is contraindicated by the currently approved product label unless
their involvement is necessary to answer a specific question and the risks to them
posed by participation are acceptable.11 The exclusion of participants for whom
more moderate safety warnings or precautions have been issued presents a more
difficult case and involves a tradeoff among several considerations: the preven -
tion of possible harm to participants, the generalizability of the trial’s findings to
patient populations in which the drug is being used, and the ability to reach an
answer to the study’s safety questions more quickly (if the participants are likely
to experience the outcome of interest at a higher rate).12
4. A comprehensive and robust safety-monitoring plan should be in
place.
Every postmarketing clinical trial should have a properly qualified data-safety
monitoring board (DSMB) in place with a written charter and a pre-specified
data-monitoring plan, which includes statistical guidelines for stopping the trial
11 I fnew information raises substantial uncertainty about the appropriateness of the current product
warning, suggesting that it may be in the interest of patients to have the warning removed, it may
be ethically acceptable to mount a trial that involves patients who are the subject of the warning to
resolve this question.
12 If such a trial is otherwise determined to be ethically justifiable, the consent process should
emphasize to potential participants the existence of safety warnings or precautions.
OCR for page 244
244 STUDYING THE SAFETY OF APPROVED DRUGS
early (Ellenberg et al., 2003; Grant et al., 2005). The frequency and intensity of
DSMB review should be determined on the basis of the seriousness, incidence,
and timing of known or possible harms. The DSMB should meet before trial onset
to review and approve the charter, protocol, and monitoring procedures and then
at regular intervals to review not just outcomes and adverse events but the various
aspects of trial conduct and data quality.
A critical issue for trial monitoring is the standard of evidence required to
halt a trial on the basis of harm. Typically, differences that cross pre-specified
boundaries of statistical significance are required to halt trials for efficacy. How -
ever, depending on the type and degree of benefit, boundaries for harm may vary.
The criteria for stopping a trial if the efficacy end point veers in the direction of
harm are typically less stringent than the criteria for stopping for efficacy differ-
ences in the direction of benefit. Modest evidence of an adverse effect on an effi -
cacy end point may be sufficient to rule out a clinically meaningful benefit even
if the point estimate does not exclude a null effect. On the other hand, if benefit
on one end point is established (for example, cardiovascular health), but the trial
is being done to assess a suspected harm on a different end point (for example,
hepatic failure), a higher standard of proof of the harm signal might be required.
The Women’s Health Initiative trial, for example, stopped its estrogen–progestin
arm because the breast-cancer outcome crossed the pre-specified safety bound -
ary and because the global index outcome just trended in the direction of harm,
effectively ruling out a substantive net benefit (Wittes et al., 2007).
Other issues that affect the evidence threshold for stopping for harm are
whether and how external information is used. This matter is not a settled meth -
odologic issue, but if an emerging signal of harm is similar to that seen in external
studies, it is ethically justified and may be ethically required to halt a trial earlier
than if such evidence did not exist (Pocock, 1996).
Although vigorous safety monitoring is crucial for minimizing risks to par-
ticipants in postmarketing trials, it is but one of multiple ethical considerations
that must be addressed and satisfied if ethical obligations to research participants
are to be fully honored.
ADDITIONAL ETHICAL OBLIGATIONS
TO RESEARCH PARTICIPANTS
In the context of FDA-required randomized controlled trials, the need for a
well-designed randomized controlled trial to determine the proper policy deci -
sion in response to a new drug-safety concern is a necessary but not sufficient
condition for a trial to be ethically acceptable. Obligations to protect the rights
and welfare of participants in a trial—to whom special duties of care and com -
passion may be owed because of illness, disability, or threat of illness—should
be respected.
The general ethical principles governing research that involves human par-
ticipants are well established and apply to the postmarketing context as they do to
OCR for page 245
245
APPENDIX B
all human research (Council for International Organizations of Medical Sciences,
2002; DHEW, 1979). In the present letter report, the committee specifies aspects
of those principles that have particular relevance to postmarketing research. In
a postmarketing study, the risks to participants should be kept to the minimum
that can be achieved while the trial is still able to answer the motivating policy
question. The risk–benefit balance should be judged to be acceptable by FDA,
participating IRBs, and the DSMB before initiation and throughout the course of
the trial. That balance should also be acceptable to trial participants. To ensure
that patients view the risks as acceptable in relation to any potential benefits, the
trial should include a meaningful informed consent process that continues over
the course of the trial and that includes prompt communication to participants
of relevant new evidence or developments in clinical practice or professional
standards that might affect their evaluation of the risks and benefits associated
with continued participation.
Although the risks to research participants in randomized controlled trials are
expected to be reasonable in relation to anticipated benefits, there is substantial
consensus in both domestic regulatory and other guidance documents that dif -
ferent ways of balancing risk and benefit can be ethically justified. For example,
both FDA regulations (21 CFR 50/56) and the Common Rule (45 CFR 46 Subpart
A) distinguish among research that does not present greater than minimal risk,
research that involves greater than minimal risk but offers the prospect of direct
benefit to individual subjects, and research that involves greater than minimal
risk and no prospect of direct benefit to individual subjects but is likely to yield
scientific knowledge about the subjects’ disorder or condition. A trial in which
the risks to participants are not outweighed by the prospect of direct medical
benefits to participants may be justifiable if a question of pressing public health
importance cannot be properly answered without the conduct of the trial and if
other conditions intended to safeguard the rights and interests of participants
are satisfied. Those conditions include but are not limited to determination by
appropriately constituted review committees that the risks are small enough to
make it ethically acceptable to ask people whether they are willing to be exposed
to the risks in the service of contributing to the public good, minimization of the
risks through careful study design and a robust monitoring plan that is in place
throughout the course of the trial, and implementation of a thorough informed
consent process that adheres to the highest standards of respect for participants.
The informed consent process should provide an accurate, comprehensible
explanation of the available knowledge about the risks and benefits associated
with being assigned to the treatment and control groups. It is a bedrock principle
of research ethics that participants who put themselves at risk in human research
should receive an understandable, unbiased, accurate, and comprehensive disclo -
sure of the potential benefits and risks attached to study participation (DHEW,
1979; ICH, 1996). A comprehensive disclosure is important to fulfill the substan-
tive moral requirement of informed consent that participants have a meaningful
understanding of what is being asked of them, including the risks and benefits
OCR for page 246
246 STUDYING THE SAFETY OF APPROVED DRUGS
(if any), not merely that information is provided to them (Faden and Beauchamp,
1986).
When a substantial amount of information indicating that a drug to be studied
may involve serious safety risks has already accumulated, there are heightened
obligations to ensure that potential participants understand the risks posed by
study enrollment. Those obligations may include special efforts to communicate
complex risk information clearly and to establish that participants have sufficient
understanding of what the risks mean to them.
The emphasis given to risk information in the consent process should
increase with the severity of risk and the level of certainty about the causal
association between the drug and the adverse outcome. At a minimum, risks that
should be disclosed should include any black-box warnings, the “major state -
ment” currently listed in television advertisements, any adverse-event findings
of an FDA advisory committee, and a summary of evidence from published
peer-reviewed studies.
Communicating complicated risk information and research findings to
participants poses challenges. It is critical that the information be conveyed in
a manner that can be understood and weighed by participants. A “kitchen sink”
approach to consent-form drafting, in which voluminous information is included
with little attempt to distill it into a short format that is useful to participants, is
unfortunately increasingly common in clinical trials and should be avoided. Par-
ticipants are likely to be overwhelmed by a long and complex form and unable to
weigh conflicting study findings or findings about different types of risk.
Verbal disclosures and written consent documents (both consent forms and
information sheets) should help potential participants to understand how experts
weigh the available evidence about the safety profile of the drug being studied.
Moreover, there is a growing set of additional resources (for example, decision
aids, videos, and interactive electronic presentations) to supplement written
materials that may enhance participants’ understanding of complex clinical
information. Although evidence about the effectiveness of techniques designed
to improve and document understanding among potential research participants
is mixed (Kass and Taylor, 2008), such interventions as engaging in additional
interpersonal conversations with potential participants and asking them to explain
the study to a friend have been shown to be helpful (Flory and Emanuel, 2004;
Kass and Taylor, 2008; Lindegger et al., 2006). Whatever efforts are employed to
communicate with potential participants, it is key that they include information
that is useful to participants about where the weight of the evidence falls with
regard to serious risks and the level of confidence that experts have in drawing
conclusions about the risks. A statement that “Some studies have found that the
drug causes X, whereas others have not” may be true but misleading if nearly all
well-designed studies have reached the same conclusion and there is little or no
reliable evidence on the other side.
OCR for page 247
247
APPENDIX B
In addition to safety 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 safety profile. A statement that if a potential participant does
not enroll in the trial, he or she is more likely to have a different drug prescribed
should be communicated in this situation. If clinical practice continues to shift
during the trial period, the 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 ensure that trial partici-
pants understand the potential consequences of study participation in addition to
what they are contributing to the advancement of public health in the regulatory
arena. They cannot, however, serve as an exclusive or sufficient ethical justifica-
tion for conducting a postmarketing trial. The other ethical bases for initiating
a trial should be independently satisfied. People should not be asked to assume
risks that are not justified in light of the 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 par-
ticipate in research. Regulators, IRBs, and DSMBs should serve as particularly
strong bulwarks against unethical experimentation in such settings.
RECOMMENDATIONS
The committee recommends that the ethical and informed consent issues
related to FDA-required postmarketing clinical trials should be evaluated accord -
ing to the considerations identified in the conceptual framework summarized in
Box 2 as explicated in this letter report.
Given the timeframe of this letter report, the committee does not detail all
the issues within the framework or discuss how various considerations should be
weighed. The committee plans to provide further details in its full report in 2011.
REFERENCES
Barton, M. B., T. Miller, T. Wolff, D. Petitti, M. LeFevre, G. Sawaya, B. Yawn, J. Guirguis-Blake,
N. Calonge, R. Harris, and U.S. Preventive Services Task Force. 2007. How to read the new
recommendation statement: Methods update from the U.S. Preventive Services Task Force.
Annals of Internal Medicine 147(2):123-127.
OCR for page 248
248 STUDYING THE SAFETY OF APPROVED DRUGS
Council for International Organizations of Medical Sciences. 2002. International ethical guide-
lines for biomedical research involving human subjects. Geneva, Switzerland: World Health
Organization.
DHEW (Department of Health Education and Welfare). 1979. The Belmont Report: Ethical principles
and guidelines for the protection of human subjects of research. Washington, DC: U.S. Depart-
ment of Health and Human Services.
Ellenberg, S. S., T. R. Fleming, and D. L. DeMets. 2003. Data monitoring committees in clinical
trials: A practical perspective. Chichester, UK: Wiley.
Faden, R. R., and T. L. Beauchamp. 1986. A history and theory of informed consent. New York City:
Oxford University Press.
FDA (Food and Drug Administration). 2010a. What does FDA do? http://www.fda.gov/AboutFDA/
Basics/ucm194877.htm (accessed June 15, 2010).
———. 2010b. NIH and FDA announce collaborative initiative to fast-track innovations to the public.
http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm201706.htm (accessed
June 21, 2010).
Fleming, T. R. 2008. Current issues in non-inferiority trials. Statistics in Medicine 27(3):317-332.
Flory, J., and E. Emanuel. 2004. Interventions to improve research participants’ understanding in
informed consent for research: A systematic review. JAMA 292(13):1593-1601.
Grant, A. M., D. G. Altman, A. B. Babiker, M. K. Campbell, F. J. Clemens, J. H. Darbyshire,
D. R. Elbourne, S. K. McLeer, M. K. B. Parmar, S. J. Pocock, D. J. Spiegelhalter, M. R. Sydes,
A. E. Walker, and S. A. Wallace. 2005. Issues in data monitoring and interim analysis of trials.
Health Technology Assessment 9(7):1-238.
Hamburg, M. A., and J. M. Sharfstein. 2009. The FDA as a public health agency. New England Journal
of Medicine 360(24):2493-2495.
HHS (Department of Health and Human Services). 2010. Challenges to FDA’s ability to monitor and
inspect foreign clinical trials. Washington, DC: Department of Health and Human Services,
Office of the Inspector General.
ICH (International Conference on Harmonisation of Technical Requirements for Registration of
Pharmaceuticals for Human Use). 1996. ICH harmonized tripartite guideline: Guideline for
good clinical practice. Geneva, Switzerland: ICH Secretariat, International Federation for
Pharmaceutical Manufacturers Association.
Ioannidis, J. P. A., and J. Lau. 2001. Completeness of safety reporting in randomized trials: An evalu -
ation of 7 medical areas. JAMA 285(4):437-443.
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 litera-
ture prepared for The National Academies’ Workshop on Collecting, Storing, Accessing, and
Protecting Data Containing Biological Measures.
Kaul, S., and G. A. Diamond. 2006. Good enough: A primer on the analysis and interpretation of
noninferiority trials. Annals of Internal Medicine 145(1):62-69.
———. 2007. Making sense of noninferiority: A clinical and statistical perspective on its application
to cardiovascular clinical trials. Progress in Cardiovascular Diseases 49(4):284-299.
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.
Lomas, J., T. Culyer, C. McCutcheon, L. McAuley, and S. Law. 2005. Conceptualizing and combining
evidence for health system guidance: Final report for the Canadian Health Services Research
Foundation. Ottawa, Canada.
NRC (National Research Council). 2004. Intentional human dosing studies for EPA regulatory pur-
poses: Scientific and ethical issues. Washington, DC: The National Academies Press.
OCR for page 249
249
APPENDIX B
Owens, D. K., K. N. Lohr, D. Atkins, J. R. Treadwell, J. T. Reston, E. B. Bass, S. Chang, and M.
Helfand. 2010. AHRQ series paper 5: Grading the strength of a body of evidence when compar-
ing medical interventions-agency for healthcare research and quality and the effective health-care
program. Journal of Clinical Epidemiology 63(5):513-523.
Pocock, S. J. 1996. The role of external evidence in data monitoring of a clinical trial. Statistics in
Medicine 15(12):1285-1293.
Temple, R., and S. S. Ellenberg. 2000. Placebo-controlled trials and active-control trials in the
evaluation of new treatments. Part 1: Ethical and scientific issues. Annals of Internal Medicine
133(6):455-463.
Vandenbroucke, J. P., and B. M. Psaty. 2008. Benefits and risks of drug treatments: How to combine
the best evidence on benefits with the best data about adverse effects. JAMA 300(20):2417-2419.
Wittes, J., E. Barrett-Connor, E. Braunwald, M. Chesney, H. J. Cohen, D. Demets, L. Dunn, J. Dwyer,
R. P. Heaney, V. Vogel, L. Walters, and S. Yusuf. 2007. Monitoring the randomized trials of the
Women’s Health Initiative: The experience of the Data and Safety Monitoring Board. Clinical
Trials 4(3):218-234.
OCR for page 250
