A multicenter trial is a collaborative
effort that involves more than one independent center in enrolling
and following study participants. Multicenter randomized clinical
trials have a long and rich history, with Hill [1] and Greenberg [2] providing general discussions of methods in
the middle of the twentieth century.
In the last four decades, there has
been a dramatic increase in the number of multicenter and
multinational trials. Multicenter studies are more difficult and
more expensive to perform than single-center studies, and they may
bring less individual professional reward due to the need to share
credit among many investigators. However, multicenter trials are
necessary primarily because single sites cannot enroll enough
participants to assess clinically important outcomes
[3]. Over 40 years ago, Levin and
colleagues provided many examples of “the importance and the need
for well-designed cooperative efforts to achieve clinical
investigations of the highest quality” [4].
The reasons for conducting multicenter
trials apply even more today, with much of medicine being global in
scope. It is common for large late-phase trials sponsored by
industry to include a wide geographical representation. Several
hundred sites might be involved, each site entering anywhere from
several to a few dozen participants. While such dispersion of sites
presents logistical challenges for training of personnel and data
quality control, the benefits of rapid participant recruitment have
generally outweighed these challenges. Another potential advantage
of multicenter trials is that investigators at multiple sites, with
standardized protocols, may be less prone to bias that could affect
trial conduct and event ascertainment, especially in open-label
trials. Participants enrolled at a single center, all under the
oversight of an investigator who is academically invested in the
hypothesis, may be subject to a greater likelihood of bias.
Much of the ground work for the
development, organization, and conduct of a multicenter trial was
laid many years ago in trials like the Coronary Drug Project
[5] and the International Studies
of Infarct Survival (ISIS) [6,
7]. This chapter will discuss the
reasons why such studies are conducted and briefly review some key
steps in their planning, design and conduct.
Fundamental Point
Multicenter trials are needed to enroll
adequate numbers of participants in care settings that are likely
to reflect diverse practice. Investigators responsible for
organizing and conducting a multicenter study should have a full
understanding of the complexity of the undertaking and the need for
systems to assure that a common protocol is followed at each
site.
Reasons for Multicenter Trials
- 1.
The main rationale for multicenter trials is to recruit the adequate numbers of participants within a reasonable time. Many clinical trials have been—and still are—performed without a good estimate of the number of participants likely to be required to adequately test the main hypothesis. Yet, if the primary response variable is an event that occurs relatively infrequently, or small group differences are to be detected, sample size requirements will be large (Chap. 8).Studies requiring hundreds of participants usually cannot be done at one center, although there are some exceptions like the Deutsches Herzzentrum, in Munich, Germany, that has enrolled over 50,000 participants in a series of single site trials [8]. This site has also successfully participated in multicenter trials [9].Some multicenter trials have been very large. The Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO) trial enrolled 41,021 patients with acute myocardial infarction at 1081 hospitals in 15 countries, with enrollment ranging from 1 to over 200 patients per center [10]. This trial had four treatment groups, and treatment with accelerated t-PA (versus the streptokinase arms) resulted in a 14% relative risk reduction (and 1% absolute reduction) in 30-day mortality, a result that changed practice. The large sample size was required to be convincingly significant (p = 0.001). The Women’s Health Initiative (WHI) [11] was an ambitious 15-year project mandated by Congress in 1991 and sponsored by the National Institutes of Health (NIH). WHI included 161,000 postmenopausal women enrolled in 40 centers across the United States. A set of clinical trials, using a partial factorial design, included 68,132 women participants, addressed dietary modification, calcium and vitamin D supplementation, or hormone replacement therapy. The WHI provided important results that changed practice. And the program was a good investment, as shown by the fact that the $260 million cost of the WHI postmenopausal therapy trial was estimated to have a total net economic return of $37.1 billion [12, 13]. This was mainly the result of a change in practice such that women were no longer being exposed to the harmful effects of hormone replacement therapy. The Heart Protection Study 2–Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE) investigators took a different approach to site selection [14]. They selected 245 high-performing centers who demonstrated they had adequate patient volume to enroll large numbers of participants. These investigators randomized 25,673 patients (of the 42,424 entering the run-in phase) with prior vascular disease over 3 years at sites in the United Kingdom (89 sites), Scandinavia (84 sites), and China (72 sites) to niacin plus laropiprant versus placebo. While these trials enroll very large numbers of participants to be able to detect modest treatment effects (15% relative risk reductions), they illustrate the importance of having many sites, and selected sites, involved. The National Cancer Institute Cooperative Group (now the National Clinical Trials Network) [15] and HIV/AIDS Clinical Trials Networks [16] provide other examples of the rich history of multicenter trials.
- 2.
A multicenter study may enable a more generalizable sample of the study population. Although no trial is completely representative, geography, race, socioeconomic status, and lifestyle of participants may be more similar to the general population if participants are enrolled by many centers. These factors may be important in the ability to generalize the findings of the trial. Concern has been raised that site selection for practical purposes like improving enrollment could negatively impact on generalizability of results [17].In the GUSTO trial, 23,000 participants were enrolled in the United States, and most of these were enrolled over a 1-year period [10]. During that year of 1992, it has been estimated that nearly 10% of all patients in the country with acute myocardial infarction treated with fibrinolytic therapy were enrolled in the trial. The participants in this “pragmatic” trial with few exclusion criteria were well represented by high-risk groups such as the elderly (12% were at least 75 years of age, and the oldest was 110 years old) [18].Another example of the need to anticipate how participant make-up may affect generalizability is in racial distribution. For instance, it is known that hypertension and its treatment response may vary according to race. A study of participants with hypertension from either a totally black or totally white community is likely to yield findings that may not necessarily be applicable to a more diverse population. Anticipating this, there was a special effort to enroll black participants in the Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Ultimately, 35% of participants in ALLHAT were black, [19] which allowed for exploration of racial heterogeneity of intervention effects.
- 3.
A multicenter study enables investigators with similar interests and skills to work together on a common problem. Science and medicine, like many other disciplines, are competitive. Nevertheless, most major research accomplishments in clinical medicine now require a collaborative team approach. A multicenter trial also gives capable, clinically-oriented persons, who might otherwise not become involved in research activities, an opportunity to contribute to science. In the early years, multicenter clinical trials typically involved only major academic centers. Now, many clinical practices based in the community successfully participate in trials, and in many trials, organized community hospitals or clinics are the best enrollers.
Conduct of Multicenter Trials
One of the earlier multicenter
clinical trials was the Coronary Drug Project [5]. This study provided an initial model for many
of the techniques currently employed. As in all active disciplines,
concepts are frequently changing and some techniques have been
refined in subsequent trials. The following series of steps, a
distillation of experience from a number of studies, is one
reasonable way to approach the planning and conduct of a
multicenter trial.
First, a planning committee should be
established to be responsible for organizing and overseeing the
various phases of the study (planning, participant recruitment,
participant follow-up, phase out, data analysis, paper writing) and
its various centers and committees. This group often consists of
representatives from the sponsoring organization (e.g., government
agencies, private research organizations, educational institutions,
private industry), with input from appropriate consultants. Use of
consultants who are expert in the field of study, in biostatistics,
and in the management of multicenter clinical trials is encouraged.
The planning committee needs to have authority in order to operate
effectively and for the study to function efficiently.
Second, to determine the feasibility of
a study, the planning committee should make a thorough search of
the literature and review of other information. Sample size
requirements should be calculated. Reasonable estimates must be
made regarding control group event rate, anticipated effect of
intervention, and participant adherence to therapy. The planning
committee also has to evaluate key issues such as participant
availability, availability of competent cooperating investigators,
timeliness of the study, possible competing trials, regulatory
requirements, and total cost. After such an assessment, is the
trial worth pursuing? Are there sufficient preliminary indications
that the intervention under investigation indeed might work? On the
other hand, is there so much suggestive (though inconclusive)
evidence in favor of the new intervention that it might be
difficult ethically to allocate participants to a control group?
Might such suggestive evidence seriously impede participant
recruitment? Since planning for the study may take a year or more,
feasibility needs to be constantly re-evaluated, even up to the
time of the actual start of participant recruitment. New or
impending evidence may at any time cause cancellation,
postponement, or redesign of the trial. In some instances, a pilot,
or feasibility study is useful in answering specific questions
important for the design and conduct of a full-scale trial.
Third, multicenter studies require not
only clinical centers to recruit participants, but also one or two
coordinating centers to help design and manage the trial and to
collect and analyze data from all other centers. There may be
regional sites, academic centers that serve as academic research
organizations, or contract research organizations that conduct site
visits and receive data from the clinical centers. Additional
centers are often needed to perform specialized activities such as
key laboratory tests, imaging, and distributing study drugs. While
the specialized centers may perform multiple services, it may not
be advisable to permit a clinical center to perform these services.
If a specialized center and a clinical center are in the same
institution, it may be important for each to have a separate staff
in order to protect against unblinding and, therefore, bias. Even
if unblinding or bias is avoided, there might be criticism that
such a bias might have occurred and thus raise unnecessary
questions about the entire clinical trial.
As reported by Croke [20], a major consideration when selecting
clinical center investigators is availability of appropriate
participants. Although this report is now old, the message remains
relevant. The trial has to go where the participants are. Clearly,
experience in clinical trials and scientific expertise are
desirable features for investigators, but they are not crucial to
overall success. Well-known scientists who add stature to a study
are not always successful in collaborative ventures. The chief
reason for this lack of success is often their inability to devote
sufficient time to the trial. In a comprehensive study of factors
associated with enrollment of eligible participants with documented
myocardial infarction, Shea et al. [21] found positive correlations with
institutions in which patients were cared for by staff other than
private attending physicians and with the presence of a committed
nurse-coordinator. While many factors have been associated with
successful enrollment, none is more revealing than prior
performance in conducting collaborative trials.
The selection of the coordinating
center is of utmost importance. This is often a single entity, but
sometimes the coordinating center functions are split between two
or more units; a clinical coordinating center, a data coordinating
center, and, often, a separate data analysis center. The
responsibilities described here apply to any of the models, but
clearly communication becomes more of an issue when there are
multiple units.
In addition to helping design the
trial, the coordinating center, or combination of centers, is
responsible for implementing the randomization scheme; carrying out
day-to-day trial activities; and collecting, monitoring, editing,
and analyzing data. The coordinating center, or, when there are two
or more units, the clinical coordinating center/data management
center needs to be in constant communication with all other
centers. Its staff has to have expertise in areas such as
biostatistics, computer technology, epidemiology, regulatory
policy, medicine, and management in order to respond expeditiously
to daily problems that arise in a trial. These might range from
simple questions, such as how to code a particular item on a
questionnaire, to monitoring clinical site conduct. The single
coordinating center, or the separate data analysis center, has
responsibilities such as preparing data monitoring guidelines,
conducting data analyses, and developing or modifying statistical
methods. The staffs at these centers must be experienced, capable,
responsive, and dedicated in order to handle their workloads in a
timely fashion. A trial can succeed despite inadequate performance
of one or two clinical centers, but a poorly performing
coordinating center or data management center can materially affect
the success of a multicenter trial. In extreme cases, a
coordinating center may have to be changed midway through the
trial. This causes serious delay and logistical problems. Thus,
proper selection of the coordinating center is extraordinarily
important.
A key element in any coordinating or
analysis center is not only the presence of integrity, but the
appearance of integrity. Any suspicion of conflict of interest can
damage the trial. This is one of the reasons that pharmaceutical
firms who support trials sometimes use outside institutions or
organizations as coordinating centers. Because the personnel in the
centers control the data and the analyses, they should be seen to
have no overriding interest in the outcome of a trial. Meinert
[22] has described the functions
of the coordinating center in detail. See also Fisher et al. for a
description of the operations of an independent data analysis
center [23].
As noted, certain functions in a
multicenter trial are best carried out by properly selected special
centers. The advantages of centrally performing laboratory tests,
reading x-rays, evaluating pathology specimens, or coding
electrocardiograms include unbiased assessment, standardization and
reduced variability, ease of quality control, and high-quality
performance. The disadvantages of centralized determinations
include the cost and time required for shipping, as well as the
risk of losing study material. Even with electronic transfer of
data, glitches may occur. It is also obvious that the centers
selected to perform specialized activities need expertise in their
particular fields. Equally important is the capacity to handle the
large workloads of a multicenter trial with research-level quality.
Despite careful selection of these centers, backlogs of work are a
frequent source of frustration during the course of a trial.
Fourth, it is preferable for the
planning committee to provide prospective investigators with a
fairly detailed outline of the key elements of the study design as
early as possible. This results in more efficient initiation of the
trial and allows each investigator to better plan staffing and cost
requirements. Rather than presenting a final protocol to the
investigators, we recommended that all or selected representatives
be given time to discuss and, if necessary, modify the trial
design. This process allows them to contribute their own ideas, to
have an opportunity to participate in the design of the trial,
strengthening their commitment to it, and to become familiar with
all aspects of the study. It may also improve the design. The
investigators need a protocol that is acceptable to them and their
colleagues at their local institution. This “buy-in” will improve
participant recruitment, data collection, and final acceptance of
the trial results. Depending on the complexity of the trial,
several planning sessions prior to the start of participant
recruitment may be needed for this process.
If there are many investigators and a
number of difficult protocol decisions, it is useful to have
specific groups or subsets of investigators address these issues
during the planning stage. Working groups can focus on individual
problems and prepare reports for the total body of investigators.
Of course, if the initial outline has been well thought out and
developed, few major design modifications will be necessary. Any
design change needs to be carefully examined to ensure that the
basic objectives and feasibility of the study are not threatened.
This caveat applies particularly to modifications of participant
eligibility criteria. Investigators are understandably concerned
about their ability to enroll a sufficient number of participants.
In an effort to make recruitment easier, they may favor less
stringent eligibility criteria. Any such decisions need to be
examined to ensure that they do not have an adverse impact on the
objectives of the trial and on sample size requirements. The
benefit of easier recruitment may be outweighed by the need for a
larger sample size. Planning meetings also serve to make all
investigators aware of the wide diversity of opinions. Inevitably,
compromises consistent with good science must be reached on
difficult issues, and some investigators may not be completely
satisfied with all aspects of a trial. However, all are usually
able to support the final design. All investigators in a
cooperative trial must agree to follow the common study
protocol.
Although a good protocol will provide
guidance for all major issues that are anticipated, investigators
will always have questions as they begin a trial that need to be
addressed in a systematic way. This information should be shared
with all investigators, in newsletters, in a question and answer
format that could exist on a website, or (when necessary) with
protocol amendments. This is part of a broader theme in multicenter
trials: the importance of effective communication. It is the
responsibility of coordinating centers to keep in frequent contact
(by telephone, e-mail, texting, visits) with all the enrolling
centers. An informative and interactive website can be helpful. The
study leaders also need to maintain contact with the various
centers and committees, closely monitoring the conduct of the
trial.
Fifth, an organizational structure for
the trial should be established with clear areas of responsibility
and lines of authority and communication. Many have been developed
[24–26]; the one outlined below has stood the test
of time.
Steering Committee—This committee
provides scientific direction for the study at the operational
level. Its membership may be made up of some or all of those who
were on the planning committee (including sponsor representation),
plus a subset of investigators participating in the trial. In
international trials, it has become conventional to have at least
one “national coordinator” investigator from each major country to
represent those investigators and to address country-specific
issues. Depending on the length of the study, some investigators
may be chosen or elected for part of the trial. Subcommittees are
often established to consider on a study-wide level specific issues
such as adherence, quality control, classification of response
variables, and publication policies and review and then report to
the Steering Committee.
It may also be important to authorize
a small subgroup to make executive decisions between Steering
Committee meetings. These committees are sometimes referred to as
executive committees or as operations committees. Most
“housekeeping” tasks and day-to-day decisions can be more easily
accomplished in this manner. A large committee, for example, is
unable to monitor a trial on a daily basis, write memoranda, or
prepare agendas. Since committee meetings can rarely be called at
short notice, issues requiring rapid decisions must be addressed by
an executive group. It is important, however, that major questions
be discussed with the investigators.
Subcommittees—Often, subcommittees of
the Steering Committee are established. For example, there is often
the need for a system for central evaluation of events, and this
could be done by an Events Classification Subcommittee.
Adjudication of events, with the participants’ identities and
intervention groups blinded, helps to assure unbiased
classification of reported events and to ensure consistent
application of criteria for particular events. Other subcommittees
might look for ways to improve participant accrual or adherence. In
some trials, the subcommittee structure has become too complex and
can lead to inefficiencies. Trials with few centers function best
with a simple structure. If committees, subcommittees, and task
forces multiply, the process of handling routine problems becomes
difficult. Studies that involve multiple disciplines especially
need a carefully thought out organizational structure.
Investigators from different fields tend to look at issues from
various perspectives. Although this variety can be beneficial,
under some circumstances it can obstruct the orderly conduct of a
trial. Investigators may seek to increase their own areas of
responsibility and, in the process, change the scope of the study.
What starts out as a moderately complex trial can end up being an
almost unmanageable undertaking.
Data
Monitoring Committee—This scientific body, which goes by
various names (see Chap. 16), should be independent of the
investigators and any sponsor of the trial. Its primary role, to
the extent possible, is to ensure participant safety and study
integrity. To accomplish this, it is charged with reviewing and
approving the protocol; periodically monitoring baseline, harmful
effects, and response variable data; and evaluating center
performance [27]. In light of
concerns about clinical-trial integrity, [28–30] the
independence of this group is especially important. It usually
reports to either the study sponsor or the chairperson of the
planning or steering committee. The coordinating or data analysis
center should present tabulated and graphic data and appropriate
analyses to the data monitoring committee for review. The committee
has the responsibility to recommend early termination in case of
unanticipated harm, greater-than-expected benefit, or high
likelihood of indifferent results (see Chap. 16). Members of this committee should
be knowledgeable in the field under study, in clinical trials
methodology, and in biostatistics. An ethicist and/or a participant
advocate may also be part of this group. The responsibilities of
the monitoring committee to the participants, as well as to the
integrity of the study, should be clearly established and
communicated to the participants. These responsibilities for
participant safety are particularly important in double-blind
studies, since the individual investigators are unaware of the
group assignments and which group is associated with various
adverse events.
Unfortunately, the organizational
structure of many trials conducted by industry excludes meaningful
involvement of independent experts in trial design, conduct, and
analysis. There is a need for academic trialists, including those
at agencies such as the National Institutes of Health, to work with
the public and health care providers to advance the conduct,
quality, and relevance of clinical trials that address health care
priorities [31].
Sixth, despite special problems,
multicenter trials should try to maintain standards of quality that
do not differ from those in carefully conducted single-center
trials (see Chap. 11). Strong emphasis should be placed
on training and standardization so that the protocol is carried out
in the intended fashion across centers and regions. It is obviously
extremely important that staff at all centers understand the
protocol definitions, and how to complete forms and perform tests.
Differences in performance among centers, as well as between
individuals in a single center, are unavoidable. They can, however,
be minimized by proper training, certification procedures,
retesting, and when necessary, retraining of staff. An attractive,
functional, interactive website with updated training materials and
other resources is an important tool in large trials. The National
Heart, Lung, and Blood Institute (NHLBI)-sponsored International
Study of Comparative Health Effectiveness with Medical and Invasive
Approaches (ISCHEMIA) trial website provides such an example
[32]. These efforts need to be
implemented before a trial gets underway (See Chap. 11 for a discussion of quality
control). In trials that require specific training and expertise, a
clinical center should not be allowed to begin enrolling
participants until it has demonstrated the capability of performing
necessary procedures. Investigator meetings are generally important
to the successful conduct of the trial because they provide
opportunities to discuss common problems and review proper ways to
collect data and complete study forms.
Seventh, there needs to be close
monitoring of the performance of all centers. Participant
recruitment, quality of data collection and processing, quality of
laboratory procedures, adherence of participants to protocol, and
loss to follow-up should be evaluated on an ongoing basis.
Regulatory requirements for investigators are outlined in Chap.
22. Table 21.1 lists some of the
major responsibilities of the principal investigator at enrolling
centers.
Table
21.1
Principal investigator’s major
responsibilities at research sites
|
1. Be familiar with ethical principles
(see Chap. 2), including as outlined in the
Belmont Report: respect for persons, beneficence, justice
|
|
2. Be familiar with US federal regulations
as defined in the Code of Federal Regulations (CFR) (see Chap.
22)
|
|
(a) Health and Human Services, for
federally funded research
|
|
i. 45 CRF 46 (Health and Human
Services, Protection of Human Research Subjects)
|
|
1. Subpart B (pregnant women)
|
|
2. Subpart C (prisoners)
|
|
3. Subpart D (children)
|
|
(b) FDA, for FDA regulated products
|
|
i. 21 CFR 50 (informed consent)
|
|
ii. 21 CFR 54 (financial
disclosure)
|
|
iii. 21 CRF 56 (IRB)
|
|
(c) Health and Human Services, for
privacy including for all human subjects research
|
|
i. 45 CFR 46, 160, 164 (HIPAA)
|
|
3. Understand the requirements of the
responsible Institutional Review Board (IRB) and the need to follow
them
|
|
4. Be responsible for oversight of the
trial and delegation of research responsibilities, with appropriate
training and experience of staff
|
|
5. Recruit participants in a fair and
equitable way (see Chap. 10)
|
|
6. Develop process of informed consent
(see Chap. 2), with IRB approval of that
process, with consent obtained by the PI or a delegated research
staff member who is identified as “key personnel” in the IRB
approval; and maintain documentation of informed consent (generally
for at least 3 years)
|
|
7. Do not enroll patients without prior
IRB approval, and not make changes to the protocol without prior
IRB approval
|
|
8. Comply with reporting requirements of
adverse events, protocol deviations, unanticipated problems
involving risks to participants or others, or irregularities (like
loss of consent documentation) (see Chap. 12)
|
|
9. Be available, or have a designated
research staff member available, to participants to answer
questions
|
|
10. Notify the IRB and seek approval for
change to a new principal investigator
|
Electronic tracking tools allow this
to be done in an efficient and systematic way, as long as
standardized reports effectively capture and display the
information. It is important to track overall performance as well
as performance by center.
Many industry-sponsored multicenter
trials that employ contract research organizations conduct
extensive auditing and quality assurance. This is quite costly and
how much benefit it provides has been questioned [33]. See Chap. 11 for further discussion of this
topic.
In most clinical trials, recruitment
of participants is difficult. In a cooperative clinical trial,
however, there is an opportunity for some clinical centers to
compensate for the inadequate performance of other centers by
exceeding their predetermined recruitment goals. The clinical
centers should understand that, while friendly competition keeps
everybody working, the real goal is overall success, and what some
centers cannot do, another perhaps can. Therefore, it is important
to encourage the good centers to recruit as many participants as
possible. There may be a limit, however, if one center, region, or
country (in the case of international trials) starts to dominate
enrollment. At some point, recruitment might need to be capped if
the study is to be seen as truly multicenter.
Eighth, publication, presentation, and
authorship policies should be agreed upon in advance. Authorship
becomes a critical issue when there are multiple investigators,
many of whose academic careers depend on publications. There is no
completely satisfactory way to recognize the contribution of each
investigator. A common compromise is to put the study name
immediately under the paper title and to acknowledge the writers of
the paper, either in a footnote or under the title, next to the
study name. All key investigators are then listed at the end of the
paper. The policy may also vary according to the type of paper
(main or subsidiary). The group authorship of manuscripts from
multicenter trials was challenged by some medical journals and
defended by others [34–36]. It
remains common, but typically with an identified writing committee
to take responsibility (see Chap. 20).
Involvement of representatives of the
sponsor as authors of the main manuscripts from a major trial can
be contentious, especially if it is a commercial firm that stands
to benefit from a favorable presentation of the trial results. Most
sponsors accept a hands-off policy and leave it to the
investigators to write the scientific papers, although including
sponsor members of the research team who provided important
intellectual contributions can be appropriate. Typically, an
industry sponsor is given 1 month to preview the main results
manuscript, to allow time to deal with patent or regulatory issues.
This review should not unnecessarily delay the publication of the
main trial results. Regrettably, there are examples of interference
that are in conflict with academic freedom. These policies should
be clearly defined in the contract between the sponsor and the
investigators.
In one four-center trial, the
investigators at one of the centers reported their own findings
before the total group had an opportunity to do so [37, 38]. Such
an action is not compatible with a collaborative effort. It
undermines the goal of a multicenter trial of having enough
participants to answer a question and, perhaps more importantly,
the trust among investigators. Academic institutions have taken a
strong stand against this principle of collaboration and in defense
of academic freedom for each investigator. However, we believe that
those unwilling to abide by the rule for common authorship should
not participate in collaborative studies.
Creating a publication charter in
advance and having all parties agree to abide by it provides
important protection against misunderstandings. However, fair
recognition of junior staff will always be difficult
[39]. Study leadership often gets
credit and recognition for work done largely by people whose
contributions may remain unknown to the scientific community. One
way to alleviate this problem is to appoint as many capable junior
staff as possible to subcommittees. Such staff should also be
encouraged to develop studies ancillary to the main trial. This
approach will enable them to claim authorship for their own work
while using the basic structure of the trial to get access to
participants and supporting data. Such ancillary studies may be
performed on only a subgroup of participants and may not
necessarily be related to the trial as a whole. Care must be taken
to ensure that they do not interfere with the main effort, either
through unblinding, by harming the participants, or by causing the
participants to leave the trial. Sackett and Naylor discuss the
issues for and against allowing publication of ancillary studies
before the main trial is completed [40].
Globalization of Trials
As noted earlier, many multicenter
clinical trials are international; there are several reasons for
this. One, it provides greater numbers of potential participants,
allowing for quicker accrual. Two, the broader populations may
allow for wider generalization of results. It is not simply people
from one country with one medical care system who are enrolled; the
data from the trial apply to many sorts of people with very
different medical systems. Three, it may be easier and less
expensive to screen people in some regions. Even in NIH-sponsored
trials, an increasing proportion of participants are being enrolled
internationally, largely due to inability to enroll enough patients
at centers in the United States [41] (see Fig. 21.1).
Fig.
21.1
International enrollment in NIH-sponsored
randomized trials of coronary disease [41]
There are, however, limitations and
concerns with globalization of trials. As discussed in Chap.
2, the ethics of enrolling
participants from underdeveloped countries or areas can be
problematic [42]. It is unethical
to enter people into a trial simply to save money, or because the
regulatory oversight is less rigorous, when there is little
likelihood that the population will benefit from or have access to
the trial intervention. Logistics of implementing international
trials may be daunting. In addition to multi-language
communication, there is the issue of translating forms and
questionnaires. Not all forms, particularly those that have been
validated in certain groups, may be usable in very different
communities and cultures. Transporting drugs and other materials
across borders may not be simple. In addition, each country has its
own regulatory structure that must be negotiated.
Some countries may present particular
challenges in regulatory approval, such as China where the process
may take over a year for drug trials. In India, concerns over
unethical trial practices have led to laws requiring trial sponsors
to cover medical costs of trial-associated adverse outcomes and to
requirements to video record the informed consent process. These
regulatory and legal requirements resulted in unwillingness to
include Indian sites in many trials. At least 35 NIH trials were
put on hold in India in 2013, although many subsequently resumed
[43, 44].
Interpretation of regionally diverse
results may be questioned. Are the overall results relevant to all
countries? Does the culture, social structure, or medical care
system (including concomitant medications and other treatment)
affect the outcome? Does each trial participant need minimal
standard background care? If so, this must be specified in advance
in the protocol. An example of a trial that examined effect by
geography is the Platelet IIb/IIIa in Unstable angina: Receptor
Suppression Using Integrilin Therapy (PURSUIT) trial
[45]. Relative reductions in the
primary response variable (a combination of death or myocardial
infarction) varied among geographic regions. In trials of beta
blockers in heart failure, there appears to have been a
consistently lesser treatment effect in the United States than in
other countries, for unclear reasons [46] (see Fig. 21.2).
Fig.
21.2
Effects of beta blockers on all-cause
mortality in major heart failure trials [46]. Panel (a) is overall, panel (b) contrasts the United States and rest
of world (ROW), with p values in panel (b) for effects in ROW, and NS referring
to non-significant p-value for the overall effect in the U.S.
The Treatment of Preserved Cardiac
Function Heart Failure With an Aldosterone Antagonist (TOPCAT)
trial included about half of patients from Russia and Georgia and
half from the Americas. The populations in these regions differed
at baseline, with more of the patients in Russia and Georgia being
characterized by prior myocardial infarction and prior heart
failure hospitalization [47].
There was a four-fold higher rate of the primary outcome of
cardiovascular death, aborted cardiac arrest, or heart failure
hospitalization in Russia and Georgia than in the Americas, and the
treatment benefit observed with spironolactone in the Americas was
not seen in Russia and Georgia. Spironolactone did not have the
degree of effect on laboratory values (potassium and creatinine) in
Georgia and Russia that it did in the Americas [48]. Related regional differences in composition
and outcome of populations with heart failure with preserved
ejection fraction have been observed in other clinical trials
[49]. These findings suggest that
the diagnosis, management, outcomes, and response to therapy of
heart failure with preserved ejection fraction may be different in
different geographic regions. This regional heterogeneity will in
turn impact on the results of clinical trials.
The Platelet inhibition and Patient
Outcome (PLATO) trial, which studied ticagrelor versus clopidogrel
after acute coronary syndromes, provides an even more striking
example of heterogeneity of treatment effect according to country
[50]. Overall, there was a 16%
relative risk reduction in the primary composite outcome. In the
United States, which included 8% of the patients, the hazard ratio
was 1.27, and in other countries, 0.81, with a p value for the
interaction of 0.0095 [51]. There
is also evidence that it may be related to, and even explained by,
the higher dose aspirin used in the United States [51].
In these examples, chance still may be
the most likely explanation. However, investigators need to
consider, in advance, whether combining results from geographically
and culturally different sites is appropriate. In any case, if a
robust and consistent treatment effect is desired to be
demonstrated in the United States population, or in any specific
population, enrolling a sufficiently large portion from that
population is important. Vickers et al. [52] found that some countries tended to produce
results more favorable to the new intervention than other
countries, though publication bias was the likely reason.
Large, Simple Trials
Large, simple trials [53] are a subset of multicenter trials that
typically involve a large number of participating centers, many of
which are non-academic institutions representative of general
practice. Education, training, and standardization may need to be
more focused and streamlined compared with other trial models. For
example, in streamlined trials background care may be left to the
caring physician such that standard of care is the goal, although
for many trials, encouraging high quality standard of care may be
important for the results to be accepted as relevant.
Clinician-investigators need to understand the basic concepts and
intent of clinical trials and how the rules of research, which may
sometimes seem arbitrary, [54]
differ from the way they practice medicine (See Chap. 2). The reliance on hard endpoints
such as all-cause mortality, and limited data collection, tends to
reduce the need for elaborate quality control procedures.
Successful conduct of streamlined
trials has become more difficult with more complex and
heterogeneous regional regulatory requirements, which have caused
large trials to be very expensive. The expense related to
complexity and various barriers that do not result in improved
quality has far reaching consequences, including resulting in an
inability to conduct many trials that are necessary to guide
clinical care. In response to these barriers, recommendations have
been made to simplify procedures for large, simple trials
[55–57].
The U.S. Food and Drug Administration
(FDA), in partnership with Duke University in the Clinical Trials
Transformation Initiative (CTTI), [58] has made a concerted effort to provide
guidance to promote streamlining when appropriate. For example, in
December 2012, a guidance was issued for simplified adverse event
reporting in large, simple trials [59] and in August 2013, another guidance was
issued for risk-based monitoring of trial conduct and data
[60].
There are examples in which randomized
clinical trials have been successfully conducted on the platform of
clinical registries [61], such as
the Thrombus Aspiration during ST-segment Elevation myocardial
infarction (TASTE) trial that randomized over 7000 patients in less
than 3 years (80% of all eligible acute myocardial infarctions in
Sweden during the enrollment period) to thrombus aspiration or
control for an estimated total US$300,000 marginal cost
[62] (see Chap. 10, Fig. 10.4). Another pragmatic trial,
INforming Fresh versus Old Red cell Management (INFORM), is
planning to randomize 31,497 patients undergoing blood transfusion
to freshest versus standard (older) blood for transfusion at five
medical centers in Canada, Australia, and the United States as of
December 2014. In this trial, consent is waived and in-hospital
data are collected using the electronic health record such that the
cost is a fraction of what would be typical for a trial of this
size [63].
Another example of a streamlined
approach to integrating clinical trials and clinical practice comes
from the NIH Health Care Systems Research Collaboratory. Launched
in 2006, this program supports demonstration projects in which
health care organizations partnered with researchers conduct
pragmatic clinical trials in everyday health care settings. One
such project was The Randomized Evaluation of Decolonization versus
Universal Clearance to Eliminate MRSA (REDUCE MRSA), a cluster
randomized trial of 43 hospitals (including 74 intensive care units
and 74,256 patients) testing whether daily antiseptic baths and a
nasal antibiotic were more effective than other procedures to
decolonize patients to prevent staphylococcal infections in
intensive care units [64]. The
Collaboratory group has outlined key steps to develop successful
partnerships between health care systems and researchers to conduct
pragmatic clinical trials that address important gaps in knowledge
to improve patient care. These steps include building partnerships,
defining the important questions, assessing feasibility, involving
stakeholders in the design, and implementing workflow
[65].
References
1.
Fleiss JL. Multicentre
clinical trials: Bradford Hill’s contributions and some subsequent
developments. Stat Med
1982;1:353–359.
2.
Greenberg BG. Conduct of
cooperative field and clinical trials. Am Stat 1959;13:13–28.
3.
Klimt CR. Principles of
multi-center clinical studies. In Boissel JP, Klimt CR.
Multi-center Controlled Trials.
Principles and Problems. Paris: INSERM, 1979.
4.
Levin WC, Fink DJ, Porter S,
et al. Cooperative clinical investigation: a modality of medical
science. JAMA
1974;227:1295–1296.
5.
Coronary Drug Project
Research Group. PL Canner (ed.). The Coronary Drug Project: methods
and lessons of a multicenter clinical trial. Control Clin Trials
1983;4:273–541.
6.
ISIS-1 (Second International
Study of lnfarct Survival) Collaborative Group. Randomised trial of
intravenous atenolol among 16027 cases of suspected acute
myocardial infarction: ISIS-1. Lancet 1986;328:57–66.
7.
ISIS-2 (Second International
Study of lnfarct Survival) Collaborative Group. Randomised trial of
intravenous streptokinase, oral aspirin, both, or neither among 17
187 cases of suspected acute myocardial infarction: ISIS-2.
Lancet
1988;332:349–360.
8.
Schulz S, Mehilli J, Schömig
A, Kastrati A. ISAR—a story of trials with impact on practice.
Circ J
2010;74:1771–1778.
9.
Schulz S, Richardt G,
Laugwitz KL, et al. Comparison of prasugrel and bivalirudin vs
clopidogrel and heparin in patients with ST-segment elevation
myocardial infarction: Design and rationale of the Bavarian
Reperfusion Alternatives Evaluation (BRAVE) 4 trial. Clin Cardiol 2014;37:270–276.
10.
The GUSTO Investigators. An
international randomized trial comparing four thrombolytic
strategies for acute myocardial infarction. N Engl J Med 1993;329:673–682.
11.
Women’s Health Initiative.
Available at http://www.nhlbi.nih.gov/whi/.
Accessed January 14, 2015.
12.
Rossouw JE, Anderson GL,
Prentice RL, et al. Risks and benefits of estrogen plus progestin
in healthy postmenopausal women: principal results from the Women’s
Health Initiative randomized controlled trial. JAMA 2002;288:321–333.
13.
Roth JA, Etzioni R, Waters
TM, et al. Economic return from the Women’s Health Initiative
estrogen plus progestin clinical trial: A modeling study.
Ann Intern Med
2014;160:594–602.
14.
HPS2-THRIVE Collaborative
Group, Landray MJ, Haynes R, et al. Effects of extended-release
niacin with laropiprant in high-risk patients. N Engl J Med 2014; 371:203–212.
15.
National Cancer Institute
National Clinical Trials Network. Available at http://www.cancer.gov/clinicaltrials/nctn.
Accessed January 14, 2015.
16.
HIV/AIDS Clinical Trials
Networks. Available at
http://www.niaid.nih.gov/about/organization/daids/Networks/Pages/daidsnetworks.aspx.
Accessed January 14, 2015.
17.
Gheorghe A, Roberts TE, Ives
JC, Fletcher BR, Calvert M. Centre selection for clinical trials
and the generalisability of results: a mixed methods study.
PLoS One
2013;8:e56560.
18.
Katz A, Cohn G, Mashal A,
Cristal N. Thrombolytic therapy for acute myocardial infarction in
a 110-year-old man. Am J
Cardiol 1993;71:1122–1123.
19.
ALLHAT Officers and
Coordinators for the ALLHAT Collaborative Research Group. Major
outcomes in high-risk hypertensive patients randomized to
angiotensin-converting enzyme inhibitor or calcium channel blocker
vs diuretic. The Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial (ALLHAT). JAMA 2002;288:2981–2997.
20.
Croke G. Recruitment for the
National Cooperative Gallstone Study. Clin Pharmacol Ther
1979;25:691–694.
21.
Shea S, Bigger JT, Campion
J, et al. Enrollment in clinical trials: institutional factors
affecting enrollment in the Cardiac Arrhythmia Suppression Trial
(CAST). Control Clin Trials
1992;13:466–486.
22.
Meinert CL. Clinical Trials.
Design, Conduct and
Analysis. Second Edition. New York: Oxford University Press,
2012.
23.
Fisher MR, Roecker EB,
DeMets DL. The role of an independent statistical analysis center
in the industry-modified National Institutes of Health model.
Drug Inf J
2001;35:115–129.
24.
Byington RP, for the
Beta-Blocker Heart Attack Trial Research Group. Beta-Blocker Heart
Attack Trial: design, methods, and baseline results. Control Clin Trials
1984;5:382–437.
25.
Meinert CL. Organization of
multicenter clinical trials. Control Clin Trials
1981;1:305–312.
26.
Carbone PP, Tormey DC.
Organizing multicenter trials: lessons from the cooperative
oncology groups. Prev Med
1991;20:162–169.
27.
Friedman L, DeMets D. The
data monitoring committee: how it operates and why. IRB 1981;3:6–8.
28.
Fleming TR, DeMets DL.
Monitoring of clinical trials: issues and recommendations.
Control Clin Trials
1993;14:183–197.
29.
Angell M, Kassirer JP.
Setting the research straight in the breast-cancer trials.
N Engl J Med
1994;330:1448–1450.
30.
Cohen J. Clinical trial
monitoring: hit or miss? Science 1994; 264;1534–1537.
31.
DeMets DL, Califf RM. A
historical perspective on clinical trials innovation and
leadership: where have the academics gone? JAMA 2011;305:713–714.
32.
International Study of
Comparative Health Effectiveness with Medical and Invasive
Approaches (ISCHEMIA) trial. Available at https://www.ischemiatrial.org/.
Accessed January 14, 2015.
33.
Eisenstein EL, Lemons PW II,
Tardiff BE, et al. Reducing the costs of phase III cardiovascular
clinical trials. Am Heart J
2005;149:482–488.
34.
Kassirer JP, Angell M. On
authorship and acknowledgments. N
Engl J Med 1991;325:1510–1512.
35.
Goldberg MF. Changes in the
archives. Arch Ophthalmol
1993;111:39–40.
36.
Meinert CL. In defense of
the corporate author for multicenter trials. Control Clin Trials
1993;14:255–260.
37.
Winston DJ, Ho WG, Gale RP.
Prophylactic granulocyte transfusions during chemotherapy of acute
nonlymphocytic leukemia. Ann
Intern Med 1981;94:616–622.
38.
Strauss RG, Connett JE, Gale
RP, et al. A controlled trial of prophylactic granulocyte
transfusions during initial induction chemotherapy for acute
myelogenous leukemia. N Engl J
Med 1981; 305:597–603.
39.
Remington RD. Problems of
university-based scientists associated with clinical trials.
Clin Pharmacol Ther 1979;
25:662–665.
40.
Sackett DL, Naylor CD.
Should there be early publication of ancillary studies prior to the
first primary report of an unblinded randomized clinical trial?
J Clin Epidemiol
1993;46:395–402.
41.
Kim ES, Carrigan TP, Menon
V. International participation in cardiovascular randomized
controlled trials sponsored by the National Heart, Lung, and Blood
Institute. J Am Coll
Cardiol 2011;58:671–676.
42.
Glickman SW, McHutchison JG,
Peterson ED, et al. Ethical and scientific implications of the
globalization of clinical research. N Engl J Med 2009;360:816–823.
43.
Kakkar AK. India puts
informed consent on camera. Science 2014;344:150–151.
44.
Reardon S. NIH makes wary
return to India. Nature
2014;506:143.
45.
Akkerhuis KM, Deckers JW,
Boersma E, et al, for the PURSUIT Investigators. Geographic
variability in outcomes within an international trial of
glycoprotein IIb/IIIa inhibition in patients with acute coronary
syndromes: results from PURSUIT. Eur Heart J 2000;21:371–381.
46.
O’Connor CM, Fiuzat M,
Swedberg K, et al. Influence of global region on outcomes in heart
failure β-blocker trials. J Am
Coll Cardiol 2011;58:915–922.
47.
Pitt B, Pfeffer MA, Assmann
SF, et al. Spironolactone for heart failure with preserved ejection
fraction. N Engl J Med
2014;370:1383–1392.
48.
Pfeffer MA, Claggett B,
Assmann SF, et al. Regional variation in patients and outcomes in
the treatment of preserved cardiac function heart failure with an
aldosterone antagonist (TOPCAT) trial. Circulation 2015;131:34–42.
49.
Kristensen SL, Køber L,
Jhund PS, et al. International geographic variation in event rates
in trials of heart failure with preserved and reduced ejection
fraction. Circulation
2015;131:43–53.
50.
Mahaffey KW, Wojdyla DM,
Carroll K, et al. Ticagrelor compared with clopidogrel by
geographic region in the Platelet Inhibition and Patient Outcomes
(PLATO) trial. Circulation
2011;124:544–554.
51.
Carroll KJ, Fleming TR.
Statistical evaluation and analysis of regional interactions: The
PLATO Trial Case Study. Stat
Biopharm Res 2013;5:91–101.
52.
Vickers A, Goyal N, Harland
R, Rees R. Do certain countries produce only positive results? A
systematic review of controlled trials. Control Clin Trials
1998;19:159–166.
53.
Yusuf S, Collins R, Peto R.
Why do we need some large, simple randomized trials? Stat Med 1984;3:409–422.
54.
Sugarman J, Califf RM.
Ethics and regulatory complexities for pragmatic clinical trials.
JAMA
2014;311:2381–2382.
55.
Yusuf S, Bosch J, Devereaux
PJ, et al. Sensible guidelines for the conduct of large randomized
trials. Clin Trials
2008;5:38–39.
56.
Reith C, Landray M,
Devereaux PJ, et al. Randomized clinical trials—removing
unnecessary obstacles. N Engl J
Med 2013;369:1061–1065.
57.
Eapen ZJ, Lauer MS, Temple
RJ. The imperative of overcoming barriers to the conduct of large,
simple trials. JAMA
2014;311:1397–1398.
58.
Clinical Trials
Transformation Initiative (CTTI). Available at http://www.ctti-clinicaltrials.org/.
Accessed January 14, 2015.
59.
Guidance for Industry and
Investigators. Safety Reporting Requirements for INDs and BA/BE
Studies. Available at
http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM227351.pdf.
Accessed January 14, 2015.
60.
Guidance for Industry.
Oversight of Clinical Investigations—A Risk-Based Approach to
Monitoring. Available at
http://www.fda.gov/downloads/Drugs/…/Guidances/UCM269919.pdf.
Accessed January 14, 2015.
61.
Lauer MS, D’Agostino RB Sr.
The randomized registry trial—the next disruptive technology in
clinical research? N Engl J
Med 2013;369:1579–1581.
62.
Fröbert O, Lagerqvist B,
Olivecrona GK, et. al. Thrombus aspiration during ST-segment
elevation myocardial infarction. N
Engl J Med 2013;369:1587–1597.
63.
INforming Fresh versus Old
Red cell Management (INFORM). Available at http://www.controlled-trials.com/ISRCTN08118744.
Accessed January 14, 2015.
64.
Huang SS, Septimus E,
Kleinman K, et al. Targeted versus universal decolonization to
prevent ICU infection. N Engl J
Med 2013; 368:2255–2265.
65.
Johnson KE, Tachibana C,
Coronado GD, et al. A guide to research partnerships for pragmatic
clinical trials. BMJ
2014;349:g6826.