Comment Letter on FDA's Request
for Information on IRB Practices ("IRB Shopping")
April 8, 2002
Dockets Management Branch (HFA-305)
Food and Drug Administration
5630 Fishers Lane Room 1061
Rockville, MD 20852
Re: 21 CFR Part 56 Docket #01N-0322 Federal Register Vol.
67 #44 10115
Dear Sir/Madam:
The Association of American Medical Colleges (AAMC) welcomes
the opportunity to comment on the recent notice in the Federal
Register, and the Agency's request for "information on
IRB practices to determine whether FDA should draft a regulation…"
The AAMC represents all 125 accredited allopathic U.S. medical
schools; approximately 400 major teaching hospitals; nearly
100 professional and academic societies; and most of the nation's
medical school faculty, students and residents. Our institutions
perform over half of the extramural research sponsored by
the NIH, as well as research sponsored by other Public Health
Service and federal agencies. Our members also participate
in a great deal of industrially sponsored research on drugs,
biologics, and devices.
Along with other organizations, the AAMC recently helped
to create the Association for the Accreditation of Human Research
Protection Programs (AAHRPP). This nonprofit organization
has established standards to guide the implementation of a
nation-wide accreditation program for human research protection
programs. It is our expectation that accreditation, when coupled
with existing FDA and other government regulations, will establish
a high standard for IRB performance that will strengthen the
protection of human participants in biomedical, behavioral
and psychosocial research. While it is presently impossible
to gauge accurately the variability of IRB approval rates
for comparable research protocols, we expect over time that
variability will be substantially reduced among accredited
research organizations. In the remainder of this letter, I
address the specific questions raised in the Federal Register
1. How significant is the problem of IRB shopping?
As the ANPRM notes, while the Office of the Inspector General
of DHHS reported "a few situations" where IRB shopping supposedly
occurred, the OIG Report offered no quantitative data on the
frequency of this behavior. In response to your query, we
have asked some of our medical schools to estimate the frequency
with which investigators or sponsors submit an IRB-disapproved
protocol to a different IRB. Investigators at our medical
schools and teaching hospitals are not able to conduct research
with human participants, which has not been approved by their
institutional IRB. Investigators based at medical schools
and teaching hospitals, therefore, have no incentive to "shop"
for a more positive review of investigator-initiated research
by seeking out another IRB. Rather, as a general rule, investigators
typically respond to their IRBs' concerns by clarifying or
modifying materials in the protocol, proposed participant
recruitment conditions, and/or the informed consent form and
process. If an investigator relocates to another institution,
her/his research would have to be reviewed (or re-reviewed)
by an IRB at the new institution.
In contrast, it is possible that sponsored multi-site trials
endorsed by a single commercial IRB may have received negative
reviews from other IRBs. Before implementing a new regulation,
however, the AAMC believes that the FDA should attempt to
document the extent of the problem. In multi-site trials,
we are aware that individual institutional IRBs may raise
objections to some aspect of a protocol that has been previously
approved by another institutional or commercial IRB. Most
typically, these concern the language or process of informed
consent. Because such judgments may be more idiosyncratic
than substantive, Drs. Greg Koski (Office of Human Research
Protection) and David Lepay (FDA) have strongly argued in
favor of central IRB reviews of multisite trials. The AAMC
supports this position, provided that local institutional
IRBs understand that they remain accountable for continuing
oversight of the research performed at their sites. Given
the position of Drs. Koski and Lepay, the AAMC would favor
the recommendation that sponsors disclose prior IRB judgments
(approval or disapproval) with regard to any multisite protocol
being reviewed by a central or local IRB. This would be consistent
with the overall guidance to sponsors that they provide all
information relevant to IRB review.
2,3. Who should make and receive these disclosures?
The AAMC believes it is the responsibility of the sponsor
of a multisite clinical trial to advise an IRB of any prior
reviews of the protocol under consideration. If the research
is being conducted at a single site, it is the responsibility
of the principal investigator to inform the IRB of prior judgments
(approval or disapproval) from reviews of the same protocol
at another institution. This issue would most likely emerge
in the case of an investigator who relocates her/his research
from one research institution to another. It is the responsibility
of the investigator to inform the new IRB of any information
relevant to its initial consideration or on-going oversight
of a clinical research protocol. IRBs should routinely request
from principal investigators all information relevant to its
initial review and on-going oversight of projects.
4,5. What information should be disclosed?
The IRB should receive all relevant information about a
study, including prior positive or negative IRB judgments,
and (if known) their rationale.
6. To permit a subsequent IRB to assess the merits of a
prior IRB decision, should information about the basis for
the prior decision be disclosed?
In general, IRBs communicate by letter to investigators
regarding the reasons for their decisions. They do not share
minutes with other IRBs. It is not unreasonable for an IRB
to request receipt of any communication received by an investigator
(or sponsor) in connection with prior IRBs' reviews of a protocol
under review. If the agency is suggesting the sharing of additional
documents (such as IRB minutes), then procedures would need
to be developed to enable an IRB to disclose the minutes surrounding
the discussion of a particular protocol. One unintended consequence
of requiring the sharing of minutes from IRB deliberations
might be to make those minutes less detailed and informative
than current practice.
In this section, you may be raising concerns regarding the
wisdom of increasing mandatory documentation requirements
for IRBs. The OIG Report of 1998, that you cite as the stimulus
for possible new rule-making on disclosure requirements, also
noted the increasing burdens facing institutional IRBs, fearing
that the latter might be placing our system for protection
of human research participants at risk of failure in its primary
mission. Before imposing additional documentation requirements
on IRBs, FDA should determine whether this will strengthen
the protection of human research participants or make it more
difficult for IRBs to conduct their work.
7. How should FDA enforce the requirement?
Absent FDA documentation of significant evidence of widespread
"IRB shopping" by sponsors (or investigators) who have received
disapprovals of proposed clinical research protocols, the
AAMC believes that the FDA should not impose a new regulatory
requirement upon institutional IRBs at this time.
The AAMC would prefer that FDA remind sponsors, investigators
and IRBs of their responsibility to provide IRBs with all
information relevant to their reviews of proposed protocols,
including prior judgments by a previous IRB. We also believe
that FDA should encourage all institutions, sponsors, and
independent IRBs to seek accreditation of their human research
protection programs, and that the accreditation process should
determine whether and how the latter are meeting this responsibility.
8. Are there ways to deal with IRB shopping other than disclosure
of prior IRB reviews?
AAMC believes that changing investigator behavior and the
institutional milieu in which human subjects research is conducted
would be, in the long run, the most effective way to deal
with this problem. We believe that a voluntary program of
accreditation of human research protection programs is the
most effective way to assure the integrity of oversight of
human research programs. IRB performance needs to be evaluated
by standards that go beyond regulatory compliance. Voluntary
accreditation begins with a set of standards, and includes
peer review of performance and enhanced quality improvement
through recurrent self-study. In our view, this type of openness
and accountability is the best protection against "IRB shopping"
and other threats to the integrity of IRB reviews.
We appreciated this opportunity to comment on the ANPRM.
We believe that the issue of "IRB shopping" needs to be better
defined and quantified before new regulations and mandates
are imposed on institutional IRBs. We also believe that the
FDA should support efforts to accredit human research protection
programs as the surest way to strengthen their performance
and overall integrity.
Sincerely,
Jordan J. Cohen, MD
|
