Jan-Feb 05 IRB.qxd
E T H I C S
HUMAN RESEARCH
&
IRB
Appelbaum and colleaguesfirst described the “thera-peutic misconception” in
1982.1 There has been much dis-
cussion since then about whethe
and why some patients who ente
clinical trials confuse research with
treatment and overestimate the
nature or likelihood of benefit to
them from research in which they
enroll, and about whether investi-
gators share in or contribute to any
misunderstanding.2 The therapeutic
misconception has been examined
empirically in surveys and inter-
views,3 some of which focus on
phase I trials,4 and in one pub-
lished examination of consent
forms for phase I oncology
esearch.5 Thus, Appelbaum and
colleagues’ original focus on
whether research subjects under-
stood how study design elements
like randomization and placebo
arms could affect them has
expanded to encompass factors
characteristic of early-phase trials,
such as translation from laboratory
and animal studies to human trials
and the implications of dose escala-
tion design. However, there has
een relatively little discussion of
these and other ethical and design
questions raised by early-phase
clinical research.6
To examine how consent forms
for early-phase trials address scien-
tific uncertainty and describe
potential benefits, we analyzed 321
consent forms for gene transfe
esearch, 99% of which were
early-phase (phase I, I/II, or II) tri-
als, and 69% of which were oncol-
ogy trials. Our goal was to assess
how consent form language might
promote or reduce the therapeutic
misconception (including misesti-
mation7 of potential benefit) in
early-phase research.
We chose to examine how the
prospect of benefit is described in
gene transfer research for several
easons. This small but rapidly
growing field of clinical research is
ased on what seems to be com-
pelling scientific logic: since genes
direct vital cellular functions, then
inadequate cellular functioning
should be treatable if new copies of
healthy genes are added.8 Just 15
years old, gene transfer research
holds out both great promise and
great uncertainty; commands con-
siderable public attention, both
positive and negative; and exempli-
fies the ethical challenges of disclo-
sure in the face of unknowns,
uncertainties, and the high failure
JA N UA RY-FE B RUA RY 2005 • VO L U M E 27, NU M B E R 1
Consent Forms and the
Therapeutic Misconception: The
Example of Gene Transfe
Research
y Nancy M.P. King, Gail E. Henderson, La
y R.
Churchill, Arlene M. Davis, Sara Chandros Hull,
Daniel K. Nelson, P. Christy Parham-Vetter,
Ba
a Bluestone Rothschild, Michele M. Easter,
and Benjamin S. Wilfond 1
The Quality of Informed Consent
in a Clinical Research Study in
Thailand
y Christine Pace, Ezekiel J. Emanuel, Theshinee
Chuenyam, Chris Duncombe, Judith D. Bebchuk,
David Wendler, Jorge A. Tavel, Laura A. McNay,
Praphan Phanuphak, Heidi P. Forster, and
Christine Grady, for the ESPRIT Group 9
BOOK REVIEW:
Case Studies in Biomedical
Research Ethics
y Jacquelyn Slomka 18
ANNOTATIONS 19
INSTRUCTIONS FOR AUTHORS 20
A PUBLICATION OF
THE HASTINGS
CENTER
Consent Forms and the
Therapeutic Misconception:
The Example of Gene Transfer Research
BY NANCY M.P. KING, GAIL E. HENDERSON, LARRY R.
CHURCHILL, ARLENE M. DAVIS, SARA CHANDROS HULL, DANIEL
K. NELSON, P. CHRISTY PARHAM-VETTER, BARBRA BLUESTONE
ROTHSCHILD, MICHELE M. EASTER, AND BENJAMIN S. WILFOND
Nancy M.P. King, Gail E. Henderson, La
y R.
Churchill, Arlene M. Davis, Sara Chandros Hull,
Daniel K. Nelson, P. Christy Parham-Vetter,
Ba
a Bluestone Rothschild, Michele M. Easter,
and Benjamin S. Wilfond, “Consent Forms and
the Therapeutic Misconception: The Example of
Gene Transfer Research,” IRB: Ethics & Human
Research 27, No XXXXXXXXXX): 1-8.
JA N UA RY-FE B RUA RY 2005 IRB: ET H I C S & HU M A N RE S E A RC H
potential for early-phase research.9
Gene transfer research also receives
extra guidance and oversight;10 its
consent forms undergo greate
scrutiny and must conform to addi-
tional standards. This increased
attention might be expected to
improve the discussion of possible
enefits to potential subjects.
Defining Benefits in Clinical
Research
To characterize the potential ben-efits described in early-phase
gene transfer research consent
forms, we applied the following def-
initions. Failure to distinguish
among the types of benefits in con-
sent forms may both reflect and
contribute to conceptual confu-
sion.11
First, benefits to subjects (benefits
from study participation) should be
distinguished from benefits to socie-
ty (future benefits to science or to
future patients from research
esults). Because benefits to socie-
ty—described in federal research
egulations as contributions to gen-
eralizable knowledge—can only be
ealized in the future, they should be
eadily distinguishable from benefits
to subjects. However, when consent
forms describe the ultimate aim of
the line of research or the mecha-
nism of action of the experimental
intervention without differentiating
these from potential benefits fo
subjects in the cu
ent trial, it may
e difficult to distinguish between
enefits to subjects and benefits to
society. The following example illus-
trates this blu
ing: “Gene therapy
works by using a virus vector to
ca
y the new gene into the patient’s
cells. Once there, the new gene
makes the protein that patients like
you lack. The investigators hope
that gene therapy will be an effective
treatment for your disease.”
Benefits to subjects are furthe
divided into two types: direct bene-
fits from receipt of the experimental
intervention, and inclusion benefits
(also called collateral or indirect
enefits), which result from partici-
pating in a study regardless of
whether the subject receives the
experimental intervention. The
potential direct benefits that may be
described in a consent form
(Table 1) depend on the nature of
the experimental intervention and
the subjects’ disease or condition.
Inclusion benefits need not be so
study-specific; descriptions in con-
sent forms can encompass such
diverse items as free goods or servic-
es provided as an enrollment incen-
tive; diagnostic testing and standard
treatments provided on-study at no
cost to subjects; the opportunity to
e monitored closely by disease
experts; and sometimes, potential
psychological benefits from “doing
everything possible” for oneself
and/or for others. Although direct
and inclusion benefits are quite dif-
ferent, they are not always distin-
guished in consent forms. A consent
form statement discussing potential
enefits to the subject from “study
participation” may refer to direct
enefits, inclusion benefits, or both,
and it may be difficult to determine
which type is meant.
Study Methods
We obtained copies of all con-sent forms and portions of
protocols for human gene transfe
studies dating from 1990 through
August 2000 that are on file and
publicly available at the Office of
Biotechnology Activities (OBA) of
the National Institutes of Health
(NIH). All documents were redacted
efore analysis to delete information
that could identify individuals, spon-
sors, and institutions. Gene transfe
studies were excluded from the
analysis if their files were confiden-
tial, unavailable, or incomplete.
Studies using healthy volunteers as
subjects were also excluded, as were
gene marking studies, which provide
data on the feasibility and efficiency
of gene transfer using only genes
with no therapeutic potential. The
esulting total of 321 consent forms
epresents over 90% of non-mark-
ing gene transfer studies submitted
to OBA during the first 10 years of
human gene transfer research.
Institutional Review Boards (IRBs)
at the University of North Carolina
at Chapel Hill and the National
Human Genome Research Institute
approved the study.
All consent forms and protocol
materials were assessed with a
94-question instrument developed
iteratively using practice protocols.12
Eight investigators working in teams
of two coded the materials, with one
2
Table 1.
Nature of Direct Benefit in Consent Forms
Nature Category Definition Consent Form Examples
Contentless no nature information you may or may not
enefit; personal benefit
is not guaranteed
Su
ogate Endpoint laboratory measurement tumor shrinkage,
that stands in statistically decrease in PSA,
for a clinical endpoint increased % Factor IX
in blood, decreased
CD4+ count
Clinical Endpoint specific benefit that can live longer, fewer bleeds,
e felt or experienced by cure, remission, less
subjects leg pain, fewer lung
infections, improved
eathing
IRB: ET H I C S & HU M A N RE S E A RC H JA N UA RY-FE B RUA RY 2005
3
investigator (NK) serving as a mem-
er of every team. Each team recon-
ciled disagreements, and Kappa
scores were calculated for each item.
Kappa scores measure the amount of
agreement beyond what would be
observed by chance (measured as 0);
perfect agreement is measured as 1.
All scores for data presented here
were in the moderate range (.41 to
.60).13
Descriptions of Potential
Benefit. We looked for descriptions
of direct, inclusion, and societal ben-
efits in the five major sections of the
consent form: Background/Purpose,
Procedures, Risks, Benefits, and
Alternatives. We examined each
description of potential direct benefit
according to the following dimen-
sions: nature, magnitude, duration,
and likelihood. We divided descrip-
tions of the nature of potential direct
enefits into contentless (no descrip-
tion, e.g., “benefit”), su
ogate end-
points (measurements substituted fo
experiences, e.g., “tumor response”),
and clinical endpoints (specific and
experienceable, e.g., “live longer”)14
(Table 1). We divided statements
about the likelihood of potential
direct benefits into likely, unlikely,
and indeterminate (probability state-
ments that could not be further spec-
ified, e.g., “may or may not”; “you
might”; “cannot be predicted”). We
also recorded when no mention was
made of benefit in a section, when
there was no such section, and when
the section said, “You will not bene-
fit.”
We coded mentions of su
ogate
and clinical endpoints only when
coders agreed that the endpoint was
described or offered as a benefit. Fo
example, “your tumors may shrink”
and “we will check to see if you
tumors shrink” were coded as su
o-
gate endpoints offered as potential
direct benefits, whereas “any
changes in tumor size during the
study will be recorded” was not.
Statements like “may help cause
emission of your disease” were
coded as clinical endpoints offered as
direct benefits, whereas statements
like “we hope that this research is
the first step toward a future cure
for this disease” were not.
Language Use. We also col-
lected information about the use of
esearch and treatment terminology
in the consent form, such as whethe
treatment terms were used to
describe the entire study, and what
terms were used to describe subjects
in conspicuous places, such as the
study title at the beginning of the
consent form or the signature line at
the end. In addition, we examined
language use in more detail in 20%
of the 321 consent forms (N=68).
Consent forms were ordered by date
and a systematic sample of every
fifth consent form was drawn. In
this sample, we counted the terms
that, from context, clearly refe
ed to
subjects, investigators, and experi-
mental gene