Ethical Issues in Intervention Studies (and any kind of research involving human subjects)
Download
Report
Transcript Ethical Issues in Intervention Studies (and any kind of research involving human subjects)
Ethical Issues in
Intervention Studies
(and any kind of research
involving human subjects)
The Nuremberg Code
(1946-47 Trials)
“Experimentation” by Nazi physicians was
exposed in 1946-47 trials. Resulting code
adopted 10 principles:
Study participants must give voluntary consent
There must be no reasonable alternative to
conducting the experiment
Anticipated results must have a basis in
biological knowledge and animal
experimentation
Procedures should avoid unnecessary suffering
and injury
The Nuremberg Code
(1946-47 Trials)
No expectation for death or disability as a
result of the trial
Degree of patient risk is consistent with the
humanitarian importance of the study
Subjects are protected against even a remote
possibility of death or injury
Study must be conducted by qualified
scientists
The subject can stop participation at will
Investigator must terminate the experiment if
injury seems likely
Declaration of Helsinki - 1964
(Revised 1975)
Issued by the World Medical Association
who adopted a formal code of ethics for
physicians engaged in clinical research
“It is the mission of the physician to safeguard
the health of the people.”
“Medical progress is based on research which
ultimately must rest in part on experimentation
involving human subjects.”
Other International Guidelines
1982 - World Health Organization (WHO)
and Council for International Organizations
of Medical Sciences (CIOMS)
Document: -- “Proposed International
Guidelines for Biomedical Research
Involving Human Subjects”
Document: -- International Guidelines for
Ethical Review of Epidemiologic Studies
High Profile Mistakes in U.S.
1936 -- Tuskegee Alabama: Study of the
effects of untreated syphilis which continued
long after an effective treatment was known.
1963 -- Brooklyn - Jewish Chronic Diseases
Hospital: Cancer cells were injected into
debilitated elderly patients to see if they
would immunologically reject the cells.
1972 - Willowbrook State Hospital in New
York: Retarded children were deliberately
infected with viral hepatitis to study its
natural history.
Modern Perspective
Institutional Review Board (IRB)
IRBs are required for all research funded
in whole or in part by the federal
government.
IRB approval is required for all drugs or
products regulated by FDA (regardless of
the funding source).
Modern Perspective
Institutional Review Board (IRB)
Composition of IRBs:
---
At least 5 members with relevant expertise
---
At least one scientist and one nonscientist
---
At least one person not affiliated with the
institution
---
Diverse racial, gender, and cultural
backgrounds
Modern Perspective
Institutional Review Board (IRB)
Research requirements judged by IRBs:
– Risks to study participants are
minimized
– Risks are reasonable in relation to
anticipated benefits
– Selection of study participants is
equitable
Modern Perspective
Institutional Review Board (IRB)
Research requirements judged by IRBs
(cont.):
– Informed consent is obtained and
appropriately documented for each
participant
– Adequate provisions for monitoring data
collection to ensure safety of the study
participants
– Participant privacy and confidentiality is
protected
Informed Patient Consent
“…the doctor should obtain the
subject’s freely-given informed
consent, preferably in writing…”
“…if the doctor considers it essential
not to obtain informed consent, the
specific reasons for this proposal
should be stated in the experimental
protocol…”
Design & Conduct of
Clinical Trials
Axiom: The randomized clinical trial
represents “The Gold Standard” for
epidemiological research.
However, other study designs can be
equally valid, and a RCT does not
guarantee a valid study.
Issues in Clinical Trials:
Cost and Feasibility
Clinical trials are expensive and unwieldy to
conduct (require collaboration between patients,
physicians, nurses, data managers,
methodologists, etc.).
Secular trends in health behavior may make
some interventions difficult to test (e.g.
antioxidants).
Large simple trials with streamlined protocols
are being advocated by many today.
Ethical guidelines and concerns continue to
evolve.
Issues in Clinical Trials:
Statistical Power
Definition of Power: The probability of
rejecting the null hypothesis of no
association when a true association
exists.
Clinical trials must have sufficient power
to reliably detect small-to-moderate yet
clinically important differences between
treatment groups that are likely to occur.
Issues in Clinical Trials:
Statistical Power
Power
of a trial may be less than
expected because:
– The number of outcomes is less
than expected
– Observation bias
– Loss to follow-up
Issues in Clinical Trials:
Selection of a Study Population
Define the reference population -- to whom
will the results of the study be applicable.
Experimental population -- the actual
group in which the trial is conducted.
Ideally, should match the reference
population.
Sufficient outcomes must be expected.
Willing participants must be screened for
eligibility according to pre-defined criteria.
Willing and eligible comprise the actual
study population.
Issues in Clinical Trials:
Selection of a Study Population
The actual study population is often a
small fraction of the experimental
population.
Be careful, the eligible and willing
participants may experience less
outcomes than expected (“volunteer
effect”).
Try to obtain baseline data for eligible but
unwilling individuals (e.g. to aid in
determining generalizability of results).
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Simple Randomization:
– Participants are assigned at random to
one of the treatment regimens (parallel
groups).
Randomization Assignment Means:
– Each individual has the same chance of
receiving each of the possible treatments.
– The probability that a subject receives a
treatment is independent of the probability
of other individuals receiving the same
treatment.
Issues in Clinical Trials:
Advantages of Random Assignment:
– Potential for bias in allocation to study
groups is removed (since no one involved in
deciding patient eligibility or procedure
allocation will know assigned treatment
group).
– Study groups will be comparable with respect
to all variables except for the intervention
being studied.
– Randomization can control for known and
unknown confounding!
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
2 x 2 Factorial Design:
Treatment B
No
Yes
Treatment A
No
Yes
A
B
C
D
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Strengths of the 2 x 2 Factorial Design:
– In some circumstances, two treatments
can be tested using the same number of
subjects and minimal additional cost.
– The interaction (effect modification) of
the two treatments can be evaluated.
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Limitations of the 2 x 2 Factorial Design:
– The two treatments must be amenable
to being administered in combination
(e.g. collective side effects can not
make the regimen hazardous)
– Not appropriate for treatments that are
believed to have the same mechanism
of action
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Randomization in Blocks:
Block #
1
1
1
1
2
2
2
2
Males
Young
Old
A
B
B
A
A
A
B
B
B
B
A
A
B
A
B
A
Females
Young
Old
B
A
B
A
A
B
A
B
A
B
A
B
B
B
A
A
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Advantages of Randomization in
Blocks:
– Improves the balance in the number of
treatment assignments in each group.
– Makes the sequence of assignments
more random.
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Crossover Design:
– Each treatment is administered to each patient
at different times in the study
Permits within-patient comparisons of
treatment effects
Each person serves as their own “control”
May improve recruitment since subjects are
offered both treatments
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Crossover Design – Disadvantages:
– Be careful -- “carry-over” effects may be
present
– Be careful -- There may be a treatment by
period interaction
– The analysis is more complex than in a
parallel groups design
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Equivalence Trials:
– Research question is whether an existing/new
therapy is as effective as a standard therapy.
– Traditional hypothesis testing is reversed:
The “null hypothesis” is that the treatments
are different.
Statistical testing is performed to test
whether the treatments are sufficiently
similar.
Issues in Clinical Trials:
Allocation of Study Regimens-Designs
Equivalence Trials:
Frequently used to test “bioequivalence”
or compare efficacy of therapies with
different costs and side effects.
– Be careful, misclassification tends to
cause the treatment groups to be more
similar – this results in bias away from
the null.
Issues in Clinical Trials:
Characteristics of Control Groups
–
Spontaneous remission
– Hawthorne Effect -- Subjects may change
behavior because they are being observed.
– Regression to the mean -- Extreme values on
initial examination show improvement during
subsequent exams, by chance alone.
– Placebo Effect -- Tendency of individuals to
report a favorable response to any therapy
regardless of the physiologic efficacy of what
they receive.
What is a placebo?
Physiologically inert substance that is
identical in physical characteristics to
the active drug.
Placebo effect: Persons taking a drug
or undergoing a medical procedure tend
to ascribe any symptom to their
treatment.
Why use a placebo?
–Ensures that all aspects of
the program offered to
participants are identical,
except for the actual
experimental treatment
Clinical Trials - Maintenance and
Assessment of Compliance
Maintaining high compliance with the treatment
protocol is a major issue.
Non-compliance is generally related to the
length of time required for adherence to the
study protocol and to the complexity of the
protocol.
Non-compliance decreases statistical power -by making the two treatment groups more
similar.
Non-compliance is a threat to the validity of the
study results (e.g. when a “null” result is
observed).
Protocol Deviations
Methods to assess non-compliance:
Give more pills than needed and count
the remaining pills during follow-up
Measure drug concentration in serum or
urine
Measure drug-induced biochemical
changes
Clinical Trials - Outcome Assessment
Goal is to obtain complete and unbiased
information.
As follow-up time increases, maintaining
complete
ascertainment of outcomes becomes more
difficult.
When the proportion of outcomes not
ascertained is large or differs among the study
groups, a biased result is likely.
The potential for observation bias is related to
non-use of masking (blinding) and subjectivity of
the study outcomes.
Blinding (Masking)
Blinding (Masking) in Clinical Trials - to
reduce observation bias
Single Blind Trials
– Patients do not know their treatment
assignment.
Double Blind Trials
– Neither patient nor investigator responsible
for assessing the outcome knows treatment
assignment (especially useful for
“subjective” outcomes).
Triple Blind Trial
– Patients, investigators, and evaluators (e.g.
Data Safety and Monitoring Committee) do
not know treatment assignment.
Issues in Analysis
Always compare baseline characteristics of the
study groups to assess the effectiveness of
randomization in balancing the treatment groups.
Which subjects should be included in the
analysis?
– Remove non-compliers? (general rule is NO).
– “Intention to Treat” (e.g. once randomized
always randomized).
– Keep in mind that the research question posed
is whether offering of the treatment program is
of benefit (not whether or not compliance is).
Issues in Analysis
Subgroup Analyses:
– A priori versus post-hoc analyses
– Problem of multiple comparisons and the
role of chance
– Overall balance achieved by
randomization is not assured when
subgroup analyses are performed
Early Trial Termination
Stopping Rules:
– Rules to decide whether a trial should be
terminated or modified before originally
scheduled.
– Used to assure the welfare of study
participants is protected.
– Interim results should be monitored by an
outside (independent) entity.
Early Trial Termination
Reasons for Early Termination:
– Data indicate a clear and extreme benefit on
the primary study outcome (e.g. would be
unethical to allow other continuation with the
less effective treatment regimen(s)).
– Treatments are found to be convincingly not
different by impartial knowledgeable experts.
– A treatment is observed to be clearly harmful.
Early Trial Termination
Reasons for Early Termination (cont.):
– The data are of poor quality.
– Subject accrual is too slow to complete the
study in a timely fashion.
– Definitive information becomes available
from outside the study, making the trial
unnecessary or unethical.
Early Trial Termination
Reasons for Early Termination (cont.):
– Adherence to treatment regimens is
unacceptably poor, preventing an answer to
the basic research question.
– Resources to perform the study are lost or
no longer available.
– The study integrity has been undermined by
fraud or misconduct.
Review of Recommended Reading
Diabetes Prevention Program
--- Multi-center trial of 3,234 persons at “high-risk” of
developing diabetes.
--- Randomly assigned to either (i) standard lifestyle +
placebo; (ii) standard lifestyle + metformin; (iii)
intensive lifestyle modification program.
--- Adherence to assigned treatment regimens assessed
quarterly.
--- Goals of intensive lifestyle program were reduction of
>7% of body weight and engage in moderate physical
activity for >150 minutes per week.
--- Primary outcome measure was incident diabetes.
--- Participants followed for average of 2.8 years.
Discussion Question 1
In table 1, does it appear that the process
of random assignment balanced the 3 study
groups in terms of potential
confounding variables?
Source: NEJM 2002; 346:393-403.
Discussion Question 2
In figure 1, does it appear that the
intensive lifestyle intervention program
was truly implemented and maintained
over the course of the study?
Source: NEJM 2002; 346:393-403.
Discussion Question 3
Interpret the results of figure 2.
(i) Was the effect of treatment assignment
consistent over the course of the study?
(ii) The “effect sizes” observed may be
underestimates of the true effect – Why?
Source: NEJM 2002; 346:393-403.
Discussion Question 4
Interpret the results of table 2.
(i) Is there evidence of effect measure
modification (i.e. the effect of treatment
assignment varied according to baseline
characteristics such as age or gender?)
Source: NEJM 2002; 346:393-403.
Discussion Question 5
Comment on the adverse events
presented in table 3. Are the results
for GI symptoms surprising?
Source: NEJM 2002; 346:393-403.