University of Utah

1
University of Utah IRB Symposium 2008 The use
of placebos in clinical trials conceptions,
misconceptions, and methodologic
rationales Howard Mann, M.D. Department of
Radiology Program Associate, Division of Medical
Ethics and Humanities
2
(No Transcript)
3
Discussion Section
4
Issues for discussion 1. What is the placebo
effect ? 2. The difference between the
measured outcomes seen in the group taking the
placebo and the placebo effect 3. The primary
rationale for using a placebo in clinical trials
4. How should one assess the need for -- and
the proposed type of placebo -- in a particular
clinical trial ?
5

• The placebo effect
• Latin - I shall please
• It is a psychologically- mediated expectancy
  effect, or conditioning effect, on subjectively
  experienced symptoms - e.g., pain mood
• The effect also derives from aspects of the
  relationship between patient and practitioner a
  meaning response to the therapeutic milieu and
  ritual.

6
Reference. Kaptchuk et al. BMJ 2008
336999-1003 (3 May)
7
(No Transcript)
8

• Other phenomena that may be seen in the group
  taking placebo that result in diminution of
  symptoms
• spontaneous fluctuation in the severity of
  disease and symptoms, including symptom
  resolution
• diminution in symptom severity due to the
  phenomenon of regression to the mean
• NB these occur in both groups in a trial

9
(No Transcript)
10
Peter Kramer, M.D. Slate January 22, 2008
11

• Regression to the mean
• The tendency of extreme values that are
  measured to be lower when measured again over
  time
• It is an invariable naturalistic phenomenon

12

• A requirement for the ethical conduct of clinical
  research
• A trial should produce results that are
  internally valid and credible.
• Investigators should avoid Systematic Bias --
  factor(s) that produce a result that deviates
  from the truth.
• Such factors include reporting, assessment,
  recruitment, and retention biases (behaviors).

13
The primary role of a placebo in a clinical trial
is to enable blinding -- particularly important
when the measured endpoints are subjective -- to
promote internal (scientific) validity
14
In any particular trial, determine who needs to
be blinded and how it is best accomplished.
15

• Types of placebo used for blinding
• Sugar pill or a molecule with a taste
• Injection or infusion of saline
• Double-dummy e.g., oral and intravenous
  administration
• A sham intervention e.g., sham acupuncture
• A sham surgical procedure e.g., anesthesia and
  skin incision only

16

• Assessing special situations
• Injected placebo in newborns
• Use sham injection (simulated injection behind a
  curtain) to blind caregivers
• Invasive sham procedures
• Judge whether the outcome measure is susceptible
  to a psychologically-mediated expectancy effect
  that is durable through the period of
  observation.
• A non-blinded pragmaticstudy may be preferable,
  followed by a proposal for a explanatory
  blinded study if 1) the outcome of the
  intervention is positive and 2) the notion that a
  placebo effect is substantially responsible for
  the positive effect is biologically plausible.

17
Outcome measure ACR 20 definition of response
20 reduction in ACR20 score
18
Question how was blinding ensured ?
19
Additional resources The placebo effect an
informative and entertaining BBC 4 audio
production http//badscience.net/files/Placebo
BBC Radio 4 Ben Goldacre.mp3 Regression to the
mean a statistical explication http//www-users.
york.ac.uk/mb55/talks/regmean.htm Placebos that
harm sham surgery controls in clinical trials.
London and Kadane www.hss.cmu.edu/philosophy/londo
n/London-ShamSurgery.pdf