Randomization

• Allocation of treatments to participants is

carried out using a chance mechanism so
that neither the patient nor the physician
know in advance which therapy will be
assigned

• Simplest Case: each patient has the same

chance of receiving any of the treatments
under study
 Simple Randomization

 Think of tossing a coin each time a subject is

eligible to be randomized
 HEADS: Treatment A
 TAILS: Treatment B
 Approximately ½ will be assigned to
treatments A and B
 Randomization usually done using a
randomization schedule or a computerized
random number generator
 Problem with Simple Randomization:

• May result in substantial imbalance in either

– an important baseline factor and/or
– the number of subjects assigned to each
group

• Solution: Use blocking and/or stratified

randomization
 Blocking Example:

• If we have two treatment groups (A and B)

equal allocation, and a block size of 4,
random assignments would be chosen from
the blocks
1) AABB 4) BABA
2) ABAB 5) BAAB
3) ABBA 6) BABA
• Blocking ensures balance after every 4th
assignment
 Stratification Example
• To ensure balance on an important baseline
factor, create strata and set up separate
randomization schedules within each stratum
• Example: if we want prevent an imbalance
on age in an osteoporosis study, first create
the strata “< 75 years” and “ 75 years”
then randomize within each stratum
separately
• Blocking should be also be used within each
stratum
 Alternatives to Randomization

• Randomization is not always possible due to

ethical or practical considerations
• Some alternatives:
– Historical controls
– Non-randomized concurrent controls
– Different treatment per physician
– Systematic alternation of treatments
• Sources of bias for these alternatives need to
be considered
 Blinding
• Masking the identity of the assigned
interventions
• Main goal: avoid potential bias caused by
conscious or subconscious factors
• Single blind: patient is blinded
• Double blind: patient and assessing
investigator are blinded
• Triple blind: committee monitoring
response variables (e.g.
statistician) is also blinded
 How to Blind

• To “blind” patients, can use a placebo

Examples
– pill of same size, color, shape as treatment
– sham operation (anesthesia and incision)
for angina relief
– sham device such as sham acupuncture
 Why Should Patients be Blinded?

• Patients who know they are receiving a new

or experimental intervention may report more
(or less) side effects
 Patients not on new or experimental
treatment may be more (or less) likely to drop
out of the study
 Patient may have preconceived notions about
the benefits of therapy
 Patients try to get well/please physicians
• Placebo effect – response to medical
intervention which results from the intervention
itself, not from the specific mechanism of action
of the intervention
Example: Fisher R.W. JAMA 1968; 203: 418-419
– 46 patients with chronic severe itching randomly
given one of four treatments
– High itching score = more itching
Treatment Itching Score
cyproheptadine HCI 27.6
trimeprazine tartrate 34.6
placebo 30.4
nothing 49.6
 Why Should Investigators be Blinded?
 Treating physicians and outcome assessing
investigators are often the same people
  Possibility of unconscious bias in assessing
outcome is difficult to rule out

 Decisions about concomitant/compensatory

treatment are often made by someone who
knows the treatment assignment
  “Compensatory” treatment may be given
more often to patients on the protocol arm
perceived to be less effective
 Can Blinding Always be Done?

• In some studies it may be impossible (or

unethical) to blind
– a treatment may have characteristic side
effects
– it may be difficult to blind the physician in a
surgery or device study
• Sources of bias in an un-blinded study must
be considered
 General Study Designs
• Many clinical trial study designs fall into the
categories of parallel group, dose-ranging,
cross-over and factorial designs

• There are many other possible designs and

variations on these designs

• We will consider the general cases

 General Study Designs

• Parallel group designs

R
A B
N
D C

control
 General Study Designs

• Dose-Ranging Studies

high dose

R
A medium dose
N
D low dose

control
 General Study Designs

• Cross-Over Designs

WASH-OUT

A B
R
A
N
D
B A
 General Study Designs

• Factorial Designs

A+ B

R
A A + control
N
D B + control

control + control
 Cross-Over Designs

• Subjects are randomized to sequences of

treatments (A then B or B then A)
• Uses the patient as his/her own control
• Often a “wash-out” period (time between
treatment periods) is used to avoid a “carry
over” effect (the effect of treatment in the first
period affecting outcomes in the second
period)
• Can have a cross-over design with more than
2 periods
 Cross-Over Designs
• Advantage: treatment comparison is only
subject to within-subject variability not
between-subject variability
  reduced sample sizes
• Disadvantages:
– strict assumption about carry-over effects
– inappropriate for certain acute diseases
(where a condition may be cured during the
first period)
– drop outs before second period
 Cross-Over Designs

• Appropriate for conditions that are expected

to return to baseline levels at the beginning of
the second period
Examples:
– Treatment of chronic pain
– Comparison of hearing aids for hearing
loss
– Mouth wash treatment for gingivitis
 Factorial Designs

 Attempts to evaluate two interventions

compared to a control in a single experiment
(simplest case)
 An important concept for these designs is
interaction (sometimes called effect
modification)
 Interaction: The effect of treatment A differs
depending upon the presence or absence of
intervention B and vice-versa.
 Factorial Designs
• Advantages:
– If no interaction, can perform two
experiments with less patients than
performing two separate experiments
– Can examine interactions if this is of interest
• Disadvantages:
– Added complexity
– potential for adverse effects due to “poly-
pharmacy”
 Factorial Designs

• Example: Physician’s Health Study

• Physicians randomized to:
aspirin (to prevent cardiovascular disease)
beta-carotene (to prevent cancer)
aspirin and beta-carotene
neither (placebo)

Stampfer, Buring, Willett, Rosner, Eberlein and Hennekens

(1985) The 2x2 factorial design: it’s application to a randomized
trial of aspirin and carotene in U.S. physicians. Stat. in Med.
9:111-116.