This document provides an overview of randomized controlled trials (RCTs). It defines RCTs as studies that compare two interventions by randomly assigning participants into groups. The key aspects covered include the importance of randomization for minimizing bias, common types of bias in RCTs, techniques for randomization, and ethical considerations. RCTs are considered the gold standard for inferring causality between an intervention and outcomes.

1. THE BASICS
Randomized controlled trials
Dr. Mamta Rath Datta,
Head Consultant,
Department of ObGy,
Tata Main Hospital, Jamshedpur

2. Key points to be covered
1. Definition of RCT
2. Types of RCT
3. How to do randomisation
4. Bias in studies
5. Blinding in RCTs
6. Analysis and reporting of results

3. Definition
RCT is a study in which a group of investigators study
two interventions in a series of individuals who
receive them in a random order

4. Key terms
Intervention to be tested is
called the experimental
group
The other intervention is
regarded as a standard of
comparison or control, and
the group of participants
who receive it is called the
control group
 can be conventional practice,
a placebo, or no intervention
at all

6. Schema of a simple trial
Eligible patients
Rx group 1
Rx group 2
Randomize

8. Why carry out RCT?
RCT are prospective longitudinal studies
 Allowing causal association between intervention
and outcomes
Other study design cannot infer causality
•

9. Why Randomize?
Compare groups at the end
of the trial
Difference is because of the
Rx
For this we need
comparable groups
Purpose of randomization
is to make the treatment
groups comparable
Ensures that only
difference in groups is due
to trial treatments

11. RCT
‘the most powerful tool in modern clinical research “
Prospective
Controlled
unbiased

12. Minimizing bias in RCT
1) Allocation bias
2) Performance bias
3) Assessment bias
4) Attrition bias
5) Allocation concealment

13. Allocation bias
Occurs when the measured treatment effect differs
from the true treatment effect because of how
participants were selected into the
intervention or control group

14. Performance bias
Occurs when participants’response to Rx is affected
by the knowledge of the group to which they are
assigned – intervention/control
Occurs when health professionals administer
treatment differently between treatment
arms

15. Assessment bias
Health professionals assessing the outcome may record
outcome measures biased by the knowledge of the
group assignments
Overestimation or underestimation of the effects
of an intervention is known as assessment bias
There might be a systematic difference in measuring
the outcomes between the two groups because of
the method of recording used
 – E.g. control group is assigned to one practitioner and the intervention
group to another, or groups are assessed at different times of the day

16. Assessment bias (cont.)
How to minimise assessment bias
 Use a standardized method of evaluation across both
groups o minimize the assessment bias?
 Avoid using subjective measures to assess the
effectiveness of a treatment which are more prone to
bias

17. Attrition bias
Also called as loss-to-follow-up bias
 Occurs when patients drop out of the study from
their respective study group
If halfway through a study the treatment has been
successful, participants may drop out and information
about the success of the treatment is then lost
Participants in the control group might be unhappy with
their lack of progress and may drop out of the study in
order to seek alternative help

18. Allocation concealment
 Bias will be minimized where the allocation schedule is
concealed
 Blinding (or masking)
 helps prevent systematic differences between comparison groups in
prognosis or responsiveness to treatments (allocation bias).
 Blinding of both participants and practitioners
 prevents performance and assessment bias by ensuring everybody
(participants, treatment admin, those measuring outcomes) do not
know which treatment was given.
 It is recommended RCT participants are blind to the
treatment they receive.
 – Control group receives placebo

19. Important factors in RCT
1) Sample size
2) Stratification
3) Trial design
4) Between group contamination
5) Ethical issues

20. 1.Sample size
Sample size dependent upon
 the power of the test
 what size of intervention impact is considered meaningful
 type of hypothesis the RCT is testing
 The smaller magnitude of difference between groups
that is to be detected and the greater the variability
in outcomes, the larger the sample size that will be
required

21. 2.Stratification
Stratification is a way of ensuring the treatment
groups are balanced on characteristics that are likely
to alter the relationship between treatment and
outcome.

22. 3.Trial design
By study design
By outcome of interest (efficacy vs.
effectiveness)
By hypothesis (superiority vs. noninferiority
vs. equivalence)

23. Classification by study design
 Parallel-group
 each participant is randomly
assigned to a group
 all the participants in the
group receive (or do not
receive) an intervention.
 Crossover
 over time, each participant
receives (or does not receive)
an intervention in a random
sequence
 Cluster
 pre-existing groups of participants
(e.g., villages, schools) are randomly
selected to receive (or not receive) an
intervention.
 Factorial
 each participant is randomly assigned
to a group that receives a particular
combination of interventions or non-
interventions
 e.g., group 1 receives vitamin X and
vitamin Y, group 2 receives vitamin X
and placebo Y, group 3 receives
placebo X and vitamin Y, and group 4
receives placebo X and placebo Y

25. By outcome of interest (efficacy vs.
effectiveness)
 Explanatory RCTs :
 test efficacy in a research
setting with highly selected
participants and under highly
controlled conditions
Pragmatic RCTs (pRCTs):
 test effectiveness in everyday
practice with relatively
unselected participants and
under flexible conditions
 in this way, pragmatic RCTs can
"inform decisions about practice”

26. By hypothesis (superiority vs.
noninferiority vs. equivalence)
differ in methodology and reporting
Superiority trials:
 in which one intervention is hypothesized to be superior to
another in a statistically significant way
Noninferiority trials :
 determine whether a new treatment is no worse than a
reference treatment
Equivalence trials :
 in which the hypothesis is that two interventions are
indistinguishable from each other

27. 4.Between group contamination
Information will be passed
between the two arms of trial
and thus alter results.
 Use cluster sampling so natural
cluster such as geographic areas
are randomized rather than
individuals

28. 5.Ethical issues
RCT is not always possible
because of ethical issues
 If one group of patients receives
treatment thought to be effective,
while another group does not, the
ethics of a trial may be brought
into question
Some trials cannot be carried out
because they may actively
encourage unhealthy practices
e.g. smoking.
 People cannot be randomized into
smoking and non-smoking group.

29. What is wrong with non-randomized studies?
Two main types of study, those with and those
without concurrent control groups

30. Non-randomized studies II
Without concurrent controls
 Uncontrolled
 cannot really make much of such studies if there is any
variation in outcomes.
 Historical controls
 type of patient may change, due to eligibility criteria
 environment changes, due to trial
 data quality often quite different between groups

31. Non-randomized studies III
Non-randomized concurrent controls
Alternation
Odd/Even hospital no. or date of birth
First letter of surname
Difficult to argue that one group is different from
another but allocation is predictable, so bias can
arise from selection of patients
so randomization must be unpredictable

32. Randomization techniques
Randomization is the process of assigning clinical
trial participants to treatment groups
all participants have the same chance of being
assigned to each of the study groups
the purpose is to keep both groups as similar to each
as possible at the start of the trial.
Successful randomization requires that group
assignment cannot be predicted in advance.

33. Simple Randomization
1. Coin Tossing for each trial participant
2. Sequence of Random Numbers from statistical textbooks
3. Computer generated sequence

34. Examples
flipping a coin. For example, with two treatment
groups (i.e., heads - control, tails – treatment)
Shuffled deck of cards (e.g., even - control, odd -
treatment)
Throwing a dice (e.g., below and equal to 3 - control,
over 3 - treatment).
 The computer generated sequence:
4,8,3,2,7,2,6,6,3,4,2,1,6,2,0,…….

35. Is coin tossing OK?
OK for big trials
For small trials, such ‘simple randomization’
can lead to imbalance in group sizes

36. Block randomization
Designed to randomize subjects into groups that result in
equal sample sizes.
 The block size is determined by the researcher and should
be a multiple of the number of groups (i.e., with two
treatment groups, block size of either 4, 6).
Example: Two treatments of A, B and Block size of 2 x 2= 4
 Possible treatment allocations within each block are
 (1) AABB, (2) BBAA, (3) ABAB, (4) BABA, (5) ABBA, (6) BAAB

37. Stratified Randomization
Trial may not be valid if it is not well balanced across
prognostic factors.
• SR means block within block For example, Age
Group: < 40, 41-60, >60; Sex: M, F
• For 6 patients in a block, Total number of strata = 3
x 2 = 6 .
• It produce comparable groups with regard to certain
characteristics (e.g., gender, age, race, disease
severity), thus produces valid statistical tests

38. • The block size should be relative small to maintain
balance in small strata
• Increased number of stratification variables or
increased number of levels within strata leads to
fewer patients per stratum
• Subjects should have baseline measurements taken
before randomization
• Large clinical trials don’t use stratification

39. Unequal Randomization
 Most randomized trials allocate equal numbers of
patients to experimental and control groups.
 most statistically efficient randomization ratio as it maximizes
statistical power for a given total sample size.
 may not be the most economically efficient or
ethically/practically feasible
Substantial cost savings with smaller randomization
ratio such as a ratio of 2:1, with only a modest loss in
statistical power
Ratio of 3:1 will lose considerable statistical power
 Ratio >3:1 is not very useful - leads to much larger loss of
sample size power

40. Inappropriate randomization methods/Pseudo-
randomisation
• Assigning patients alternately
to treatment group is not
random assignment
• Assigning the first half of the
population to one group is
not random assignment
• Assignments by methods
based on patient
characteristics such as date of
birth, order of entry into the
clinic or day of clinic
attendance, are not reliably
random

41. Types of RCTs/Hierarchy of Blinding
• Open label: no blinding
• Single blind: patient blinded to treatment
• Double blind: Patient and Physician (and data
collectors) blinded to treatment
• Complete blind: Everyone involved in the study
blinded to treatment

42. Blinding
protects a trial against bias is by keeping the people
involved in the trial unaware of the identity of the
interventions for as long as possible

43. Open Label Studies
 Useful for
 • Pilot studies
 • Dose ranging studies
 Open label studies are not recommended for comparative
trials

44. Single Blind Studies
usually the patient is blinded
justification is usually that double-
blind is "impractical" because of
need to adjust medication,
medication affecting laboratory
values, potential side effects,
critical condition of the patient etc.
can also be used when it is
unacceptable ethically to give a
placebo treatment to a patient, and
in such a case, the assessor (not the
patient) should be the one blinded
to the treatment

45. Double Blind Studies
both the subjects and the
investigators are kept from
knowing who is assigned to
which treatment
serve as a standard by which all
studies are judged, since it
minimizes both potential
patient biases and potential
assessor biases

46. Double Blinding:Techniques
Coded treatment groups
 Placebo for each possible treatment.
 Tablets identical in physical appearance.
 Tablets with similar taste and smell
 IV infusions would normally be the same carrier as
used for active medications.
 Other treatments "shammed" as far as possible: eg.
Minimal power ultrasound therapy when testing
effect of physical therapy in back pain.

47. Disadvantages: Double Blinding is not always
feasible??
When intervention is surgery- It is unlikely that
sham surgery would be considered ethical in a study
It would be hard to blind a patient to the therapy
given in an exercise study
it might not be possible to blind a patient while
comparing utility of different invasive procedures

48. Double Blind Studies: Difficulties
Side effects:
 Side effects (observable by
patient) are much harder to
blind
 in general, there are
significant ethical problems
using placebos to induce side
effects in patients
 a way to avoid it is that the
side effects of all the potential
therapies be combined into a
single list, so that knowledge
of side effects would not
indicate therapy (at least to
patient).
Efficacy:
 A truly effective treatment
can be recognized by its
efficacy in patients
 Some new treatments ARE
VERY EFFECTIVE and when
this happens, it is becomes
clear which treatment a
patient is receiving, at least
for the health care providers
involved in the trial

49. Complete Blinding
 Patient and the investigator, all members of the
clinical project team of the sponsor including CRA,
statistician, programmer, and data coordinator are
blinded.
May require two groups for data processing, one
group to encode the data/analysis and one group to
perform the analysis
 Normally only available in major drug company
studies, and not routinely used

50. Complete Blinding:Techniques
Analysis uses coded treatment groups
 Analysis uses coded side effects (e.g., side effects
coded using non-standard scheme, with only
numeric codes available at time of analysis)
 Analysis uses coded laboratory tests (e.g., name of
test coded numerically at time of analysis, using non-
standard code)

51. Coding of drugs
Assigning a random number.
 As many as different code.
Participants: unique code.
If only one code is used: disclosure for 1 will disclose
for all.
Many side effects in many people: decode
Efficient coding: should not confuse the prescriber and
stocking of drugs

52. Unblinding of study
The carton must contain slip of drug inside it
 Should not be disclosed to patients while storage
 Official unblinding may be necessary during
emergency

53. Blinding is difficult
Having placebo in the same shape , formula and
taste is very costly, and time consuming
The drug side effects e.g. local reaction at the site of
injection would partially unblind
Impossible if surgical and medical treatments are
compared
The need for urgent unblinding code in case of
serious side effects

54. Follow up
Adherence to the study protocol
Patients compliance with treatment and follow up
Sufficiently long and complete

55. Disadvantages of RCT
expensive: time and money
 volunteer bias
 ethically problematic at
times
high dropout when the
intervention has
undesirable side-effects or
there is little incentive to
stay in the control arm
ethical consideration may
mean that a research
question cannot be
investigated using RCT
design
 for a descriptive overview
it may be cheaper and
easier to use an
observational design.
 Prior knowledge is
required for sample size
calculation
 the level of improvement
that is clinically meaningful
 expected variation of
improvement in the sample

56. Analysis of data
 RCT are experiments set up to test hypotheses.
 Null hypotheses – the intervention will have no impact
on the outcome measure
 Outcome will be similar in both the test and control groups
 Alternative hypotheses – intervention will have
meaningful effect and statistically significant
 Statistical method (e.g.);
 Pre and post intervention differences = paired t-test
 Mean differences pre and post between two group = independent t-
test
 Mean differences pre and post between more than two groups =
ANOVA

57. Analysis of data
For dichotomous (binary) outcome data:
 logistic regression (e.g., to predict sustained virological
response after receipt of peginterferon alfa-2a for hepatitis C)
and other methods can be used.
For continuous outcome data:
 analysis of covariance (e.g., for changes in blood lipid levels after
receipt of atorvastatin after acute coronary syndrome[57]
) tests
the effects of predictor variables.
For time-to-event outcome data :
 censored, survival analysis (e.g., Kaplan–Meier
estimators and Cox proportional hazards models for time
to coronary heart disease after receipt of hormone replacement
therapy in menopause[58]
) is appropriate.

58. Interim Analysis
Done in large multicenter RCTs
To explore the results after recruiting of half of the
participants
If marked difference is recognized , then trial should
be stopped
Examples: WHI trial
 Breech Trial
 AIDS trial

59. Reporting of results
The CONSORT 2010 Statement is "an evidence-based,
minimum set of recommendations for reporting RCTs
 CONSORT 2010 checklist contains 25 items focusing on
"individually randomised, two group, parallel trials"
which are the most common type of RCT
For other RCT study designs, "CONSORT extensions"
have been published, some examples are:
 Consort 2010 Statement: Extension to Cluster Randomised Trials
 Consort 2010 Statement: Non-Pharmacologic Treatment
Interventions

60. Conflict of interest dangers
Some RCTs are fully or partly funded by the health care
industry
A systematic review published in 2003 found a correlation of
industry sponsorship and positive study outcome
A 2004 study of 1999-2001 RCTs determined that industry-
funded RCTs "are more likely to be associated with
statistically significant pro-industry findings."
These results have been mirrored in trials in surgery, where
although industry funding did not affect the rate of trial
discontinuation it was however associated with a lower odds
of publication for completed trials
One possible reason for the pro-industry results in industry-
funded published RCTs is publication bias.
Declaration of conflict of interest
essential

61. Take home message
The gold standard for clinical /interventional/drug
research
Basis for Meta-analysis for policy making and EBM
Familiarity with sample size, blinding and statistical
analytical tools is essential for interpretation of
outcome
Changes the way we practice medicine
Trial registration may be important/essential for
future studies