Lecture 5: Randomised Controlled Trials

1

Epidemiology

• Lecture 5 Randomised Controlled Trials

2
Learning outcomes

• Describe the key features of randomized
  controlled trials
• Calculate measures of association used in
  randomized controlled trials

3
Randomised Controlled Trials

• Ideal design for evaluating both the
  effectiveness and the side effects of new forms
  of intervention
• RCTs can be used to evaluating new drugs and
  other treatments of disease including tests of
  new health and medical care technology
• RCTs can also be used to assess new programs for
  screening and early detection or new ways for
  organising and delivering health services

4

• RCT is an experiment in which subjects in a
  population are randomly allocated into groups
• to receive an experimental preventative or
  therapeutic procedure or intervention
  (intervention group)
• or
• to receive a placebo, no intervention or usual
  care (control group) and the outcomes are
  compared.

5
RCTs
6
Randomised Controlled Trials

• All subjects free from outcome factor/disease at
  the commencement of the study
• Study pop defined geographically, temporally and
  by other exclusion criteria RCTs
• Subjects randomly allocated to an intervention
  group (study factor) or non-intervention (control
  placebo or established therapy)
• Followed over time to determine the outcome.

7
Randomised Controlled Trials

• Best evidence as to whether exposure results in
  specific outcome
• Used to evaluate the efficacy of drugs, clinical
  and preventive intervention programs

8
Example of Randomised Controlled Trials

• A placebo-controlled randomized trial might
• compare
• the effect of vitamin E treatment in
• schizophrenia patients (the treatment group)
• AGAINST
• the effects of a placebo on a separate
• group of schizophrenia patients (the control
• group).

9
Phases of clinical trials

• Phase 1 Clinical pharmacologic studies
• Small studies 20-80 patients look at toxic
  pharmacological effects
• Phase 2 Efficacy and safety
• 100-200 patients
• Phase 3 large scale RCTs for effectiveness and
  relative safety. Often multicentred. Followed by
  licensing of drug
• Phase 4 Post-marketing surveillance for
  monitoring new drugs

10
Issues in designing RCTs

• Selection of subjects
• Random Allocation of subjects to treatment and
  control groups
• Data collection
• Blinding (or masking)
• Sample size
• Crossover trials
• Noncompliance
• Contamination

11
Selection of subjects

• Criteria precisely defined
• Volunteers
• Representative?

12
Random Allocation of subjects

• Randomization avoids bias by
• Ensuring unpredictability of next assignment
• Reduces baseline differences in risk between
  treatment and control groups.
• should make both groups similar in terms of the
  distribution of risk factors
• larger the randomized groups, the greater the
  probability of equal baseline risks.

13
Random Allocation of subjects

• Use computer generated random lists
• Or random number tables (example in text)
• Reduces bias caused during selection

14
Data collection

• It is essential that data collected for each of
  the study groups be of the same quality
• Treatment (assigned and received)
• Outcome criteria explicitly stated and measured
  comparably in all groups
• Otherwise ? observer bias

15
Blinding (or masking)

• Knowledge of whether the participant is in a
  treatment or control group can influence
  behaviour
• Overcome through blinding the participants and/or
  observer and and/or data reviewer

16
Blinding (or masking)

• Single blinding subject (participant) not given
  any information about whether allocated to
  treatment or comparison group
• usually via a placebo (inert agent usually
  indistinguishable from the active treatment)

17
Subjects suspicion about their treatment
Gordis, 2001
18
Blinding (or masking)

• Double Blinding - neither subject or observer
  have any information about allocation of subject
  to treatment or comparison groups
• Minimises bias during assessment and care

19
Blinding (or masking)

• Triple blinding neither subject, observer or
  person analysing the data have any information
  about allocation of the subject to treatment or
  comparison groups
• Unblinded or open label studies no attempt at
  blinding

20
Blinding (or masking)
21
Sample size

• Need to calculate adequate sample size to obtain
  meaningful results
• Use sample size computer program
• Sample size tables and statistical formulas

22
Possible trial outcomes
23
Crossover trials

• There are 2 types
• Planned crossover trials
• subjects are randomised to therapy A or B and
  after being observed for a period they switch
  therapy.
• Subject become their own controls
• Useful as it allows constant characteristics
  between the treatments

24
Cross over trials
25
Planned Crossover trials

• Advantages
• Comparisons within individuals
• Subjective outcomes can be used (with blinding)
• Reduces sample size needed
• Disadvantages
• Carryover
• Washout period
• Need to consider effect of the order of
  interventions (psychological response)
• Dropouts from 2nd treatment

26
Unplanned crossover in study of bypass surgery
Randomised
Surgery
Medical
Refuse Surgery
Require surgery
Surgery
No surgery
27
Noncompliance

• Reduces the power for the detection of an effect
• Limiting the analysis to only those who comply
  may bias the results

28
Noncompliance - Example

• Trial of cancer-prevention diet vs placebo diet.
• Treatment Group
• some subjects had GI symptoms (these were
  actually precursors of the cancer)
• they did not comply with the diet
• If these non compliant people were excluded from
  the analysis
• ? Greater effect of the cancer-prevention diet

29
Intention to treat

• Performing the analysis of RCTs according to
  original randomised assignment
• Not according to whether they comply with
  treatment or not
• Optimal estimate of the true benefit of the
  intervention

30
Contamination

• that people in control group will receive part or
  all of intervention that is used for the
  intervention group
• ? reduces any differences between the two groups
• ? decreases the likelihood of identifying these
  differences.

31
Contamination

• may be caused by
• service providers or trials inadvertently
  applying trial interventions to the control
  group
• individual participants seeking additional care
  from providers outside of the trial.
• other influences on usual care which are out of
  the control of the trial organisers.

32
Contamination

• cannot usually be prevented
• it may be reduced by modifications to trial
  design (eg. using randomization by service
  providers).
• where it cannot be prevented or reduced then it
  should be measured and the information used when
  assessing the results.

33
Advantages Disadvantages

• Best evidence for causality
• ensures that individuals are allocated to the
  intervention or control groups without prejudice
  
• Minimizes/ eliminates unequal distribution of
  factors that influence clinical outcome between
  groups
• Facilitates statistical analysis.

• Expensive (time and money).
• Organisationally difficult.
• Difficult to recruit health professionals to
  participate.
• Not always generalisable.
• Sometimes ethical problems.

34
Non-randomised Controlled Trial (NRCT)

• an experiment in which subjects in a population
  are allocated into groups (using methods other
  than randomisation) to receive or not receive a
  therapeutic procedure or intervention and the
  outcomes are compared.
• all other features may be similar

35
Example
BCG vaccination for TB in children from
tuberculous families in New York, 1946
36
Advantages Disadvantages

• Poorer evidence for causality
• Can minimizes unequal distribution of factors
  that influence clinical outcome between groups
• Facilitates statistical analysis
• Less expensive than RCTs

• Individuals may be allocated to the intervention
  or control groups with prejudice
• Expensive (time and money).
• Not always generalisable.

37
Example
BCG vaccination for TB in children from
tuberculous families in New York, 1946
Incidence/Event rate of outcome (intervention
group) 3/445 0.67 Incidence/Event rate of
outcome (control group) 18/545 3.3
38
Measures of Association used in Intervention
Trials

• When comparing 2 groups of participants in a RCT,
  the control group (those who did not receive
  treatment) and the intervention group, (those who
  were exposed to the treatment), you can use the
  following measures to help you judge the size of
  the effect of the intervention
• The relative risk reduction (RRR)
• The absolute risk reduction (ARR)
• The numbers needed to treat (NNT)

39
Relative Risk Reduction (RRR)

• Is the extent to which a treatment reduces a
  risk, in comparison with patients not receiving
  the treatment of interest.
• is the commonest reported measure of dichotomous
  treatment effect
• however does not discriminate between huge
  absolute treatment effects and trivial ones.

40
Relative Risk Reduction (RRR)

• RRR
• Event rate Event rate
  in control grp in intervention grp
• Event rate in the control group

41
Absolute Risk Reduction (ARR)

• is the difference in the absolute risk (rates of
  adverse events) between study and control
  populations
• ARR
• Event rate Event rate
  in control grp in intervention grp

42
Number needed to treat (NNT)

• Is the number of patients who must be exposed to
  an intervention before the clinical outcome of
  interest occurred
• for example, the number of patients needed to
  treat to prevent one adverse outcome.

43
Examples

• NNT 1
• ARR
• Example control intervention
• mortality rate 17 12
• NNT 1 1 20
• 17-12 0.05
• We need to treat 20 people to prevent one death

44
DCCT Diabetic Neuropathy
45
Measuring the precision of the results

• Tests for Statistical Significance
• show you the precision of the results of a study
  by examining the confidence intervals or the p
  values help you to evaluate whether the study was
  statistically significant.

46
Confidence intervals

• show a range within which the true effect of the
  intervention is likely to be.
• a confidence interval that includes the value of
  no effect (e.g. RR1 or RRR0) shows that the
  intervention group is not statistically
  significantly different from the control.

47
Confidence intervals

• Where the confidence interval does not include
  the no effect value this shows that there is a
  statistically significant different between the
  intervention and control group.
• Statistical significance is usually measured
  using a 95 confidence interval, meaning that if
  the study is repeated multiple times, 95 of the
  studies will have result within that range.

48
Confidence intervals

• When a study does not show a statistical
  significance, it may still have a real effect.
• For example, small studies will often report no
  statistical significance but they may show very
  important clinical effects.
• Is the preferred way of calculating precision but
  in some studies only p values may be presented.

49
p-value

• reflects the degree of certainty about the
  existence of a true effect.
• based on the supposition that the null hypothesis
  is true i.e. that there is no true difference
  between the intervention and control groups.
• Using the p value we calculate the likelihood
  that our null hypothesis is true.

50
p-value

• If p is small, it is unlikely that the difference
  is due to chance and we reject the null
  hypothesis.
• If p is large then it is likely that the
  difference is due to chance and we do not reject
  the null hypothesis.
• Statistical significance is usually set at
  plt0.05 or plt0.01.

51
Consort (Consolidated Standards of Reporting
Trials) statement

• First published in 1996
• Available at
• http//www.consort-statement.org/revisedstatement.
  htmchecklist
• Aim - improve quality of conduct reporting of
  RCTs
• Developed by - international group of clinical
  trialists, statisticians, epidemiologists and
  biomedical editors

52

• Supported by a growing number of medical and
  health care journals, and editorial groups
• Endorsed by The Lancet, JAMA, Annals of Internal
  Medicine
• Comprises checklist flow diagram for reporting
  RCTs
• Used in the writing, reviewing, or evaluating
  reports of simple two-group parallel RCTs

53
Consort Flow Diagram
54
CONSORT Checklist

• INTRODUCTION
• Background
• METHODS
• Participants
• Interventions
• Objectives
• Outcomes
• Sample size
• Randomisation sequence generation, allocation
  concealment, implementation
• Blinding (masking) Statistical methods

55
CONSORT Checklist

• RESULTS
• Participant flow
• Recruitment
• Baseline data
• Numbers analyzed
• Outcomes and estimation
• Ancillary analyses
• Adverse events
• DISCUSSION
• Interpretation
• Generalisability
• Overall evidence

56
QUESTION 1

• The purpose of double blinding or double masked
  study is to
• Achieve comparability of treated and untreated
  subjects
• Reduces effects of sampling variation
• Avoid observer and subject bias
• Introduce observer and subject bias in the study

57
QUESTION 2

• The major purpose of random assignment in a RCT
  is to
• Facilitate double blinding
• Facilitate measurement of outcome variables
• Ensure study groups are comparable on baseline
  characteristics
• Reduce selection bias in allocation of treatment

58
QUESTION 3

• A RCT comparing the efficacy of 2 drugs showed a
  difference between the two (plt0.05). Assume in
  reality, however, the 2 groups do not differ.
  This is therefore an example of
• Type 1 error (a error)
• Type 2 error (ß error)
• 1- a
• 1- ß

59
QUESTION 4

• In a RCT, a planned crossover design
• Requires standardisation
• Must take into account possible residual effects
• Enhances generalisability of the study
• Eliminates need for monitoring compliance/non
  compliance

60
Question 5

• Phase 1 trials are
• Trials with large number of subjects are
  recruited
• Done for post marketing surveillance
• Multi-centered
• Small studies with 20-80 patients looking at
  toxic pharmacological effects of drugs

61
ANSWER 1

• The purpose of double blinding or double masked
  study is to
• Achieve comparability of treated and untreated
  subjects
• Reduces effects of sampling variation
• Avoid observer and subject bias
• Introduce observer and subject bias in the study

62
ANSWER 2

• The major purpose of random assignment in a RCT
  is to
• Facilitate double blinding
• Facilitate measurement of outcome variables
• Ensure study groups are comparable on baseline
  characteristics
• Reduce selection bias in allocation of treatment

63
ANSWER 3

• A RCT comparing the efficacy of 2 drugs showed a
  difference between the two (plt0.05). Assume in
  reality, however, the 2 groups do not differ.
  This is therefore an example of
• Type 1 error (a error)
• Type 2 error (ß error)
• 1- a
• 1- ß

64
ANSWER 4

• In a RCT, a planned crossover design
• Requires standardisation
• Must take into account possible residual effects
• Enhances generalisability of the study
• Eliminates need for monitoring compliance/non
  compliance

65
ANSWER 5

• Phase 1 trials are
• Trials with large number of subjects are
  recruited
• Done for post marketing surveillance
• Multi-centered
• Small studies with 20-80 patients looking at
  toxic pharmacological effects of drugs