Epidemiological Study Designs:
Randomized Controlled Trials
Dr. Zaeema Ahmer
Learning Objectives
By the end of this session, the students will be able to:
Explain the application of experimental studies
Classify types of experimental studies
Describe characteristics, types and advantages of randomized controlled
trial (RCT)
Discuss role of blinding in RCT
Calculate relative risk and efficacy to express results of experimental
studies
Epidemiological Study Designs
Observational Experimental
studies Studies
Descriptive Analytical Randomized
Studies Studies Controlled Trials
Case Report,
Case Series, Cohort Non-Randomized
Case Control
Incidence Study, Study Trials
Study
Cross-sectional
Study,
(Ecological Study)
Generate hypothesis Test Hypothesis
Hierarchy of Epidemiologic Study Design
When Experimental designs are preferred over
observed designs?
When exposures/factors are beneficial to health
Ideal design for evaluating the effectiveness and side effects of
new interventions
Iron supplements and reduced anemia
New surgical procedure and reduction in complications
Types of Experimental Studies
Randomized Controlled Trials
Non-Randomized/ Quasi Experimental
Randomized Controlled Trials
Intervention - Investigator manipulates exposure (s)
Randomization - allocation in two groups by chance
Controls for confounding
Control group or reference - placebo or standard
treatment
Control Arm
Standard care (e.g. low-dose aspirin and coronary heart disease)
Placebo (e.g. Hormonal Replacement Therapy vs. placebo)
Active alternative (e.g. Antihypertensive and Lipid-lowering
Treatment to Prevent Heart Attack Trial)
Defined population
Randomized
INTERVENTION NO INTERVENTION
IMPROVED NOT IMPROVED NOT
IMPROVED
IMPROVED
Design of an Experimental Study
Example: 3-day versus 5-day course for Amoxicillin to treat
pneumonia in under-5 children
Cure (a)
3-day course
No Cure (b)
Random
Under-5 pneu
Allocation
monia
children
Cure (c)
No pneumonia 5-day course
(excluded)
No Cure (b)
Period of follow-up
Randomized Controlled Trial: Types
[Link] Trials
Classical clinical trial with one group given drug and other
placebo or old treatment
[Link] Trials (Prophylaxis trials)
Vaccination Trials
Health education trials
Nutrition
3. Health Service evaluation trials
New health initiative in an area vs. non-intervention area
Types of Intervention
Pharmaceutical (Therapeutic or Preventive)
Device
Procedure
Behavior modification
Why Randomize?
Distribute important (and unknown) characteristics evenly
across the study groups
Eliminates selection bias
Makes intervention and control group comparable
Provides the strongest inference for causal inference
Done by table of random numbers, or computer generated
list etc.
Types of Randomization
Simple randomization (First option)
Stratified randomization
Classified into subgroups before randomization (males and
females)
Randomize within subgroups
Cluster Randomization
SIMPLE RANDOMIZATION STRATIFIED RANDOMIZATION
Cluster Randomisation
Cluster Randomisation
Intent-to-treat Analysis
Keep the original groups formed by randomization intact for comparison
of outcomes
Ignore non-compliance, crossing over & late exclusions to keep
randomization
Typically dilutes treatment effect, as a result the estimate of treatment
effect is generally conservative
ITT analysis is usually described as “once randomized, always analyzed”
Randomized Control Trial: Blinding
•Hiding information about the allocated study regimes
Single Participants do not know about the study
Double Participants + Observer do not know who is in which group
(BEST)
Triple Participants + Observer +analyzer all are blind
Only study designer (PI) knows about the study
Exercise: Identify the study design
A survey was conducted in 1996 in a slum and
anthropometric measurements of 5000 children <5years
were taken. 2000 children were found to be malnourished.
Cross Sectional
Exercise: Identify the study design
Prevalence of malnourishment in population is 40%. 2000
malnourished and 3000 well nourished children were followed for 2
years. 280 among malnourished and 220 among well nourished children
developed pneumonia in the period.
COHORT
Exercise: Identify the study design
A study of 1000 people in which 500 are treated with a new drug and
500 are treated with placebo. Assignment to new drug and placebo
groups is assigned randomly.
EXPERIMENTAL-RANDOMIZED CONTROLLED TRIAL
Ways of Expressing Results in Experimental Studies
Relative risk
Efficacy of new intervention
1. Relative Risk
Relative Risk= Incidence in intervention/Incidence in non intervention
A study of 1000 people in which 500 are treated with a new drug and 500 are treated with
placebo for six months. Assignment to new drug and placebo groups is assigned randomly.
100 were cured with new drug while 50 recovered with placebo
Diseased Disease negative
New Drug 400 100 500
Placebo 450 50 500
RR= a/a+b / c/c+d
RR= 400/500 /450/500
RR=0.8/0.9= 0.88
Interpretation: Patients who received the intervention are less likely to be diseased as
compared to those who received the placebo
Efficacy=Rate in those who did not receive intervention-Rate in those who received the intervention
2. Efficacy of
Ratenew
in thoseIntervention
who did not receive the intervention
A study of 1000 people in which 500 are treated with a new drug and 500 are
treated with placebo for six months. Assignment to new drug and placebo
groups is assigned randomly. 100 were cured with new drug while 50
recovered with placebo
Efficacy= 0.9-0.8/0.9
Efficacy=0.1/0.9
Efficacy=11.1%
Interpretation: There was 11% reduction in diseases due to use of new drug
References:
Gordis Chapter 7,8 and 17
Gupta S. K. (2011). Intention-to-treat concept: A
review. Perspectives in clinical research, 2(3), 109–112.
doi:10.4103/2229-3485.83221
THANK YOU