Dr Sridevi NH
MBBS (MD)
Community medicine
Chikmagalur
 Introduction
 Different study designs.
 The basic steps with an example
 Number needed to treat
 Types of RCTs
 Blinding and contamination
 Summary
 References
 The randomized controlled trial (RCT) is one of the
simplest, most powerful tools of research.
 The RCT is a study in which people are allocated ‘at
random’ to receive one of several interventions.
 The intervention is controlled by the investigator
 Interventions include varied actions such as preventive
strategies, diagnostic tests, screening programs, and
treatments.
 Simple, two arm RCT.
 Cross over RCT design.
 Factorial RCT design.
HWilliams & Wilkins, 2001 ulley et al. Designing Clinical Research. 2ndEdition.
Lippincott
 Drawing up a protocol
 Selecting reference and experimental populations
 Randomization
 Manipulation or intervention
 Follow up
 Assessment of outcome
 The protocol specifies the aims and objectives of the
study.
 Criteria for the selection of study and control groups,
sample size, procedure for allocation.
 The exposure variable, outcome variable, confounding
variable and co interventions should be defined clearly.
 Inclusion and exclusion criteria should be enumerated.
 The protocol aims at preventing bias and to reduce the
sources of error in the study.
 Objective: To evaluate effect of stand-alone Care for
Child Development, stand-alone Enhanced Nutrition,
and Combined Care for Child Development and
Enhanced Nutrition.
 All children aged birth to 2.5 months were enrolled and
followed to age 2 years.
 Reference/target population: population to which the
findings of the trial, if found successful, are expected to
be applicable.
 Experimental population: actual population that
participates in the experimental study.
◦ Representative of the population
◦ Qualified or eligible for the trial.
◦ Informed consent.
 Experimental population :children less than two years in
Naushero Feroze district of Sindh Province, Pakistan.
Reference population is children less than two years in
other places of Pakistan
 It is a statistical procedure.
 Assignment is predetermined
 Each individual has an equal chance of being allocated
to all groups
 Best way to allocate participants to different intervention
groups
 Ensures that the only difference between the two
intervention groups is the intervention
 Makes the Study and Control groups comparable
 Minimizes chance of confounding even against unknown
confounders
 Selection Bias minimized
 Facilitates blinding / masking of treatment from
participants and investigators
 Unrestricted randomization:
Simplest method – is tossing a coin, Use random
number tables, Can use computers to generate random
numbers.
 Restricted randomization
Also called block randomization or balanced
randomization
 Randomization is done in blocks – each block will have
equal number of subjects assigned to each group at the
end of the block
 Stratified randomization
 Block randomization may not protect against obtaining
different distributions of important risk factors for the
outcome in the study groups
 Used only when the results of the trial are thought to be
related to a particular factor
 Ensures that different subgroups are balanced between
treatment groups
Matched pair randomization:
It can be used when the experiment has only two
treatment conditions. Subjects are grouped into pairs
based on some variable and then are randomly
assigned to different intervention
Cluster randomization:
 All the methods can be applied to either individuals or
communities
 All individuals belonging to a specific group (village,
ethnic, school etc.) communities are allocated
intervention/placebo
 The Trial was a cluster randomized controlled trial.
 80 LHW catchment areas were selected for the study.
 Stand-alone Care for Child Development
 Stand-alone Enhanced Nutrition
 Combined Care for Child Development and Enhanced
Nutrition
 Control group
 Introduce the Intervention and Placebo Control
Modalities.
 Deliberate application or withdrawal or reduction of the
suspected causal factor as laid down in the protocol.
 Examination of both the group subjects at defined
intervals of time, under the same given circumstances, in
the same time frame till final assessment of outcome.
 Follow up may be short or may require many years
depending upon the study undertaken.
 The study team undertook in-depth assessments when
enrolled children were 6, 12, 18 and 24 months old.
 The incidence of positive and negative results is
rigorously compared in both the groups, if any
differences are found: tested for statistical significance.
 Child development:
 Measured at 12 and 24 months using the Bayley Scales
of Infant and Toddler Development
 Four scales were used:
◦ the cognitive scale,
◦ the communication scale,
◦ the motor scale and
◦ the social-emotional scale
 Child growth and nutrition status was measured by
data collectors using multiple tools.
 Anthropometry (length/height, weight, mid-upper-arm
circumference and head circumference) was taken at
baseline (when each child was enrolled) and when
children were 6, 12, 18 and 24 months old
Subject variation: participants may feel better or report
improvement if they knew they were receiving a new form
of treatment
Observer bias: Investigator measuring the outcome of a trial
may be influenced if he knows beforehand the particular
procedure to which the subject has been subjected
Bias in evaluation: Investigator may subconciously give a
favourable report of the outcome of the trial.
Number of patients that need to be treated/ receive
intervention in order to have an impact on one person
For example, if 80% of patients in the control group got better
and 90% of patients in a treatment group got better, the
absolute risk of not getting better if denied the more effective
treatment is 10%.
NNT = 1/ (IE - INE) (where IE and INE are measured as
proportions out of 1)
For every 10 patients who get this treatment, 1 more would get
better compared to the control group.
 Number of patients that need to be treated/ receive intervention
to in order to have one adverse outcome
 NNH = 1/ARI Absolute risk Increase
 Drug to decrease tumour size in colorectal cancer
 EER = 55/75=0.73 (Experimental Event Rate)
 CER=35/75=0.47 (Control Event Rate)
 ARI = 0.73 – 0.47 = 0.26
 NNH= 1/ ARR = 1/ 0.26 = 4
 Every four patients treated one will die
 Clinical trials: aspirin on cardiovascular mortality
 Preventive trials:vaccines and chemo prophylactic drugs
 Risk factor trials :on CHD
 Cessation experiments: cigarette smoking and lung
cancer
 Trial of etiological agents: oxygen for retrolental fibroplasia
 Evaluation of health services: domiciliary treatment for TB
 Single Blind – Participant does not know which
treatment received
 Double Blind – Doctor and Participant, both do not know
which treatment received
 Triple Blind – Person analyzing the data, Doctor and
Participant, all do not know which treatment received.
 Study participants do not receive their allocated
intervention / placebo
 Study participants receive the intervention even though
they were in control group
 Tends to weaken the observed association
 In summary, RCTs are quantitative comparative
controlled experiments in which a group of investigators
study two or more interventions in a series of individuals
who are randomly ‘allocated’ to receive them.
Thank you
1. Park K. Randomized contolled trials . Parks text book of Preventive and
social medicine. 23rd edition. Jabalpur, M/S Banarsidas Bhanot Publishers.
January 2015 page 80-87
2. Leon Gordis. Epidemiology 4th edition. Elsevier. 2009 page 131-163.
3. Roger detels. Oxford textbook of public health. 5th edition. Oxford 2009
page 527-556
4. Textbook of public health and community medicine 1st edition 2009. AFMC
page 151-158
5. Hulley et al. Designing Clinical Research. 2ndEdition. Lippincott Williams
& Wilkins, 2001

randomised controlled trial

  • 1.
    Dr Sridevi NH MBBS(MD) Community medicine Chikmagalur
  • 2.
     Introduction  Differentstudy designs.  The basic steps with an example  Number needed to treat  Types of RCTs  Blinding and contamination  Summary  References
  • 3.
     The randomizedcontrolled trial (RCT) is one of the simplest, most powerful tools of research.  The RCT is a study in which people are allocated ‘at random’ to receive one of several interventions.  The intervention is controlled by the investigator
  • 4.
     Interventions includevaried actions such as preventive strategies, diagnostic tests, screening programs, and treatments.
  • 7.
     Simple, twoarm RCT.  Cross over RCT design.  Factorial RCT design.
  • 10.
    HWilliams & Wilkins,2001 ulley et al. Designing Clinical Research. 2ndEdition. Lippincott
  • 11.
     Drawing upa protocol  Selecting reference and experimental populations  Randomization  Manipulation or intervention  Follow up  Assessment of outcome
  • 12.
     The protocolspecifies the aims and objectives of the study.  Criteria for the selection of study and control groups, sample size, procedure for allocation.  The exposure variable, outcome variable, confounding variable and co interventions should be defined clearly.  Inclusion and exclusion criteria should be enumerated.  The protocol aims at preventing bias and to reduce the sources of error in the study.
  • 13.
     Objective: Toevaluate effect of stand-alone Care for Child Development, stand-alone Enhanced Nutrition, and Combined Care for Child Development and Enhanced Nutrition.  All children aged birth to 2.5 months were enrolled and followed to age 2 years.
  • 14.
     Reference/target population:population to which the findings of the trial, if found successful, are expected to be applicable.  Experimental population: actual population that participates in the experimental study. ◦ Representative of the population ◦ Qualified or eligible for the trial. ◦ Informed consent.
  • 15.
     Experimental population:children less than two years in Naushero Feroze district of Sindh Province, Pakistan. Reference population is children less than two years in other places of Pakistan
  • 16.
     It isa statistical procedure.  Assignment is predetermined  Each individual has an equal chance of being allocated to all groups  Best way to allocate participants to different intervention groups  Ensures that the only difference between the two intervention groups is the intervention
  • 17.
     Makes theStudy and Control groups comparable  Minimizes chance of confounding even against unknown confounders  Selection Bias minimized  Facilitates blinding / masking of treatment from participants and investigators
  • 18.
     Unrestricted randomization: Simplestmethod – is tossing a coin, Use random number tables, Can use computers to generate random numbers.  Restricted randomization Also called block randomization or balanced randomization
  • 19.
     Randomization isdone in blocks – each block will have equal number of subjects assigned to each group at the end of the block
  • 20.
     Stratified randomization Block randomization may not protect against obtaining different distributions of important risk factors for the outcome in the study groups  Used only when the results of the trial are thought to be related to a particular factor  Ensures that different subgroups are balanced between treatment groups
  • 21.
    Matched pair randomization: Itcan be used when the experiment has only two treatment conditions. Subjects are grouped into pairs based on some variable and then are randomly assigned to different intervention Cluster randomization:  All the methods can be applied to either individuals or communities  All individuals belonging to a specific group (village, ethnic, school etc.) communities are allocated intervention/placebo
  • 22.
     The Trialwas a cluster randomized controlled trial.  80 LHW catchment areas were selected for the study.  Stand-alone Care for Child Development  Stand-alone Enhanced Nutrition  Combined Care for Child Development and Enhanced Nutrition  Control group
  • 23.
     Introduce theIntervention and Placebo Control Modalities.  Deliberate application or withdrawal or reduction of the suspected causal factor as laid down in the protocol.
  • 28.
     Examination ofboth the group subjects at defined intervals of time, under the same given circumstances, in the same time frame till final assessment of outcome.  Follow up may be short or may require many years depending upon the study undertaken.
  • 29.
     The studyteam undertook in-depth assessments when enrolled children were 6, 12, 18 and 24 months old.
  • 30.
     The incidenceof positive and negative results is rigorously compared in both the groups, if any differences are found: tested for statistical significance.
  • 31.
     Child development: Measured at 12 and 24 months using the Bayley Scales of Infant and Toddler Development  Four scales were used: ◦ the cognitive scale, ◦ the communication scale, ◦ the motor scale and ◦ the social-emotional scale
  • 32.
     Child growthand nutrition status was measured by data collectors using multiple tools.  Anthropometry (length/height, weight, mid-upper-arm circumference and head circumference) was taken at baseline (when each child was enrolled) and when children were 6, 12, 18 and 24 months old
  • 35.
    Subject variation: participantsmay feel better or report improvement if they knew they were receiving a new form of treatment Observer bias: Investigator measuring the outcome of a trial may be influenced if he knows beforehand the particular procedure to which the subject has been subjected Bias in evaluation: Investigator may subconciously give a favourable report of the outcome of the trial.
  • 36.
    Number of patientsthat need to be treated/ receive intervention in order to have an impact on one person For example, if 80% of patients in the control group got better and 90% of patients in a treatment group got better, the absolute risk of not getting better if denied the more effective treatment is 10%. NNT = 1/ (IE - INE) (where IE and INE are measured as proportions out of 1) For every 10 patients who get this treatment, 1 more would get better compared to the control group.
  • 38.
     Number ofpatients that need to be treated/ receive intervention to in order to have one adverse outcome  NNH = 1/ARI Absolute risk Increase  Drug to decrease tumour size in colorectal cancer  EER = 55/75=0.73 (Experimental Event Rate)  CER=35/75=0.47 (Control Event Rate)  ARI = 0.73 – 0.47 = 0.26  NNH= 1/ ARR = 1/ 0.26 = 4  Every four patients treated one will die
  • 39.
     Clinical trials:aspirin on cardiovascular mortality  Preventive trials:vaccines and chemo prophylactic drugs  Risk factor trials :on CHD  Cessation experiments: cigarette smoking and lung cancer  Trial of etiological agents: oxygen for retrolental fibroplasia  Evaluation of health services: domiciliary treatment for TB
  • 40.
     Single Blind– Participant does not know which treatment received  Double Blind – Doctor and Participant, both do not know which treatment received  Triple Blind – Person analyzing the data, Doctor and Participant, all do not know which treatment received.
  • 41.
     Study participantsdo not receive their allocated intervention / placebo  Study participants receive the intervention even though they were in control group  Tends to weaken the observed association
  • 42.
     In summary,RCTs are quantitative comparative controlled experiments in which a group of investigators study two or more interventions in a series of individuals who are randomly ‘allocated’ to receive them.
  • 43.
  • 44.
    1. Park K.Randomized contolled trials . Parks text book of Preventive and social medicine. 23rd edition. Jabalpur, M/S Banarsidas Bhanot Publishers. January 2015 page 80-87 2. Leon Gordis. Epidemiology 4th edition. Elsevier. 2009 page 131-163. 3. Roger detels. Oxford textbook of public health. 5th edition. Oxford 2009 page 527-556 4. Textbook of public health and community medicine 1st edition 2009. AFMC page 151-158 5. Hulley et al. Designing Clinical Research. 2ndEdition. Lippincott Williams & Wilkins, 2001

Editor's Notes

  • #14 The PEDS Trial tested the effects on child development and growth of two intervention packages, Care for Child Development and Enhanced Nutrition, along with a third package that combined both interventions The Pakistan Early Development Scale-up (PEDS) Trial was implemented in Naushero Feroze district in Sindh Province, Pakistan from June 2009 through March 2012.