Experimental study

design

Dr. Shumaila Shabbir

Assistant Professor
Community Medicine
 Learning Objectives

• To identify different types of interventional studies

• To describe the steps of randomized control trial

• To differentiate between the designs of randomized control trial

• To classify the types of randomized control trial

• To understand various challenges of randomized control trial

• To define quasi experimental studies and determine its types
 Epidemiological studies
types

Observational Interventional studies

Studies

Descriptive Analytical Non-

studies studies Randomized randomized

Cross Cross
Case Case Case ecologi
section section cohort
report series control cal
al al
 Interventional study design

• Randomized control trials

• Involves random allocation

• Also called experimental study designs

• Non-randomized trials
• No random allocation

• Also called non-experimental study designs

 Randomized Control Trial
• A real epidemiological experiment and most reliable method of evaluation of
new programs and therapies

• Involves some action, intervention or manipulation

• Conditions are under control of the investigator

• Same advantages and disadvantages as cohort study but three additional issues:
• Cost

• Ethics

• Feasibility
 Randomized Control Trial
AIMS:

• To provide scientific proof of etiological or risk factors

• To provide a method of measuring the efficacy, effectiveness and

efficiency of health services
 Randomized Control Trial Steps
1. Drawing up a protocol

2. Selecting reference and experimental population

3. Randomization

4. Manipulation or intervention

5. Follow-up

6. Assessment of outcome
 1. The protocol

• Specifies aims, objectives, questions to be answered, criteria of

selection of participants, sample size, procedure of allocation,
treatments to be applied up to the evaluation of outcome

• Prevents bias

• Reduces error

• Pilot studies to be made before original research

 2. Selecting reference and
Experimental population
• Reference/ target population:
• Population to which the findings of the trial are expected to be applicable

• May include whole world or may be limited to a certain group. E.g. age, sex,
occupation
 2. Selecting reference and
Experimental population
• Experimental / study population:
• It is derived from reference population ( ideally random selection)

• It is the actual population that participates

• It should have same characteristics as reference population

• Participants are different from non participants

• Must fulfil three criteria:

• Must give “ Informed consent”

• Should be representative of reference population

• Should be eligible for the trial

 3. Randomization

• It is a statistical procedure of allocating participants into “ control

group” and “ Study group”

• Study group : to receive an experimental, preventive or therapeutic

procedure, maneuver or intervention

• Control group: not to receive an experimental, preventive or

therapeutic procedure, maneuver or intervention
 3. Randomization

• Every individual gets an equal chance of being allocated in either

group

• It eliminates bias ( e.g. selection bias)

• Both the study and control groups should be alike with regards to
certain variables that might affect the outcome

• Allows for comparison

 3. Randomization

• How to randomize?
• Flip coin method

• Table of random number

• Computer generated assignments

 4. Manipulation

• To intervene or manipulate the study (experimental) group

• Deliberate application or withdrawal or reduction of the suspected

causal factor. e.g. Drug, vaccine, dietary component, a habit, etc)
 5. Follow up

• Examination of study and control groups at defined intervals of time

• Follow up may be short or may take several years

• Limitation: Attrition - losses to follow up, death, migration or loss of

interest in study by the participants
 6. Assessment

• Assessment of outcome in terms of positive or negative results

• Positive results : benefits of intervention . E.g. reduces incidence or

severity of disease, cost to health services

• Negative results: severity and frequency of side effects and

complication, death
 Bias
“An error in the conception and design of a study – or in the collection,
analysis, interpretation, reporting, publication, or review or data –
leading to results or conclusions that are systematically (as opposed to
randomly) different from truth”- (International Journal of Epidemiology)
 Biases
• Three sources of Bias

• First-Subject variation:
• Participant may subjectively feel better

• Second-Observer bias (Selection bias) :

• Researcher can be influenced if he knows beforehand the intervention to
which the patient has been subjected

• Third-Bias in evaluation :
• Investigator may give subconsciously a favorable report on outcome
 Biases
• Solutions for biases:
Blinding
 Blinding
• Three ways

• Single blind trial : participant is not aware of his group, control or

study

• Double blinded trial : neither the doctor nor the patient is aware of
group allocation. Most frequently used.

• Triple blind trial : participant, doctor and the data analyzer all are
blind. This is ideal blinding. Only the principle investigator knows
 Types of RCTs

1. Clinical trials
• Concerned with evaluation of therapeutic agents

• e.g. drugs (folate for NTDs),tonsillectomy ( for recurrent throat infections)

2. Preventive trials
• usually primary preventive trials

• e.g. vaccines and chemoprophylactic drugs

 Types of RCTs
3. Risk factor trials

type of preventive trials

modification of risk factor

e.g. reducing blood cholesterol for hypertension

4.Cessation Experiment

type of preventive trial

termination of a habit or removal of suspected agent

e.g. cigarette smoking and lung cancer

 Types of RCTs

5. Trial of etiological agents

confirms or refute etiological hypothesis

e.g. retrolental fibroplasia in premature babies

6. Evaluation of health services

to assess effectiveness and efficiency of health services

e.g. some of the physicians tasks can be performed by the nurses

 Challenges in RCTs

• Recruiting a sufficient number of eligible participants

• Volunteers are different from non volunteers

• Retaining participants for full duration of study

• Loss to follow up

• Non compliance

• Cost

• Contamination
 Analysis

• Relative risk

• Efficacy

• Number needed to treat - NNT

• Number needed to harm- NNH

• Survival curves
 Quasi experimental research

• Quasi means “ resembles”

• Quasi experimental research resembles experimental research but

not a true research

• Lacks randomization
 Quasi experimental research

• Usually conducted when an RCT cannot be conducted due to

• ethical issues. e.g. Cigarette smoking and lung cancer or depriving one group
from treatment or prophylaxis ( COVID vaccine)
• practical issues. e.g. intervention is expensive, funds are not available for the
research
 Quasi experimental research

• Advantages:
• Easier to conduct than RCT

• Establishes cause and effect relationship

• Less expensive than RCT

• Disadvantages:
• Confounding variables can distort the results

• Lacks randomization
 Quasi vs Randomized Control Trial
• EXAMPLE 1:
• You want to study the impact of a new psychological therapy on depression
through RCT but the HOD doesn’t give you the permission. So you conduct
quasi experimental study

• EXAMPLE 2:
• Dr. Ali is giving new oral hypoglycemic agent to his patients while Dr. Sajjad is
using the old oral hypoglycemic. You observe the outcome in both the groups
and compare them.
 Types of Quasi experimental
research
• Non equivalent group design

• Natural experiments

• Pre test post test design

 Non equivalent group design

• Researcher chooses existing groups which are similar but only one
group experience treatment or intervention

• In RCT both the groups are equivalent in all they ways except the
treatment

• In quasi experimental study the groups are non random, they may
differ in other ways, hence non-equivalent
 Non equivalent group design

• You hypothesize that a new after-school program will lead to higher

grades. You choose two similar groups of children who attend
different schools, one of which implements the new program while
the other does not.

• By comparing the children who attend the program with those who
do not, you can find out whether it has an impact on grades
 Natural experiments

• An external event or situation (“nature”) results in the random or

random-like assignment of subjects to the treatment group

• Natural circumstances that resemble experiment

 Natural experiments

• Examples:

• Death due to cardiac issues in earth quake victims vs death due to

cardiac issues in non victims

• Victims of bomb blast in Hiroshima and Nagasaki vs non victims

• Smokers and non-smokers

• Different religious or socio economic groups

 Pre test post test design

• Also called before and after comparison studies

• Two types:
• Before and after comparison studies without control

• Before and after comparison studies with control

 Pre test post test design

• Before and after comparison studies without control :

• Compare the incidence of outcome before and after the introduction

of the intervention in the absence of a control group

• e.g. seal belt legislation in city A

 Pre test post test design

• Before and after comparison studies with control :

• Compare the incidence of outcome before and after the introduction

of the intervention with a natural control group

• e.g. seal belt legislation in city A with no seat belt legislation in city B

• The two comparison groups should be as similar as possible except

the presence or absence of intervention
 Sources
• Park’s textbook of preventive and social medicine- 24th edition
• Ilyas Ansari’s book of Public Health and Community Medicine
• Gordis epidemiology-7th edition
• https://www.scribbr.com/methodology/quasi-experimental-design/
• https://opentextbc.ca/researchmethods/chapter/quasi-experimental-
research/
THANK YOU