Intervention Studies

Dr. Birhanu J
 Definition
• Intervention study – an investigation
involving intentional change in some
aspects of the subjects.
• E.g introduction of a preventive or
therapeutic regimen
 Key Features of Experimental Design

1) Investigator manipulates the

condition under study
2) Always prospective
 Types of interventions

1. Prophylactic - focus on prevention (e.g.

vaccines)

2. Diagnostic - focus on evaluation of new

diagnostic procedure

3. Therapeutic - focus on treatment (e.g.

drug testing, evaluation of new surgical
technique, etc.)
 Other ways of classifying
intervention studies
A. Classification based on the population studied
A.1 Clinical trial
• usually performed in clinical settings
• the subjects are patients
A.2 Field trial
• used in testing medicine for preventive purpose
• subjects are healthy people e.g. vaccine trial
A.3 Community trial
• unit of the study is group of people/community
• e.g. Dental caries study where one entire community is allocated at
random to receiving sodium fluoride added to the water supply, while
the other continued receiving water without supplementation
B. Classification based on design

B.1 Uncontrolled trial

• no control group
• control will be past experience (history)
B.2 Non-randomized controlled
• there is control group
• allocation to either group is not randomized
B.3 Randomized controlled
• there is control group
• there is random allocation of subjects to
either group
C. Classification based on objective

C.1 Phase I
 trial on small subjects to test the safe dose range
& side effects;
20-80 healthy volunteers needed.
C.2 Phase II
 trial on small group to determine the therapeutic
effect
 100-200 ill volunteers needed.
C.3 Phase III
• study on large population
• usually randomized controlled trial
 D. Classification based on the
method of treatment administration
D.1 Parallel treatment
Individuals in each group simultaneously receive
one study treatment

Group 1 →Treatment A →Follow up & outcome assessment

Group 2→ Treatment B→ Follow up & outcome assessment

D.2 Crossover design

• Two or more study treatments are administered
one after another to each group
• All trial participants receive all of the treatments
and only the order of the treatment differs
 Cross over design cont…

• The groups usually switch treatments at

the same time

• there is often a washout period between

the end of one treatment and the start of
another
 Cross over design cont..

Group 1 →Treatment A → Washout →Treatment B →

Follow up & outcome assessment

Group 2→ Treatment B→ Washout → Treatment A →

Follow up & outcome assessment
 Cross over design cont..
• Cross over design is appropriate to study the
efficacy of therapies intended to reduce the
frequency or severity of chronic, recurrent
problems, such as seizures, arthritic pain

• the subject can serve as his/her own control

• The rates of outcome events (e.g seizures) or
levels of symptoms (e.g joint pain) that are
present during or at the end of the periods of time
are compared within individual study subjects
 E. Classification based on the
number of treatments being tested
E.1 Simple design
• Each group receives a treatment
consisting of one component
• One hypothesis tested

E.2 Factorial design

• Two or more treatments are combined
• More than one hypotheses are tested
 Factorial Design cont…

• A clinical trial for two hypotheses can utilize a two-

by-two factorial design

• subjects are first randomized to treatments α and

β to address one hypothesis, and then within each
treatment group there is further randomization to
treatment A and B to evaluate a second question
 α β

A B A B
 Factorial Design cont…

The advantage of factorial design is its

ability to answer two or more questions in
a single trial for only a marginal increase in
cost
 Factorial Design cont…

Ideally the additional treatments in a

factorial design should not:
• complicate trial operations,
• materially affect eligibility requirements, or
• cause side effects that could lead to poor
compliance or losses to follow up
 Factorial Design cont…
• the possibility of interaction between
treatment regimens must be considered.

• Factorial design allows an investigator to

test the separate and combined effects of
several agents
 Features of Randomized Trial
• Avoid bias in assignment into alternative
groups

• Avoid bias in the assessment of the

endpoints by the researcher

• Avoid bias in behavior of the participants

due to knowledge of their intervention
allocation
 Methodological Issues in
Intervention studies

• Ethics

• Randomization

• Blinding

• Placebo use
 Challenges of Intervention Studies

• Experiment on human beings is an

ethically charged undertaking

• Potential interventions must be safe

and acceptable for human use

• Dose schedule of potential

interventions must be established
 Preventive trials
Primary prevention trials
• Are trials that prevent or delay the onset of
disease among healthy individuals
• While therapeutic trials are virtually always
conducted among individuals, primary
prevention measures can be studied
among either individuals or entire
population
 Preventive trials cont…
Secondary prevention trials
are trials that prevent or delay
progression among diseased individuals
 Preventive trials cont…
In some preventive trials, the alleged
causal factor is reduced or removed. E.g
removing lead-containing soil from around
children‟s homes and replacing with clean
soil to test whether it reduces blood lead
levels and their risk of lead poisoning
 Preventive trials cont…
 In other preventive trials, agents are
studied to determine if they are effective in
reducing disease occurrence or recurrence

 E.g adding vitamins, minerals etc to

prevent different diseases related to micro-
nutrient deficiency
 Preventive trials cont…
Prevention trials usually take many years to
conduct & requires tens of thousands of
participants because they often focus on
reducing the incidence of diseases that
typically occur at a yearly cumulative
incidence of 1% or less
Studies of diagnostic and screening tests

• Diagnostic tests – tests performed in

persons with a symptom or sign of an illness

• Screening tests- test done in individuals with

no symptoms & signs
Studies of diagnostic and screening tests cont..

• To assess a test‟s validity, the results it

provided can be compared either
 with a “true” measure of the physiologic,
biochemical, or pathologic state the test is
seeking to characterize or
with the occurrence of disease progression or
a disease complication that the test result
seeks to predict
 Studies of diagnostic and
screening tests cont..
• For example, arterial blood pressure as
assessed via a sphygmomanometer can
be compared either
 with reading obtained from direct intra-
arterial pressure measurements or
with the subsequent occurrence of stroke
or other forms of cardiovascular disease
 Studies of diagnostic and
screening tests cont..
• The validity of the clinical criteria/test could be
described in terms of the degree to which
persons with and without the condition under
study are correctly categorized – sensitivity,
specificity
• Sensitivity – percentage of persons with the
disease who tested positive by the clinical
criteria/test
• Specificity – percentage of persons without the
disease who were correctly categorized as
negative by the clinical criteria/test
 Studies of diagnostic and
screening tests cont..
• Alternatively, the validity of the clinical
criteria/test could be expressed as the extent to
which being categorized as positive or negative
actually predicts the presence of the disease –
predictive value of a positive test (PV+),
predictive value of a negative test (PV-)
• PV+ - the percentage of persons who were
deemed clinically positive/positive by the test
and who were found to have the disease
• PV- - the percentage of persons who were
clinically negative/negative by the test and who
truly had no the disease
 Therapeutic trials
• Therapeutic trials are commonly called
clinical trials because they are conducted in
a clinical setting among diseased patients
and use a clinical outcome measure such as
cure, recurrence, side effect etc
• A new therapy is usually compared to the
standard treatment or placebo
 Therapeutic trials cont..
• Because therapeutic trails attempt to reduce
the occurrence of relatively common
outcomes (e.g recurrence) they usually only
take one or two years to conduct & generally
involve only a few hundred patients
 Intended degree of
effectiveness of the new drug
Superiority trials
• seek to show that a test drug is superior to the
control (placebo, no treatment, or a lower dose
of the test drug).
• Superiority trial also includes comparison with
an active treatment (standard treatment) if the
intent is to show that the new drug is more
effective than the control.
Intended degree of effectiveness of the
new drug cont..
Equivalence trial
• the intention of an equivalence trial is to
show that two treatments have about the
same therapeutic effect.
• the goals set for an equivalence trial are
less ambitious than the goals set for a
superiority trial.
Intended degree of effectiveness of the
new drug cont..
While one might strive for new treatment to be
superior to existing treatment it is more
realistic to hope for an equally good
performance regarding outcomes of disease
and to focus on other possible benefits of the
new treatment, such as side effects, cost,
ease of administration
Intended degree of effectiveness of the
new drug cont..
Non-inferiority trials
are trials that seek to show any difference
between the two treatments is small
enough to allow a conclusion that the new
drug has an effect that is not too much
smaller than the active control.
 Studies of diagnostic and main steps in an RCT
1. Identify new drug/intervention
2. Identify comparison – e.g standard treatment versus
placebo
3. Define eligible patient population/ exclusions
4. Define the outcomes and how to assess them
5. Write the protocol
6. Obtain research ethics committee approval

7. Recruit & consent required

8. Randomize to treatment, then treat

9. Follow-up & compare/analyze outcome data

10. Publish/disseminate findings

 Small group work
• In southern part of Ethiopia, a plant called
“Shiferaw” is used by many people to treat
hypertension
• How will you conduct scientific research to
assess the effectiveness of “Shiferaw” in
treating hypertension?
 Step 2 in an RCT: identify
comparator to the exposure
• RCT aim: to ensure that subjects being compared
differ only in their exposure to the intervention being
considered.
• Thus, all other differences between groups are due
to chance alone.
• E. g comparator = standard treatment or placebo
• Placebo as comparator = dummy treatment.
i.e belief/expectancy of benefit, although treatment
itself is inert.
 Comparing with historical controls
• Previous (historical) patients are compared with
patients now receiving new treatment

• Problem- Current patients and previous patients

may not be similar

• There can be differences in:

 referral patterns & case-mix

 treatment environment and supportive care

 quality of past data & detection of all variables

 Step 3 in an RCT: Defining the
target population/sampling frame
• Who is eligible to be included in the RCT?
• Who, as a result, will be excluded?
• Trade-off: narrow inclusion criteria Vs external
validity (generalizability) –
i.e -excluding those to whom treatment will be
eventually offered
• E.g surgical trials exclude older people due to
greater risk, likelihood of refusal etc
Step 4 in RCTs: Outcome measures
• What is the primary outcome of interest?
• What are the secondary outcomes?
• What is the best method of detecting them?
• A good outcome measure (means of detecting
outcomes) is:
– Valid (i.e measures what it should measure)
– Reliable (i.e produces repeatable results)
– Objective (not amenable to subjective interpretation)
– Feasible (resources are available/easy administration)
– Acceptable
 Relating outcomes to exposure

The role of blinding

• It avoids bias due to preference

• Help in avoiding bias during treatment,

monitoring and endpoint assessment
 Blinding
• Conditions in which blinding is not
necessary:
– If outcome is clearly defined and objectively measured

– If participants are less likely to be influenced by

knowledge of which intervention is given
 Types of blinding/ masking
1. Single blinding - the observer is aware but the
subject is not aware of treatment assignment

2. Double blinding - Neither the observer nor the

subject is aware of treatment assignment

3. Triple blinding - The observer, subject, and

data analyst are not aware of treatment
assignment
 Method of Blinding
• Use of identical placebo

• Blinding which treatment is active when

an identical placebo cannot be
obtained
 Steps 5 and 10 in an RCT: the
protocol & publishing
• Critical appraisal of the quality of clinical trials
is possible only if the design, conduct, and
analysis of RCTs are thoroughly, and
accurately described in published articles.

• A group of scientists and editors developed

the CONSORT (Consolidated Standards of
Reporting Trials) statement to improve the
quality of reporting of RCTs.
 Steps 5 and 10 in an RCT: the
protocol & publishing cont..
• The CONSORT Statement is intended to improve
the reporting of a randomized controlled trial (RCT),
enabling readers to understand a trial's design,
conduct, analysis and interpretation, and to assess
the validity of its results.

• It emphasizes that this can only be achieved through

complete transparency from authors.
• Refer the evidence published from RCTs which
reflect the latest CONSORT guidelines:
http://www.consort-statement.org/
Step 6 in an RCT: ethics committee
approval
• Getting approval from the ethics
committee is mandatory

• Ethics and invitation to participate – issue

of informed consent.
 Ethical issues in Trials

• Informed consent: with full understanding of the

nature of the study, special attention to those not competent
to give consent…

• Confidentiality: keeping securely all information

collected in research

• Coercion: unacceptable under any circumstances. Be

careful in using local authorities and health care providers in
research …
 Ethical issues in Trials
cont..
• Scientific merit: appropriate methods, taken
into account available knowledge, appropriate
sample size, appropriate duration, … results must
lead to action

• Equitable selection of subjects: equal

distribution of benefits and potential harm
 Ethical issues in Trials
cont..
• Obtaining communal consent:
– eliciting consent through intermediary of trusted community leader;
useful in large preventive field trials
– … but ensure freedom to refuse to participate or withdraw at any
time without penalty

• Potential benefit and the risk of harm:

– any harm that results from deliberate medical intervention is
unacceptable even when the balance is in favor of the intervention;
– research should not take-over or undermine existing services
 Ethical issues in Trials
cont..
• Incentives: avoid monetary payment except in
covering expenses or lost income

• Feedback of results: to participants, relevant

authorities

• Anonymity of communities: keep identity of

the community in discussing sensitive behavioral issues
in publications
 Step 7 in an RCT: recruitment of
sample
 Sample size estimation
– Make sure that your study is large enough to confidently answer the
question and generalize
– That any effect you see is not just due to chance
 Power
• Ability/chance of a study ( e.g 80%) to detect an effect of treatment, if
in reality that effect exists, having „excluded‟ the role of chance
 Number of subjects
• Consult standard Biostatistics text books
•  ?Need to go multi-centre to recruit large sample but….
• Depends on:
– patient/clinical effect size sought
– Estimate of prevalence
– Level of confidence required (usually 95%)
 Step 8 in an RCT: randomize

• Randomization – equal chance of allocation

to any treatment i.e no allocation bias
• BUT, randomization is not always feasible –
e.g due to patient preferences
Advantages of Randomization
• guarantees objectivity of allocation

• the effect of factors other than those under

study is balanced out
– the larger the groups, the more effective this balance
 Randomization procedure

1st stage 2nd stage

Allocating each selected
Selecting participant into groups
participants on the based on equal
basis of the criteria opportunity
 Advantages and disadvantages of RCTs:

• Advantages • Disadvantages
– Prospective – Unnatural situation
– Randomization – Human behavior may be
difficult to control
– Clear temporal sequence
– Best evidence for causation – Ethical constraints
– Exclusions may limit
generalizability
– Expensive in time, personnel,
facilities, and budget
 Step 9 in an RCT: analysis
• Results can be arranged in two-by-two
table
• Calculate measures of association
 Example
Table 1: Reduction of Maternal-Infant transmission of HIV with
Zidovudine

Treatment Infant HIV infected

Yes No Total
Yes (Zidovudine) 13 167 180
No (Placebo) 40 143 183
 Example cont…
• Relative Risk (RR) = 0.33 i.e Infants whose
mothers took zidovudine had one third the risk of
becoming HIV infected than did the infants whose
mothers took placebos
• Or there was a 67% relative reduction in the risk
of HIV transmission among the zidovudine group
 Analysis cont...
• Analysis must deal with inevitable refusals,
drop-outs/ loss to follow-up, non-
compliance
• Bias due to withdrawals/drop-outs can be
avoided by comparing outcomes of all the
subjects originally allocated to each group
(in countries with vital event registration)
Intent-to-treat analysis/ treatment
assignment analysis
• All individuals who were randomly allocated to a
treatment are analyzed, regardless of whether
they completed the regimen or received the
treatment

• It gives information on the effectiveness of a

treatment under everyday practical conditions
 Intent-to-treat analysis cont…
Purpose of intent-to-treat analysis
1) it preserves the benefit of randomization
• i.e it preserves baseline comparability of the groups for known and
unknown confounders
2) it maintains the statistical power of the original study population
3) because good and poor compliers differ from one another on
important prognostic factors, it helps ensure that the study results
are unbiased
4) it gives information on the effectiveness of a treatment under
everyday practical conditions
• It answers the question, “How well does the treatment work among
those who are offered it?”
 Efficacy analysis
• Is an alternative to intent-to-treat analysis
• It determines the treatment effects under ideal
conditions, such as when participants take the full
treatment exactly as directed
• It answers the question, “How well does the
treatment work among those who take the full
treatment as directed?”

• Some times investigators conduct both intent-to-

treat and efficacy analysis
 Enroll eligible and willing patients

Random assignment

Treatment 1 Treatment 2

Completed Did not complete Completed Did not complete

Treatment 1 Treatment 1 Treatment 2 Treatment 2
Group A Group B Group C Group D

Intent to treat analysis considers A, B, C and D

Efficacy analysis considers only A and C

 Stopping rules in clinical trial
Early termination of a clinical trial can be
considered if:
1. Interim results show a clear and extreme
benefit on the primary end point due to the
intervention
2. If one treatment is clearly harmful
 Reference
• Kenneth J. Rothman, Sander Greenland, Timothy L.
Lash. Modern Epidemiology. Philadelphia:Lippincott
Williams & Wilkins, 2008, 3rd edition.
• Anne Aschengrau, Geoge R. Seage III. Essentials of
Epidemiology in Public Health. Canada: Jones and
Bartlett Publishers, 2008, 2nd edition
• Charles H. Hennekens, Julie E. Buring. Epidemiology in
Medicine. Philadelphia:Lippincott Williams & Wilkins,
2008, 3rd edition.
• Ann Bowling. Research Methods in Health. Investigating
Health and Health Services. Buckingham: Open
University Press, 1997