Clinical Study Design
Sandipan Dasgupta, [Link], Ph.D
Associate Professor
Faculty, MAKAUT, WB
Definition of Clinical trial
• It is a systematic study of new drug(s) in human
subjects to generate data for discovering
and/or verifying the clinical, pharmacological,
and/or adverse effects with the objective of
determining their safety and/or efficacy of the
new drugs.
Key elements of Clinical Trials
• The acronym PICO is used by health professionals
to convey all elements of the clinical scenario in
an orderly fashion:
• P - Patient, Population of patients,Problem
• I - Intervention (a therapy, test)
• C - Comparison (another therapy,
Bias
• Bias is a systematic error contained in the study
design, conduct or interpretation of a study.
Whereas extensive lists of particular bias forms
exist,
there are two basic forms of bias:
1. Selection bias occurs if study
• populations are selected in an erroneous way
that comparison groups are not comparable.
2. Information bias occurs if
• measurements are different
Techniques to avoid bias
The two most important techniques are:
1. Randomisation of subjects.
2. Blinding of subjects as well as investigator.
3. Monitoring of clinical trial.
4. Checking original source documents.
5. Source data verification.
6. Clinical data management.
7. Quality control (QC) and Quality assurance
(QA) procedures.
Randomisation
• It aims to obviate systematic differences between
groups due to factors other than intervention.
• It gives each patient a known (or equal)chance of being
assigned to any of the groups.
• The most common methods of
• randomisation are:
1. Simple randomization
2. Blocked randomization
3. Stratified randomization
4. Cluster randomization
Blinding
• Single- blind design:
• Double- blind design:
• Triple- blind design:
Types of Study Designs
Research study design
Observational study
Types of study
• Descriptive study: is to generate hypothesis
• Analytical study: is to test hypothesis
• Experimental study: is to prove hypothesis.
Case report and case series
Analytical Studies
• Case-control study:
Case-control study is a retrospective study
Advantages:
• Less expensive
• Easier to do and take less time compared to most prospective studies
• Can be useful when obtaining follow-up data that is difficult to obtain due to the nature
of population being studied.
• When disease being studies is either rare or when there is a long period of time between
exposure to the time of manifesting the outcome, this study design can be more
efficient.
• Depending on the exposure and outcome of interest, this design may be the only ethical
way to evaluate something.
Disadvantages:
• Potential recall bias
• Subject to selection bias
• Generally do not allow investigators to calculate an incidence or absolute risk
Cohort study
It is a Prospective study.
A cohort study is a study, starts with exposure (risk factor) and follow for the
outcome (disease). That means the study consists two group’s first group is
exposed to some risk factor and second group free from exposure
Cohort study can measure risk (it is the probability of developing an outcome
when exposed to a drug) it is calculated by the relative risk (RR), incidence
(absolute risk) and risk difference (attributable risk).
Relative risk
Incidence =
Cross Sectional Study
Cross sectional study is a study in that measures exposure and outcome at the same
time in randomly selected population, that means ask the questions and get the
answers regarding exposure and outcome (determines presence or absence of
exposure and presence or absence of disease outcome for each individual), follow up
is not necessary for this study (no time sense, people examine only one point of time).
Advantages :
• data available readily, good for assessing the
prevalence,
• follow up not necessary,
• fast and inexpensive.
Disadvantages :
• not suitable for acute disease,
• no comparison group,
• susceptible to bias and causality cannot be
determined.
Experimental study design
• The study Involves intervention
A. Randomized : Participants are randomly allotted
i. Control study
a. Parallel
b. Cross over
ii. Uncontrolled study
• Non randomized: No randomization
i. Non randomized controlled study
ii. Non randomized uncontrolled study
How to categorize clinical
trials?
Trial format:
1. Traditional designs for clinical trials
A. Parallel group trials
B. Cross over trials
C. Factorial design
D. Add on design
E. Randomized withdrawal design
F. Early escape design
2. Special Design issues for small clinical trials
A. N- of- 1 design
B. Sequential design
C. Decision analysis- based design
D. Adaptive design
E. Risk based allocation design
Miscellaneous Designs
A. Cluster randomized design
B. Enrichment design
C. Placebo Challenging design
D. Blind Reader design
E. Trial with Zelen’s design
F. Trial with Wennberg’s design
G. Trial with Comprehensive cohort design
H. Design using historical controls
I. Rolling design
Randomized Clinical Trial
• A randomized clinical trial is also known as parallel group
randomized trials or randomized controlled trials.
• Randomized clinical trials involve randomizing subjects
with similar characteristics to two groups (or multiple
groups):
a. the group that receives the intervention/ experimental
therapy and
b. the other group that received the placebo (or standard
of care).
This is typically performed by using random number tables,
a computer software, manually or by other methods.
Parallel control design
Treatment group Control group
Test drug
• Most common clinical design.
• Complete randomized design in which each patient receives one and only one
treatment in a randomized fashion.
• In this study interventional group receives interventional drug and control group
receive either old drug or placebo .
For randomized controlled study is also named as:
a. Clinical trail
b. Gold standard study.
Advantages:
a) It’s simple and easy to implement.
b) It is universally acceptable.
c) It is applicable to acute conditions.
d) Analysis is less complicated and interpretation is
straight forward.
Disadvantages:
a) It does not into account the interindividual
variability.
Matched Pair Parallel Design
Matching of patients
B is done before
randomization
In this method, pairs of D
subjects are formed
possessing the same
characteristics and
who might be expected E
to respond similarly to
the
treatments.
Advantages:
a) Requires small study population.
b) Can reduce variability from treatment
comparison (compared with parallel from designs).
Disadvantages:
a) The prognostic characteristics are not easily defined.
b) Patient recruitment is slow.
c) When the number of co-varieties is large, this design
is difficult to implement.
Cross over design
A crossover design is a modified randomized
block design in which each block receives more
than one treatment at different dosing periods.
Advantages:
1. It allows a within-patient comparison between
treatments, since each patient serves as his or
her own control.
2. It removes the interpatient variability from the
comparison between treatments.
3. With a proper randomization of patients to the
treatment sequences, it provides the best unbiased
estimates for the differences between treatments.
Disadvantages:
1. Carry- over effects: The residual influence of
treatments on subsequent treatment periods.
Avoided by wash out period.
2. Order effects: Order in which the test drug are
administered affects the outcome.
3. Period effects: The diff. between the study
periods.
4. Drop-outs can be higher.
II. Randomized Uncontrolled Study
• In this study there is no control group present
only one group present with randomization.
• It is so weaker when this study compared with
the randomized control study.
Result
Non - randomized Study
Non - randomized controlled Study
Non - randomized Uncontrolled Study
• This study is also same as that of randomized
uncontrolled study but without randomization
and control group.
• Outcomes measured finally after the collection.
Different Randomized and Non-Randomized Trails
1. Cluster Randomized Study
Cluster means group and this study itself consists
subjects as randomization. The study goes on group
wise, not on the basis of individuals.
Latin Square Design
Latin square design helpful to study more than the one treatment
simultaneously to understand the usefulness of each treatment by
the comparison. For example, four treatments namely – A, B, C, D.
then Latin square is applied (four replications of each treatment) as
follows and do comparison.
A B C D
B C D A
C D A B
D A B C
Factorial Randomized Study
In this study, only one control group and two interventional
groups are present. The results of each group comparing with
the control group. And with the help of this study we can assess
at a time two factors (treatment and dosage form).
Interventional Interventional
Control
group I group II
Add on design is especially useful for testing of an experimental
interventions that have mechanism of action different from that of
established effective treatment.
Study designs for small populations
Defined as <50 possible patients recruited in 5years
with multicentre∕ multinational recruitment.
1. Rare diseases
2. Unique study populations (e.g. Astronauts)
3. Individually tailored therapies
4. Environments that are isolated
5. Emergency situations
6. Public health urgency
7. Restricted resources coupled with an important need
N of 1 design
1. They are cross over trials in which one participant receives the experimental and
the control interventions.
2. Typically the number of pair of interventions varies from two to seven.
3. The number of interventions is not pre specified so that the clinician and the
patient can decide to stop at will.
