RANDOMIZED CONTROLLED TRIALS

Dr.S.Babyvanitha
Professor
Department of Pharmacology
St.James College of Pharmaceutical Sciences.
DEFINITION
• Clinical trial is research in humans expected to reveal or confirm the clinical,
pharmacological, and or other pharmacodynamics effects of a research products and
or to determine any adverse reactions to a research products, and or to study
absorption, distribution, metabolism and excretion of a research products with the
aim to discover its safety and or efficacy.
• RCT provides highest level of evidence
Procedure:
Formulate a specific research question to be answered
Decision on the method of selecting the subjects
Ensure the maximum possible number of the subjects during the study period.
After sampling, randomization is required in order to divide the subjects into at least
two investigational group (Exposed and unexposed or control)
Unexposed group receive either placebo if it is a healthy subject, or a gold standard if
it is a subject with illness.
RANDOMIZATION
• Process that allows unpredictability of the distribution of subjects in the exposed
group and group of unexposed ones.
• Allows us to eliminated the effect of confounding factors.
• Achieve comparability of a group of subjects with respect to certain significant
characteristics of the subjects
• Can be accomplished in many ways: using random number tables, using sealed
envelopes, drawing red or green balls and today it is easier to do it with the
computer using statistical programs.
• Stratified randomization approach ensures complete comparability of subjects in
groups according to very important confounding factors such as age and sex.
MONITORING AND DATA COLLECTION

• Data collection must follow a predefined protocol and must be performed equally
well in both groups of subjects in order to avoid the occurrence of a
discontinuation of the diagnostic procedure.
• Ensured by using blinding, masking which can be single, double or triple
• Single blinding- neither the subjects nor the researchers who collect the data
during the follow-up period of the subjects do not know who from the subjects is
in which group.
• Triple-blinding neither the subjects, the researcher nor the person analysing the
data do not know who from the subjects is in which group.
• The blinding can only be done if placebo, or gold standard drug are of the same
appearance (Color, shape) weight, taste and odor as the drug being investigated.
• It is possible to have a placebo effect when the ineffective substance also
improves health status as well as the appearance of Hawthorne’s effect, which is
marked by the positive reaction of subjects because researchers care about them.
MEASUREMENT OF HEALTH OUTCOMES
• Depends on the research question and the disease we are investigating
• The outcomes are divided into primary and secondary
• The primary health outcome in research is usually one and that is the answer to
the most important part of the research question.
• The secondary outcome may be drug side effects but also recurrence of the
disease, functional impairment, disability etc.
TYPES OF RCT DESIGN

• Besides most commonly used design where subjects are randomized to one of two
or more arms and where each arm allocated different treatment so called parallel
design.
• Other possible approaches in research in experimental epidemiology is the use of
so-called cross - over design or the use of factorial design.
• Basic principles of cross-over type of design is to put the subjects in the groups by
randomization, application of the investigated drug in the exposed group and
placebo/gold standard drug in the unexposed group and follow-up of the health
outcomes over time.
• What is specific to this approach is the change in exposure in the investigated
groups and the exposed group in the second part of the research becomes
unexposed and vice versa.
• This change in exposure and non-exposure status of each subject can be
accomplished after a certain amount of time has passed required for the previously
applied substance to be completely excreted from the body.
• In this design every subject is control to itself.
Factorial design:
We simultaneously investigate the effect of multiple drugs.
These drugs have different pharmacokinetic and activity must be independent.
Presentation of results:
We can calculate the relative risk for health outcomes using the
incidence in the exposed and non-exposed group of subjects.
We can show survival of the subjects in the groups for which we can use
Kaplan-Meier survival curve which shows the proportion of survivors
during the follow-up time.
For the analysis of survival, we can use the multivariate analysis of
survival - Cox regression model and as a result we get the hazard ratio.
Unlike Kaplan-meier’s method, Cox-regression takes into account the
effect of the confounding factors.
PRESENTATION OF RESULTS

Finally we can calculate efficacy;

Efficacy = incidence in unexposed group – incidence in exposed group/ incidence
in unexposed group x 100
Exposed group is the group that received the investigated drug, while the unexposed
group is the one who received a placebo or gold standard drug.
We can calculate the number of people you need to treat in certain way in order to
prevent one undesirable outcome (Number need to treat – NNT)
NNT = 1/ incidence in unexposed group – incidence in exposed group.