M9: Introduction to Experimental Study

MT BIOEPI LEC BIOSTATISTICS AND EPIDEMIOLOGY (LEC) MS. ARLENE JOY M. CANASA, RMT

INTRODUCTION TO EXPERIMENTAL STUDY CLASSIFICATION SCHEMES OF EXPERIMENTAL

STUDIES
★ EXPERIMENTAL STUDY
● Also known as intervention study. 1. ACCORDING TO TYPE OF EXPOSURE / INTERVENTION
● A prospective study comparing the effects and
THERAPEUTIC PROPHYLACTIC
value of intervention/s against the control of
Agent or procedure is Agent is administered or a
human beings.
given in an attempt to procedure is performed to
● It provides the strongest link between the
relieve the symptoms determine its effect on
relationship of an exposure and outcome among
and/or improve the preventing disease.
the analytic designs.
survivorship of those with
a. Prospective Nature of Inquiry
the disease.
b. Uses One or More Intervention Techniques
c. Use of Control Group
Population: Persons with Population: Persons without
d. Conduct of study in humans.
the disease. the disease.
KEY ASPECTS OF AN EXPERIMENTAL STUDY
AZT (azidothymidine) Hepatitis B vaccination for
A. The participants are followed forward
treatment for AIDS. prevention of hepatitis and
Prospective in time.
hepatocellular carcinoma.
Nature of
Inquiry
2. ACCORDING TO THE UNIT IN WHICH TREATMENT IS
ASSIGNED.

Study participants are subjected to a CLINICAL Treatment is assigned to individuals in

single or a combination of TRIALS clinical or hospital settings.

B. intervention techniques.
Simple mastectomy for breast cancer.
Uses One or
More FIELD Treatment is assigned to individuals in
Intervention TRIALS community settings.
Techniques
Field trial of COVID-19 from UK, US
Pharmaceutical Companies to be
conducted in Cavite.

Must be sufficiently similar and

COMMUNITY Treatment is assigned to a group of
comparable in relevant aspects to the
TRIALS individuals in community settings.
C. intervention group.
Use of Control Evaluation of fluoridation of community
a. Experimental Group water supply for preventing dental decay.
Group
- Has an assigned exposure.
b. Control Group 3. ACCORDING TO THE SETTING IN WHICH EXPOSURE IS
MANIPULATED.
- Has no assigned exposure.
EFFICACY ● Demonstrates how an intervention
TRIALS works in ideal conditions.
● Does not allow drop-outs and
non-compliance.
● Less generalizable to individuals
outside the study or reference
population.

1
 Studies in which the exposure is DESIGNING A RANDOMIZED CONTROLLED
TRIAL
manipulated in a contrived, artificial
setting. 1. SELECT A SAMPLE FROM THE POPULATION.

Reference ● Group of ultimate interest.

EFFECTIVENESS ● Demonstrates how an intervention
Population ● The groups in which study
TRIALS works in practice (i.e. “real-world”
results will be applied.
scenario).
● Drop-outs are taken into Study ● Group that is studied or
consideration. Population observed.
● More generalizable to the ● Should be identical to the
reference population. reference population.

Studies in which the exposure is To avoid biases, the reference and

manipulated on test units in their study population must be equal in
natural setting. terms of different variables such
as:
4. ACCORDING TO TYPE OF COMPARISON ➔ Age
BETWEEN SUBJECT DESIGN WITHIN SUBJECT DESIGN ➔ Gender Distribution
➔ Similar Cases / Diseases
● Compare outcomes ● Effects of the
observed in two or experimental Assembling the Draw ip criteria of eligibility for the
more groups of intervention and the Study Cohort trials as to identify a group of people
subjects/patients who non-experimental or patients in which the influence of
receive different intervention on the various extraneous factors can be
interventions. outcome compared avoided:
within one person.
● All subjects experience Inclusion Criteria

all treatments in ● Appropriate with the

succession, though the research problem/question.

order varies.
Exclusion Criteria
● Entails repeated
● Helps in controlling errors.
measurements of the
● Shall identify the risk or
dependent variable in
factors that will give errors in
each subject.
the experimental study.

PARALLEL CONTROLLED CROSS-OVER TRIAL ● Exclusion of other risks to

TRIAL prevent bias or errors.

Type of clinical study in Type of clinical study where 2. MEASURE BASELINE VARIABLES.
which two or more groups participants receive
Purpose It is essential that all the relevant
of participants receive multiple treatments in a
baseline information is collected for
different intervention or specific sequence. each subject:
treatment. They are both the control
a. To confirm comparability of the
and experimental group.
groups.
b. To enable a more detailed
examination of the results of the trial
in a search for confounding factors
and their interaction with the
treatment under test.

2
 Baseline A type of independent variable that is Blinding Whenever possible, the investigator
Variable usually used as a starting point or should design interventions in such a
reference for comparison. fashion that neither the subjects nor
anybody who has contact with them,
An information collected at the beginning
has any knowledge of the group
of a study that is used to assess change
assignment.
brought by a particular intervention.
Blinding can protect the study from:
● Understand what factors can affect
✿ ● Confounding variables that
the specific variables being
investigated upon the study. arise during follow-up.
● Be more specific to the hallmarks and ● Observation Bias
factors, but at the same time,
maintain uniformity in getting the All subjects and other contacts have
reference and study population.
no idea who among the participants
received the treatment nor placebo.
3. RANDOMIZE
Open Label There is no blinding employed.
● Hallmark or sine qua non of a true
✿
experiment. Single Blind Either the subject or observer is

● Refers to allowing chance to unaware of treatment being received

determine the assignment or but the researcher has this

allocation of subjects into either information.

the experimental or control group.

Double Blind Neither the subject nor the observer is
● Ensures that age, sex, and other
aware of the treatment being
baseline characteristics that could
received.
confound an observed association
will be distributed equally Triple Blind The patient, researcher, and data
between the randomized groups. analyst are unaware of the treatment
being received.
4. & 5. APPLY INTERVENTIONS AND FOLLOW-UP ● Only the pharma-companies
COHORT who develop the treatment
5. a

know.
Unintended This can occur when:
● After clinical trials, that is the
Intervention ● Investigator gives extra attention
only time they will reveal the
Co-Intervention to patients receiving the active
experimental and control
drug.
groups.
● Subjects who discover that they
are receiving placebo seek out
6. MEASURE OUTLINE VARIABLES
other treatments. 7.

ANALYZING THE RESULTS

This creates a bias among the
✿ Qualitative ● Test for Two Independent
subjects that creates a problem and
causes erroneous results due to Variables Proportions
resignation of the subject.
- Ex.) Z-Test
Prevention Apply blinding. ● Chi-Square Test for Homogeneity /
Association
● FIsher's Exact Test

3
Quantitative 1. Test for Two Independent Sample C. TESTING EFFECT
Variables Means ● Occurs when the mere taking of
✿
● Compare two different means pre-test measurement sensitizes
from unrelated groups (control subjects.
and treatment group). ● Therapeutic effect of the drug may
2. Analysis of Variance coincide with improvement from the
● Used to compare three or more natural course of the disease.
different means.
3. Kruskal–Wallis Test Happens when during the pre-tests, the
subject already identifies that nothing is
Treatment happening or they are already feeling
Effect sick leading to curiosity which may cause
them to resign in the study.

Possible Solomon Four-Group Design

ISSUES AND THREATS TO INTERNAL VALIDITY Solutions ● A type of experiment where
participants get randomly
A. ETHICAL CONCERNS
assigned to either one of four
● Potential harm for those receiving the groups that differs whether
✿
participants receive the treatment
experimental intervention.
or not, and whether the outcome
● Control group might be denied or of interest is measured once or
deprived of potentially-beneficial twice in each group.
intervention.

Possible 1. Use of standard treatment/care in

Solutions the control group.
● Use a substitute medicine or
intervention found beneficial for
the disease being studied.
2. Interim Analysis D. PLACEBO EFFECT

● Check for errors, effects, and Refers to the beneficial effect resulting
✿
conditions of the subjects. solely from administration of the
intervention and not strictly through the
B. ATTRITION EFFECT
intervention’s mechanism of action.
● Occurs when the subject
✿
characteristics that discontinue the It does not necessarily mean that those in

study are markedly different the control group will not receive any

between the groups compared, treatment at all instead they will not just

● It is a form of selection bias. receive the treatment being studied.

Possible Obtain baseline measurements. Possible Blinding

Solutions ● Allows the evaluation of the Solution ● The participants and observers

degree or extent of bias do not have an idea what

introduced by attrition. treatment they are receiving.

● The control and the treatment

group should be similar in all
aspects except for the medicine
received.

4
 ADVANTAGES AND DISADVANTAGES OF
INTERVENTION STUDIES

ADVANTAGES 1. Experiments can produce strongest

evidence for cause and effect.
2. Experiments can be the only
possible design for some research
questions.
3. Experiments can sometimes
produce faster and cheaper
answers to research questions
than observational studies.

DISADVANTAGES 1. Experiments are often costly in

time and money.
2. Many research questions are not
suitable for experimental designs.
3. Standardized interventions may be
different from common practice
which can reduce generalizability.
4. Experiments tend to restrict the
scope and narrow the study
question.

If you’re scared, you do it scared.

TONGOL, LANA LAINE N. | MT 2D