This document discusses experimental study designs in epidemiology. It defines experimental studies as those where the investigator manipulates the exposure and randomly assigns subjects to treatment groups. Randomized controlled trials are considered the gold standard. The key features of experimental studies are randomization, manipulation of the exposure, and use of a control group. Some examples of experimental designs discussed include randomized controlled trials, non-randomized controlled trials, pre-post studies, factorial designs, and crossover studies. The document also discusses how experimental studies can provide stronger evidence of causality than observational studies by minimizing bias and variation between groups.
2. Introduction
• A study design is a specific plan or protocol for conducting the
study, which allows the investigator to translate the
conceptual hypothesis into an operational one.
• A major goal of epidemiological research is to explain patterns
of disease occurrence and causation (etiology).
• Epid’l measurements are aimed at quantifying 3 things:
exposures, confounders & outcomes.
• Once quantified, the association between exposure and
outcome is the central focus of epid’l studies.
• The different study designs are ways of evaluating the
association between an exposure and an outcome
3.
4.
5. Experimental study
designs/intervention studies
• A study in which a population is selected for a planned trial of
a regimen, whose effects are measured by comparing the
outcome of the regimen in the experimental group versus the
outcome of another regimen in the control group.
• Such designs are differentiated from observational designs by
the fact that there is manipulation of the study factor
(exposure), and randomization (random allocation) of subjects
to treatment (exposure) groups.
• RCTs represent the “gold standard” of research designs.
They thus provide the most convincing evidence of
relationship between exposure and effect.
6. Hierarchy of Epi Design Strategies
Case Reports
Case Series
Cross-Sectional Surveys
Case-Control Studies
Cohort (Follow-Up) Studies
Randomized Controlled Trials
Complexity Confidence
7. Benefits of experimental studies
• Provide stronger evidence of the effect (outcome)
compared to observational designs, with maximum
confidence and assurance.
• Yield more valid results, as variation is minimized and
bias controlled.
• Determine whether experimental treatments are safe
and effective under “controlled environments” (as
opposed to “natural settings” in observational designs),
especially when the margin of expected benefit is
doubtful / narrow (10 - 30%).
8. time
Study begins here (baseline point)
Study
population
Intervention
Control
outcome
no outcome
outcome
no outcome
baseline
future
RANDOMIZATION
9.
10. Types of Experimental study designs
• Randomized Control Trial
In an RCT, a group of participants fulfilling certain inclusion
and exclusion criteria is “randomly” assigned to two
separate groups, each receiving a different intervention.
Random assignment implies that each participant has an
equal chance of being allocated to the two groups.
The use of randomization is a major distinguishing feature
and strength of this study design. A well-implemented
randomization procedure is expected to result in two
groups that are comparable
11.
12. Types of Experimental study designs
• Nonrandomized controlled clinical trials
In this design, participants are assigned to different
intervention arms without following a “random” procedure.
For instance, this may be based on the investigator's
convenience or whether the participant can afford a particular
drug or not.
Although such a design can suggest a possible relationship
between the intervention and the outcome, it is susceptible to
bias – with patients in the two groups being potentially
dissimilar – and hence validity of the results obtained is low
13. Types of Experimental study designs
• Interventional studies without concurrent controls
When a new intervention, e.g., a new drug, becomes
available, it is possible to a researcher to assign a group
of persons to receive it and compare the outcome in
them to that in a similar group of persons followed up in
the past without this treatment (”historical controls”).
This is liable to a high risk of bias, e.g., through
differences in the severity of disease or other factors in
the two groups or through improvement over time in the
available supportive care
14. Types of Experimental study designs
• Before–after (pre–post) studies
In this design, a variable of interest is measured before
and after an intervention in the same participants.
Examples include measurement of glycated hemoglobin
of a group of persons before and after administration of a
new drug (in a particular dose schedule and at a
particular time in relation to it) OR number of traffic
accident deaths in a city before and after implementation
of a policy of mandatory helmet use for two-wheeler
drivers.
15. Types of Experimental study designs
• Factorial study design
If two (or more) interventions are available for a particular disease
condition, the relevant question is not only whether each drug is
efficacious but also whether a combination of the two is more
efficacious than either of them alone.
The simplest factorial design is a 2 × 2 factorial design. Let us think
of two interventions: A and B. The participants are randomly
allocated to one of four combinations of these interventions – A
alone, B alone, both A and B, and neither A nor B (control). This
design allows (i) comparison of each intervention with the control
group, (ii) comparison of the two interventions with each other,
and (iii) investigation of possible interactions between the two
treatments (whether the effect of the combination differs from the
sum of effects of A and B when given separately).
16. Types of Experimental study designs
• Crossover study design
This is a special type of interventional study design, in
which study participants intentionally “crossover” to the
other intervention arm.
Each participant first receives one intervention (usually by
random allocation, as described above). At the end of this
“ first” intervention, each participant is switched over to
the other intervention. Most often, the two interventions
are separated by a washout period to get rid of the effect
of the first intervention and to allow each participant to
return to the baseline state.
17. Crossover study design
For example, in a recent study, obese participants underwent two
5-day inpatient stays – with a 1-month washout period between
them, during which they consumed a smoothie containing 48-g
walnuts or a macronutrient-matched placebo smoothie without
nuts and underwent measurement of several blood analytes,
hemodynamics, and gut microbiota.
This design has the advantages of (i) each participant serving as
his/her own control, thereby reducing the effect of interindividual
variability, and (ii) needing fewer participants than a parallel-arm
RCT.
However, this design can be used only for disease conditions
which are stable and cannot be cured, and where interventions
provide only transient relief
18. Features of a good experimental study
• Randomization (in placing study subjects in of
intervention/treatment or control group)
• Manipulation (intervention)
• Blinding (may be single, double, or triple
blinded study)
• Inclusion of a control group (no control, no
conclusion – NCNC)
19. Causation and Association in
epidemiologic studies
• When considering the relationship between
exposures and health outcomes, it is important
to distinguish between association and
causation.
• Epidemiologists ultimately want to be able to
draw conclusions about causation, but most
epidemiologic studies focus on establishing
associations
20. Association
• The concurrence of two variables more often
than would be expected by chance.
• Is a specified health outcome more likely in
people with a particular "exposure"? Is there a
link? Association is a statistical relationship
between two variables
21. Causation and Association in
epidemiologic studies
• Types of Association
– Spurious (Artifactual)
– Indirect
– Direct (causal) association
• One-to-one causal association
• Multifactorial causal association
22. Example….
• A researcher in his observational study found the
presence of Helicobacter pylori in patients of
duodenal ulcer!
• Can we say that
– H.pylori causes duodenal ulcers?
• Hypothesize that
– H.pylori may have a role in etiology of duodenal ulcers.
• For final proof there has to be a ‘comparison’.
Comparison would generate another summary
measure whic h shows the extent of ‘Association’ or
‘Effect’ or ‘risk
23. Association
• Bias and Confounding
• If an association is observed, the first question
asked must always be …
“Is it real?”
24. Cause: definition
• A factor is a cause of an event if its operation increase the
frequency of the event
• Types of causal relationships:
– Necessary – the outcome occurs only if the causal factor has
operated
– Sufficient – the operation of the causal factor always results in
the outcome
– Both (necessary and sufficient) – the causal factor & the
outcome have a fixed relationship, neither occurs without the
other
– Neither (i.e. neither necessary nor sufficient) – the operation of
the causal factor increase the freq. of the outcome; but the
outcome does not always result; and the outcome can occur
without the operation of the causal factor.
25. Causation
• Causation means that the exposure produces the effect. It
can be the presence of an adverse exposure, e.g., increased
risks from working in a coal mine, using illicit drugs, or
breathing in second hand smoke.
• Causative factors can also be the absence of a preventive
exposure, such as not wearing a seatbelt or not exercising.
• To conclude that lack of exercise is a cause of heart disease,
one needs to review the body of evidence suggesting a
causal relationship and also consider other criteria
26. From Association to Causation
Bias in selection or
measurement
Confounding
Hill’s criteria for Causality
Association
Yes No
Likely Unlikely
Chance
Yes No
Yes
27. Sir Austin Bradford Hill,1965
In what
circumstances
can we pass from
[an] observed
association to a
verdict of
causation? Upon
what basis should
we proceed todo
so?
28. Guidelines for judging whether an association is
causal
Sir Austin Bradford Hill criteria
Most Important criteria
1. Temporality: cause precedes effect.
2. Strength of association: large relative risk.
3. Consistency: repeatedly observed by different.
persons, in different places, circumstances, and times
29. Guidelines for judging whether an association is
causal
4. Biological gradient (dose response): larger
exposures to cause associated with higher
rates of disease. And reduction in exposure is
followed by lower rates of disease
(reversibility).
5. Biological plausibility: makes sense, according
to biologic knowledge of the time.
6. Experimental evidence
30. Evidence for a causal relationship:
The Henle-Koch Postulates
• The organism is always found with the disease
• The organism is not found with any other
disease
• The organism, isolated from one who has the
disease, and cultured through several
generations, produces the disease (in
experimental animals)
31. Causal Relationship in the context of
medicine & public health
• A causal relationship would be recognized to
exist whenever evidence indicates that the
factors form part of the complex of
circumstances that increases the probability of
the occurrence of disease and that a
diminution of one or more of these factors
decreases the frequency of the disease
32. Judgment of a cause-effect
relationship
• 1. Temporal relationship (time sequence)
• 2. Strength of association
• 3. Biologic credibility (plausibility)
• 4. Consistency (replication) of the findings
• 5. Dose-response relationship
• 6. Reversibility (cessation of exposure)
• 7. Specificity of the association
• 8. Consideration for alternate explanations
• 9. Type of study design providing evidence
34. Factors in Causation
• Predisposing factors e.g. age, sex, and previous illness
may create a state of susceptibility to a disease agent
• Enabling factors e.g. low income, poor nutrition,
• inadequate medical care
• Precipitating factors e.g. exposure to a specific
• disease agent or noxious agents may be associated
• with the onset of disease
• Reinforcing factors e.g. repeated exposure and unduly
hard work may aggravate an already established
disease or state