Sunteți pe pagina 1din 46

Epidemiological

Study Design
Prof. Bhisma Murti

Masters Program in Public Health


Postgraduate Program
Universitas Sebelas Maret
Research Problem
• Previous belief:
Garlic consumption is the key to good health

• Research question:
Does garlic reduce the incidence of coronary heart
(artery) disease?

• Hypothesis
Garlic intake decreases the risk of coronary artery
disease (CAD)
Study Design

• Descriptive study
• Analytic study
Descriptive Studies

• Individual level:
• Cross-sectional studies
• Case reports

• Case series

• Population level:
• Ecologic studies
Descriptive Studies: Aims
1. Describe the health status of a
population
2. To assess the public health
importance of diseases
3. To describe the natural history of
disease
4. To describe other study variables
including exposure status
Descriptive Studies: Uses

• Planning health care


program

• Hypothesis generating
suggesting associations
Analytical Studies

• Observational

• Experimental
• aka. Intervention study
Observational Studies

• Individual level:
• Cross-sectional
• Case-control
• Cohort
• Population level:
• Ecologic studies (bias known as “ecologic fallacy”)
Cross-Sectional Study
• Exposure status and disease
status are measured at the
same point of time
• Data collected at a single
point in time (point
prevalence)
• Data can also be collected
over a time period (period
prevalence)
• Prevalence data can be used
for health program planning
• Generate hypothesis on
associations to be tested in
analytic studies A “Snapshot
Prevalence vs. Incidence
• Prevalence
• Exposure status and disease status are measured at
the same point of time
• Generates the total number of cases at a point in
time
• Includes both new and old cases
• Data can be gathered in a period of time (period
prevalence)

• Incidence
• Exposure status is measured initially among disease-
free study subjects, then after a period of time
disease status is measured
• Generates the number of new cases over time
• Investigator must follow study subjects and identify
each developing case
Example of a Cross-Sectional
Study

Association between garlic

consumption and CAD in

the Family Practice Clinic


Cross-Sectional Study

Sample of Population

Garlic Eaters Non-Garlic Eaters

Prevalence of CAD Prevalence of CAD

Time Frame = Present


Cross-Sectional Study

Garlic
+
Consumption
-

C + 10 90
A
D

- 90 10
Cross-Sectional Study
• Strengths
• Quick
• Cheap

• Weaknesses
• Cannot establish cause-effect
because temporal sequence
cannot be ascertained
Observational Studies
Case-Control Study
• Begin with determining subjects on
the basis of disease status
• Start with people who have disease
• Match them with controls that do
not
• Look back and assess exposures
Case-Control Study Design
Exposure Disease Observer

Choose groups with and without disease,


look back at what different exposures they
may have had
Case Control Study
(Retrospective)
Study
Exposure subjects
The

? Disease Investigator’s
Starting
? Point of
Controls Observation

Retrospective nature
Case Control Study
(Prospective) Study
subjects
Exposure
The Disease
Investigator’ ?
s Starting ?
Point of Controls
Observation
Retrospective nature

Time of
Observation
Case-Control Study
Cases
High Garlic Diet

Subjects with CAD


Low Garlic Diet

Controls
High Garlic Diet

Subjects w/o CAD


Low Garlic Diet

Past Present
Case-Control Study
• Research Question:
• Are those with CAD less likely to have
consumed garlic?

• This question leads to use of odd

• Rephrased question:

• Do the odds of consuming garlic differ between


those with and without CAD?
Odds vs. Risk
• Odd is a concept different from risk
• Odd is the ratio of probabilities of two contradicting
events (e.g. disease/ well, alive/ dead)
• Whereas risk is the probability of the event per se
• Odd is used in a case control study instead of risk
because it is not possible to measure risk (incidence) in
a case control study
• Odds are generally greater than risk, but this
difference will lessen as the risk is smaller
Case-Control Studies:
Strengths
• Good for rare diseases (outcomes): e.g. cancer
• Investigator can just use all the existing rare disease cases
and select control subjects from the same source
population
• Can examine multiple exposures for a single disease
• Useful to generate hypothesis to be tested further in a
methodologically more rigorous study design, i.e. cohort
or experimental studies
• Fast
• Cheap
• Provides Odds Ratio
Case-Control Studies:
Weaknesses
• Cannot measure
• Incidence
• Prevalence
• Risk Ratio (aka. Relative Risk)

• Can only study one outcome


• High susceptibility to bias
• Particularly selection bias in retrospective case control studies
when determining disease status of the subjects is influenced by
exposure status
• Recall bias also possible
Cohort
Studies
following up subjects
towards outcomes (disease
status)
Example of a Cohort Study

Does garlic use reduce


CAD?
Example of a Cohort Study

Population Population
#1 #2
Eat garlic Don’t eat garlic

Disease
(Outcome):
Does the risk of having
coronary artery disease
(CAD) differ?
Cohort Study

• Begin with disease-free patients

• Classify patients as exposed/unexposed

• Record outcomes in both groups

• Compare outcomes using Risk Ratio (aka.

Relative Risk)
Cohort Study Design
(Prospective)
Exposure Observer
Disease

Start with two groups of people who


are exposed and unexposed, follow
them to see who gets disease.
Prospective Cohort Study
Disease
Exposure Study starts occurrence

time

Disease
Study starts Exposure occurrence

time
Cohort Study Design
(Retrospective)
Exposure Disease Observer

Start with two groups of people who


are exposed and unexposed, find out
who got the disease.
Retrospective Cohort
Studies
Disease
Exposure occurrence Study starts

time
Prospective Cohort Study
CAD
Garlic Free
No
CAD

CAD
Garlic Eaters
No CAD

Present Future
Cohort Study: An
Example
Do the infants of mothers with
good nutritional status have better
outcomes at one year of age than
infants of mothers with poor
nutritional status?
Cohort Study
Mother’s
nutritional
status:
Survival of
child to
Good one year?

Survival of
child to
Poor
one year?
Cohort Study: Strengths
• Provides incidence data

• Establishes time sequence for causality

• Eliminates recall bias

• Allows for accurate measurement of exposure


variables
Cohort Study: Strengths

• Can measure multiple outcomes

• Can adjust for confounding variables

• Can calculate relative risk


Cohort Study: Weaknesses

• Expensive

• Time consuming

• Not suitable to study rare outcomes (since


sample size would be prohibitively large)

• Confounding variables
Cohort Study: Weaknesses

• Exposure may change over time

• Disease may have a long pre-clinical phase


(therefore time consuming)

• Attrition of study population (aka. loss-to


follow up bias)
Experimental Studies

Randomized control trials (RCT) provide the

“gold standard” of determining the causal

relationship between garlic and

cardiovascular disease prevention.


Example of an Experiment
Population #1 Population #2
(Treatment Group) (Control Group)

Ascorbic Acid Placebo

Outcomes:
Do the average number of colds differ?
Do their average lengths of colds differ?
Randomized Controlled Trial
• Randomized
• Control all known and unknown potential
confounding factors

• Double-blind
• Prevent information bias

• Placebo-controlled
• Control placebo effect
Randomized Controlled Trial

Treatment
Outcomes
Group

Study
Randomi
Population
zation

Control Group Outcomes


Randomized Controlled Trial

Garlic Pill No CAD

CAD

Study Random
Population ization

No CAD
Placebo

CAD
Randomized Controlled Trial
Strengths:
• Best measure of causal relationship
• Best design for controlling bias
• Can measure multiple outcomes

Weaknesses:
• High cost
• Ethical issues may be a problem
• Compliance
Analytical Studies:
Summary
Aspect Cross- Case- Cohort RCT
Sectional Control
Cost + ++ +++ ++++
Duration + ++ +++ +++
Sample Varies Small Large Varies
Size
Incidence, Prevalence None (Odd) Incidence Incidence
Prevalence
Multiple Yes No Yes Yes
Outcomes
Multiple Yes Yes No No
Exposure
Bias Prone Yes Yes Possible Possible
Causality Weak Moderate Strong Very
strong
Questions?
Thank you for your time and attention.

S-ar putea să vă placă și