Lecture-3 - Three - Study Design Select
Lecture-3 - Three - Study Design Select
1
Learning objectives
2
A study design
Specific plan or protocol for conducting the study
Allows the investigator to translate the conceptual hypothesis
into an operational one.
The research design of a study spells out the strategies that
researchers adopt to answer their questions and test their
hypotheses
Epidemiologic Study Designs
• Descriptive studies
• Analytical studies
• Experimental studies
Descriptive studies
Can help to give perspective to the burden of disease and
may assist in planning services.
Examine patterns of disease
Aims
To measure the importance of, and monitor changes in,
diseases in a community
To describe the frequency of different diseases within a
community
Study Design Details: Descriptive studies Methods
• Case report
• Case series
• Ecological
• Cross sectional
2.1 Case report and case series
• The infant's T cell count was low, suggesting AIDS. There was no
family history of immunodeficiency, but one of the blood donors
was found to have died of AIDS. This led the investigators to
hypothesize that AIDS could be transmitted by blood transfusion.
Example of case series
• In 1974, Creech and Johnson reported a case series
of three men with angiosarcoma of the liver among
workers at a vinyl chloride plant.
• This number in such a small population during the
time period studied was clearly in excess of what
was expected
• this led to the formulation of the hypothesis that
occupational exposure to vinyl chloride caused
hepatic angiosarcoma.
• Later the same year, this hypothesis was
substantiated by data from two analytic studies
Example 2 of case series
• In 1980 -1981 four previously healthy young men were diagnosed with Pneumocystis
carinii pneumonia, an unusual "opportunistic" infection that had only been seen in
immune compromised people with hereditary disorders or in people with immune
compromise due to chemotherapy. The medical histories didn't suggest any preexisting
immunodeficiency, but all four men had decreased immune responses and low T cell
counts.
35 Finland
30
25 Sweden
2
Norway R = 0.53
20
UK US Denmark
15
Netherlands
10 NZ Canada Austria
5 Australia France
Japan
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Average annual coffee consumption (kg/person)
1978-1982
Strengths and Limitations of correlational studies
Strength
can be done quickly and inexpensively, often
using already available information
Limitation
1) Since the unit of analysis is a population or
group, the individual link between exposure
and effect cannot be made
Cont’d…..Limitations of correlational studies
2) lack of ability to control for the effects of
potential confounding factors
3) correlational data represent average
exposure levels rather than actual individual
values.
This may lead to - Ecological fallacy- the
association observed between variables at
the group level does not necessarily
represent the association that exists at
individual level
Analytic Studies
Aims
To test hypotheses on disease causation by showing the degree
of correlation between possible determinants and disease
To assist health service planning by measuring the burden of
disease in subgroups and identifying those in greatest need of
services
Cohort studies
29
What is exposure of interest?
Exposure in usual sense
E.g. Ingestion of contaminated food
E.g. Droplets from someone with active pulmonary
tuberculosis
Behaviors
E.g. Sharing needles, drinking alcohol, fatty food eating
habit etc
Treatment
E.g. Intervention - education program
30
What is outcome of interest?
Disease
E.g. Diabetes, TB,
Event
E.g. Injury from land mine
Condition
E.g. Blindness
Death
Others
31
Sources of exposure and outcome information
Pre-existing records
Information supplied by the study subjects
Direct physical examination or screening tests
32
Possible outcomes in cohort studies
No disease
Disease
Death from the computing risk
Lost to follow up
33
Follow up period of cohort studies
The follow-up is the most critical and demanding part of a cohort study
Changes in the level of exposure to key risk factors, after the initial survey
and during the follow-up period, are a potentially important source of
random bias
34
Follow up time in cohort studies
The outcome has not occurred at the beginning of the study
Length of follow-up time required is dependent on:
the incidence rate of the outcome
the size of the population at risk
the latency period of the outcome
35
Types of cohort studies
36
Cohort study design
Classical (prospective)
Historical (retrospective)
38
Retrospective cohort studies
39
Ascertainment of outcome of interest
The aim of good case ascertainment is to ensure that the
process of finding cases, whether deaths, illness episodes, or
people with a characteristic, is as complete as possible using:
Operational definition
Standard definition
Eligibility criteria
Case definition
Case finding
Case ascertainment
40
Strength of cohort studies:
Particularly efficient when exposure is rare
Can examine multiple effects of a single exposure
e.g. (smoking > lung cancer, COPD, larynx cancer)
Minimize bias in outcome measurement if prospective
Allows direct measurement of incidence (risk) and prognosis
(natural history)
Can elucidate temporal relationship between exposure and
outcome of interest (if prospective )
41
Limitation of cohort studies:
42
Case-Control Studies
43
Case-control studies
Direction of inquiry
Exposed
Cases
Non-exposed
Population
Exposed
Controls
Non-exposed
Time
44
Steps of case-control studies
1. Define cases and controls
2. Identify group of cases
3. Identify group of controls
4. Review both groups for previous history of exposure to risk
factors under study
5. Measure frequency of exposure to risk factors occurrence in
both groups
6. Compare frequency of exposure to risk factors between cases
and controls
7. Conclude that previous history of exposure to risk factors
contributed for the cases more than controls or not
46
Applications of case-control studies
47
What is case?
Establish a clear operational definition or use standard definition of
disease (outcome) of interest in order to have a clear
understanding of exposure-disease association
It is the outcome of interest under study
It can be:
A disease
E.g. HIV status, malaria case
A behavior
E.g. Alcohol drinking habit, cigarette smoking
Occurrence of an event
E.g. migration
48
Selection of case …
Define ‘disease’ and how it will be measured
Selecting the source population of cases (homogeneous cases)
Sources of cases are commonly:
49
Sources of cases
Hospital-based:
Easy and in-expensive to conduct
It is prone for selection bias
Population-based:
Avoids selection bias
Allows the description of a disease in the entire population
and the direct computation of rates of disease in exposed
and non-exposed groups
50
What is control?
It is the comparison group (referent)
It should be free of the disease (outcome of interest under
study)
It should be as similar as the cases in all aspects except for the
disease of interest under study
Controls must have the same opportunity of getting exposure to
risk factors as cases and should be subjected to the same
inclusion and exclusion criteria
51
Sources of controls
1. Population-based controls
2. Hospital-based (health institution) controls
3. Specials controls: neighbourhood, friends, spouses or
relatives (siblings)
o Potential bias regarding ethnic backgrounds,
socioeconomic status and environmental exposures
E.g. If diet and smoking as exposure factor will
underestimate the effect of exposure to the occurrence
of disease
52
Hospital-based controls
Advantages:
Minimize re-call bias
Convenient
Cooperative non-case patients (minimize non-response
bias)
Disadvantages:
Control disease may be linked to exposure of interest
Hospitalized controls differ from general population
53
Ratio of controls to cases
A single control group is optimal in most of the times
However, ratio of controls to cases may vary from 1:1
to 4:1
Trade-off: cost versus power of the study
Decision based on power calculation
More than one control groups may be recommended
when the single control is not appropriate or has a
specific deficiency
54
Strengths of case-control studies
55
Limitation of case-control studies
In-efficient for rare exposures
No calculation of rates and risks possible
In some situations, the temporal relationship between
exposure and disease may be difficult to establish
Temporal exposure–disease uncertainty
Prone to selection and information bias
Selection of controls difficult some times
56
Interventional Studies
57
Monitoring and Evaluation: Evaluation Designs
Objectives of the Session
Experimental group O1 X O2
RA
Control group O3 O4
Experimental group X O2
RA
Control group O4
A Experimental Design: Example
• Health providers in a Native Nation want to find a method for
raising awareness about HIV and risky behaviors in the
adolescents in the community. They have identified two
programs that can be delivered in the classroom; however,
there is a considerable difference in the cost of the two
programs in terms of not only materials, but also staff time.
Before making a final recommendation, they decide to pilot
each program in a different school setting. Since the goals and
objectives of both programs were there, both groups receive
the same post-test at the conclusion of the program year.
Changes in the KABB of the students can be attributed to their
respective program.
A Non-Experimental Design
Time
Experimental group
O1 X O2
Only people who are participating in program get pre- and post-test.
Steps
1. Pre-test everyone in the program
2. Deliver the intervention
3. Post-test the same individuals
• This design does not provide any information about what kinds of results
might have occurred without the program and is the weakest in terms of
scientific rigor
Another Factor that May Lead to Invalid Conclusions
Time
Experimental groupO1 O2 O3 X O4 O5 O6
Time
Experimental group O1 X O2
---------------------------------
Comparison group O3 O4
• A nurse working with women in an OB/GYN clinic wanted to see if her HIV-
education programs were increasing awareness among pregnant women
and new mothers.
• She asked a colleague working with women at another OB/GYN clinic to
help her. Women in their second trimester of pregnancy in both clinics were
asked to complete a pre-test. Visits by women in the experimental clinic
also included regular HIV-education messages, delivered during their
routine clinic visits.
• At the post-partum visit, six weeks after the birth of their child, participants
at both clinics completed the post-test. If the intervention proved
successful, the information could be used to train other nurses to deliver
prevention messages during the regular course of their work.
Threat to Validity
79
Thank You !
80