Professional Documents
Culture Documents
4a-Observational Study Designs
4a-Observational Study Designs
2
Core Knowledge
• The evidence for evidence-based medicine is
all collected via research, which uses a variety
of study designs.
• Evidence-based medicine (EBM) is the process
of systematically reviewing, appraising and
using clinical research findings to aid the
delivery of optimum clinical care to patients.
3
Different study designs provide information of different
quality. Of course, we always try to use the best
possible design, but sometimes this is not practical or
ethically acceptable (you cannot do an experiment to
expose some people to a harmful substance to see
what effect it has). Therefore, you need to
understand the strengths and limitations of each type
of study design, as applied to a particular research
purpose. The purposes we will consider include
(1) describing the prevalence of health problems;
(2) identifying causes of health problems,
(3)evaluating therapy, including treatment and
prevention.
4
Did investigator assign exposures?
Yes No
Yes No Yes No
Exposure &
Outcome
Exposure to outcome Direction? at same time
Outcome to
exposure
Cohort study Case-control Cross-sectional
study study
5
Classification of types of Epidemiological research
Investigation Sequences
Analytic
Clinical trials
Epidemiology
Preventive Therapeutic
Case Cross
Cohort
Control Sectional
6
Case Reports
• Detailed presentations of a single case
7
Types of case Reports
• Previously undescribed syndrome or
disease
10
Case series
• Prevalence survey of a group of
individuals with a particular disease or
characteristic, performed at a single point
in time.
11
Example: Case Series
First 1000 cases of AIDS in USA
Complications % Risk Group %
• Pneumocystis 49.7 Homo- 72. 7
carini or bisexual
• Kaposi’s s 28.4 IV drug users
15.5
• PCP and KS 8.3 Haitian natives 5
• Other opport 13.6 Haemophilics
0.7
12
• In observational studies, the researcher observes and
systematically collects information, but does not try to
change the people (or animals, or reagents) being
observed.
16
Examples of observational studies:
17
Examples of experiments:
18
When do you do an observational study?
19
Observational Study Designs
By the end of this lesson the students of 4th year
MBBS should be able to:
• Describe the key characteristics of cross-sectional,
case-control and cohort studies regarding subject
selection, data collection and analysis.
• Identify the design of a particular study.
• Discuss the factors that determine when a
particular design is indicated.
• Discuss the strengths and limitations of each study
20
Lots of Observational study designs are there but you
need to know about 3 main ones:
Cross- sectional
surveys
Observational studies
Case-control studies
Cohort studies
21
Defined population
23
Example:
What is the prevalence of diabetes in this community?
Here, you draw a random sample of people and record
information about their health in a systematic manner.
You can also compare people with, and without, diabetes
in terms of characteristics (such as being overweight) that
may be associated with the disease. The problem is that
you cannot be sure which came first: the diabetes or the
weight problem, so this is a very weak design for drawing
conclusions about causes.
24
Analysis of Cross Sectional Studies
• Distribution of characteristics
• Frequency
• Means
• Proportions
• Prevalence
25
Main outcome is Prevalence. Using a 2 x2 table:
Diseas Non-
e diseased
Exposed
a b a+b
Non-
exposed
c d c+d
a+c b+d
27
Cross-Sectional Study
Disease Status
Yes No Total
Exposed A B A+B
Not Exposed C D C+D
A+C B+D N
c/ (c +
28
d)
Cross-Sectional Study
Example:
Hypothesis:
• Obesity is a risk
factor for knee
osteoarthritis.
Sample:
• 100 retirees living
at “University
Village”
29
Cross-Sectional Study
Medical exam + X-rays to diagnose
osteoarthritis of the knee
Osteoarthritis
Yes No Total
Yes 40 10 50
Obesity
No 20 30 50
60 40 100
30
Cross-Sectional Study
Prevalence of osteoarthritis among
obese subjects:
40/50 = 0.8
31
Cross-Sectional Study
32
Advantages of cross-sectional studies
1. Easy to carry out & economical to conduct.
2. Describe the distribution of both exposure and outcome in a
population.
34
35
Case-Control Studies
36
Case-Control Studies (CCS)
• Question:
– How do diseased (cases) differ from non-diseased
(controls) with respect to prior exposure history?
– Effect cause.
37
Schulz KF and Grimes DA 2002. Case-control studies. Lancet 359:431-34. 38
Selecting cases
• Select cases after the diagnostic criteria and
definition of the disease is clearly established
• Study cases should be representative of all
cases
• Cases may be located from hospitals, clinics,
disease registries, screenings etc
39
Selecting Controls
• Conceptually, controls should come from the
same population at risk of disease from which
cases develop
• But practically, controls are often selected to
be similar to cases on key factors but without
the disease—because it is difficult to define
the population at risk of disease
• Different types of controls may be used, and
they have different limitations
40
Types of Controls
• Hospital controls
− Have similar quality of information and are convenient to select,
but they may have characteristics or diseases that led to
hospitalization
• Dead controls
− If cases are dead, information of past exposures will be given by
surrogates, such as spouse or children. Dead controls share the
same limitation
• Best friend or neighbor controls
− May share similar characteristics (too similar?)
• Population controls
− Random digit dialing (RDD) is often used
41
Three Qualities Needed In Controls
42
COMPARABILITY VS. REPRESENTATIVENESS
43
Case Definition (example)
• Incident(=newly diagnosed case of breast
cancer)
• Between 1/1/92 and 31/12/93
• Resident of Geneva
• Aged < 75 yrs
• Identified: All pathology labs of Geneva
44
Control Definition
• Never diagnosed with breast cancer
• Between 1/1/92 and 31/12/93
• Resident of Geneva
• Aged <75 yrs
• Randomly from Geneva population
45
First, select
46
Analysis:
Case-Control Study: Example
Cases (CHD) Controls
(without CHD)
Smoked cigarettes 112 176
Did not smoke 224
cigarettes 88
47
Odds Ratio
In a CCS we cannot calculate the Relative Risk (RR) directly
to determine whether there is an association between the
exposure & the disease, coz having started with cases &
controls rather than with exposed & non-exposed
persons, we do not hv info about the incidence of disease
in exposed & non-exposed persons.
48
The OR as a measure of association
49
Referring to e.g. of CHD & smoking on slide 29
OR = 112x224/176x88 = 1.6
Interpretation:
1. OR = 1.0 (implies equal odds of exposure - no effect)
2. If the exposure is +vely related to the effect or
disease, the OR will be > 1
3. If the exposure is –vely related to the disease, the
OR will be < 1
OR of 1.6 means???
50
An odds ratio of
• 1.0 (or close to 1.0) indicates that the odds of exposure
among case-patients are the same as, or similar to, the odds
of exposure among controls. The exposure is not associated
with the disease.
• Greater than 1.0 indicates that the odds of exposure among
case-patients are greater than the odds of exposure among
controls. The exposure might be a risk factor for the disease.
• Less than 1.0 indicates that the odds of exposure among
case-patients are lower than the odds of exposure among
controls. The exposure might be a protective factor against
the disease.
51
Strength of the Association
The magnitude of the odds ratio is called the
“strength of the association.” The further away
an odds ratio is from 1.0, the more likely it is
that the relationship between the exposure and
the disease is causal. For example, an odds ratio
of 1.2 is above 1.0, but is not a strong
association. An odds ratio of 10 suggests a
stronger association.
52
Can Prevalence be estimated from a case control study?
Example:
53
CCS - Advantages
• Quick and cheap (relatively)
– so ideal for outbreaks
http://www.cdc.gov/eis/casestudies/casestudies.htm
55
Case-control Studies - Disadvantages
• uncertain of E D relationship (esp.
timing)
• cannot estimate disease incidence
• worry about representativeness of controls
• inefficient if exposures are rare
• Bias:
– Selection
– Confounding
– Measurement (especially recall bias)
56
Problems in CCS
1. Limitations of Recall: much of the info relating to
exposure often involves collecting data from subjects
by interviews. Because virtually all human beings are
limited to varying degrees in their ability to recall
information, limitations in recall are an important issue
in such studies.
2. BIAS: “Any systematic error in the design, conduct or
analysis of a study that results in a mistaken estimate
of an exposure’s effect on the risk of disease.”
Cont’d
57
Recall Bias
Recall Bias occurs when the recall is better
among cases than controls because of the
presence of the disease – consequently, a
false association may be found between
exposure & disease.
58
Recall Bias (e.g.): suppose we are studying the possible
relationship of congenital malformations to prenatal infections.
We conduct a CCS & interview mothers of children with
congenital malformations (Cases) & mothers of children
without malformations (Controls).
Each mother is questioned regarding infections she may have had
during the pregnancy.
A mother who has had a child with a birth defect often tries to
identify some unusual event that occurred during her
pregnancy. She wants to know whether the abnormality was
caused by something she did. Such a mother may even recall an
event, such as a mild respiratory infection, that a mother of a
child without a birth defect may not even notice or may have
forgotten entirely. This type of bias is known as RECALL BIAS.
59
Selection Bias: Cases and Controls may not be
representative – (prevented by matching), which is the
process of selecting controls so that they are similar to
the cases with regard to certain key characteristics –
such as age, sex & race.
60
Key Facts About Observation Bias
Observation error is an error that arises from systematic differences in the
way information on exposure & disease is obtained from study gps.
This type of bias results in participants who are incorrectly classified as either
exposed or unexposed or as diseased or not diseased.
It occurs after subjects have entered the study
Specific types include recall bias, interviewer bias, & differential &
nondifferential misclassification
Methods for avoiding observation bias include masking interviewers and
subjects to the study hypothesis (interviewer & recall); using a control gp that
is composed of diseased individuals (recall bias); carefully designing the study
questionnaire (interviewer & recall); relying on noninterview data
(interviewer & recall); using multiple measurements, the most accurate
information source, & sensitive & specific criteria to define the exposure &
disease (misclassification)
61
Confounding
• Confounding comes from the Latin confundere- (to
mix together)
• Imagine you notice that some of your patients who
drink coffee get heart disease. This is worrying: could
coffee be causing the heart disease?
Beware the issue of confounding.
62
Confounding occurs when a factor other than the one you are
studying is associated both with the disease and the factor
you are studying (drinking coffee). This can make it seem as
though the factor you are studying causes the disease even if,
in reality, it does not.
For example, if drinking coffee is associated with smoking
(perhaps people who drink coffee also tend to smoke) the
smoking forms a confounding variable that can make it falsely
appear as though coffee drinking leads to heart disease.
The result is that there appears to be an association between
coffee drinking and heart disease when, in reality, this may be
occurring just because the people who drink coffee also
smoke, and it is their smoking (not the coffee drinking) which
actually causes the heart disease.
63
W
Y
X
64
Confounder
(e.g. smoking )
Risk Factor
Association
Relationship of Interest
The solution?
You have to do an analysis that statistically adjusts (or
controls) for the smoking. For example, you could
examine the association between coffee drinking and
heart disease at each level of smoking: first, among non-
smokers, is there an association between coffee and
heart disease? Next, among moderate smokers? And
among heavy smokers? If the original association
between coffee and heart disease was due to
confounding by smoking, you would no longer see an
association when you sub-divide people according to
their smoking level.
66
A definition... For those who want a more formal
definition:
"In a study of the effect of variable X on disease Y,
a third variable, W, may be a confounder if
(a) it is a risk factor for disease Y, and
(b) it is associated with (but does not cause) variable X."
67
Assignment
What are the key characteristics of a
confounding variable?
68
Consider each of the following scenarios & state
whether the variable in question is a confounder:
A. A study of the relationship between contact lens use &
the risk of eye ulcers. The crude RR is 3 & the age-
adjusted RR is 1.5. Is age a confounder in this study?
YES
B. A case-control study of the relationship between cigarette
smoking & pancreatic cancer. In this study, coffee drinking
is associated with smoking and is a risk factor for
pancreatic cancer among both smokers & nonsmokers. Is
coffee drinking a confounder in this study?
YES
69
C. A study of the relationship between exercise & heart
attacks that is conducted among men who do not
smoke. Is gender a confounder in this study?
NO
70
An association between exposure to welding fumes and
COPD was explored in a case-control study. The
following data were reported for 399 COPD patients: 37
currently employed as welders; the remainder had no
occupational exposure. Among 800 controls. 48 were
employed as welders. Set up a 2x2 table, calculate the
OR and interpret it.
COHORT STUDIES
72
Cohort Studies
• Cohort studies begin with a group of people (a
cohort) free of disease.
• The people in the cohort are grouped by
whether or not they are exposed to a
potential cause of disease.
• The whole cohort is followed over time to see
if the development of new cases of the
disease (or other outcome) differs between
the groups with and without exposure.
73
Example-1: you could do a cohort study if you suspect there
might be a causal relationship between the use of a certain
water source and the incidence of diarrhea among children
under five in a village with different water sources.
You select a group of children under five years, either all
children of that age in the village, a random sample taken from
the population register, or e.g. children living in the same area,
or attending the same clinic. Then you classify them as either
using the suspected water source or other water sources. You
check e.g. after two weeks whether the children have had
diarrhea.
You can then calculate the incidence of diarrhea among those
children using the suspected water source and those using
other sources of water supply
74
Example-2: You want to see whether using a cell phone leads to
brain cancer.
1. Draw a sample of people (medical students, for e.g.) who do
not have the disease you are interested in, and collect
information on the factor you have hypothesized to be a cause
of the disease.
2. Collect information on how many minutes each student uses
their phone each week (you might get permission to obtain
this from their phone company bills), and collect this
information over a long time, and then eventually collect
information on who gets brain cancer. You could then see
whether the cases of brain cancer arose among the people
who used their cell phones most often.
3. In technical terms, you record the incidence of cancer among
those who use their phones more than a pre-determined
amount and compare this to the incidence in the non-users.
75
1. Type of Analytic study
2. Unit of observation and analysis: Individual (not groups)
3. Also called follow-up studies, incidence studies,
longitudinal studies, or prospective studies
76
Steps in designing cohort studies
1. Identify group of exposed subjects
2. Identify group of non-exposed subjects
3. Follow-up both groups for disease
4. Measure frequency of disease occurrence in both
groups.
5. Compare risk between exposed and non-exposed
group
78
Cohort study design
Diseased
Exposed
Not diseased
Target
Disease free
Population Cohort Diseased
Unexposed
Not diseased
79
Framingham Study
80
Cohort study: Calculate
Then, follow to see whether & compare
Disease Disease Incidence
First, develops does not of Disease
develop Totals
identify
Exposed a b a+b a/a + b
Not exposed c d c+d c/c + d
83
Attributable risk
• the amount of disease incidence that can be
attributed to a specific exposure
– Difference in incidence of disease between exposed
and non-exposed individuals
= a/a+b – c/c+d = 0.028 – 0.0174 = 0.0106
• In general, all individuals, whether they have or have not
been exposed to a risk factor, have some chance of
developing a disease if no prevention measures have been
taken. The AR estimates the risk above and beyond this
baseline risk that all people have.
• AR can be used to indicate the health benefit of removing
the exposure.
84
Attributable Risk
Excess Risk
Risk
100
40
Risk among risk
20 factor negatives
0
+ -
Risk Factor
85
Advantages of cohort studies
• Measure incidence and thus permit direct
estimation of risk of disease.
• Do not rely on memory for information about
exposure status, hence avoid bias due to selective
recall.
• Cohort studies provide a logical approach to studies
of causation or effects of treatment.
• Cohort approach can yield information on
associations of exposure with several diseases
• Efficient for studying rare exposures
86
Disadvantages of cohort studies
• Require large samples to yield the same number of cases
that could be studied more efficiently in a case-control
study.
• Particularly inefficient for studies of rare diseases.
• Logistically difficult – long follow-up period, often serious
attrition to study subjects.
• Direct observation of participants may cause changes in
health behavior.
• Possible bias in ascertainment of disease due to changes
over time in criteria and methods.
• Very costly.
87
Types of Cohort studies
• Prospective Cohort study
– Concurrent Cohort study or Longitudinal study
• Retrospective Cohort study
– Non-concurrent Cohort or historical Cohort study
• Ambidirectional Cohort study
– To look both ways
88
Differentiating b/w Prospective & Retrospective
90
91
Timing of Cohort Studies
Prospective Cohort Study
* Exposure *
* Disease *
Study starts
Study starts
Ambidirectional Cohort Study
* Exposure *
* Disease *
Study Starts
Time 92
What are the factors that influence the
decision of the researcher to carry out a
particular type of Cohort study?
93
Hierarchy of Study Designs
Time-series analysis
94
Points Regarding Study
Designs
Do not confuse scientific accuracy
with clinical relevance
95
Study Designs
M R
• Observational No No
• Quasi-experimental Yes No
• Experimental Yes Yes
96
Spotting The Study Design – General
Rules
When were the outcomes determined?
• Some time after the exposure?
– Yes… cohort study (‘prospective study’)
• At the same time as the exposure or
intervention?
– Yes… cross sectional study or survey
• Before the exposure was determined?
– Yes… case-control study (‘retrospective
study’)
97
Case-Control Studies:
Question: What Happened?
100
Did investigator assign exposures?
Yes No
Yes No Yes No
Exposure &
Outcome
Exposure to outcome Direction? at same time
Outcome to
exposure
Cohort study Case-control Cross-sectional
study study
101
Classification of types of Epidemiological research
What do Cohort Studies Tell You:
103
Comparison of Cohort & Case-Control study designs
104
Retrospective cohort studies are characterized by all of
the following except:
105
In cohort studies of the role of a suspected factor in the
etiology of a disease, it is essential that:
A.There be equal numbers of persons in both study groups
B. At the beginning of the study, those with the disease and
those without the disease have equal risks of having the
factor
C. The study group with the factor and the study group
without the factor be representative of the general
population
D. The exposed and unexposed groups under study be as
similar as possible with regard to possible confounding
factors
E. Both B and C
D
106
References & preparing for exams
1. http://www.med.uottawa.ca/sim/data/Study_Designs_e.htm
2. http://www.snz.unizg.hr/ph-see/full/u3-t2-roshi-burazeri.pd
f
3. Gordis L. Epidemiology. 5th Edition. Canada: Elsevier
Saunders; 2014.
4. Aschengrau A, George R, Seage III. Essentials of Epidemiology
in Public Health. 2nd Edition. USA: Jones and Bartlett
Publishers; 2008.
5. Park K. Preventive and Social Medicine. 20th Edition.
India: M/s Banarsidas Bhanot Publishers; 2009.
6. http://ocw.jhsph.edu/courses/FundEpiII/PDFs/Lectur
e13.pdf
107