1

CASE-CONTROL STUDY DESIGN

 Learning objectives
2

 At the end of this module, you will be able to:

 Describe the basic concepts of case-control study
 Describes how to conduct and analyse case-control
study
 Describes the types of bias and confounding in
case-control and how to control in this study design
 Describe the strength and weakness of case-control
study
 Case-control study design
3

 The case-control study is an analytic epidemiologic

research design in which the study population
consists of groups, who have (cases) and do not
have (controls) a particular health problem or
outcome.
 The investigator looks back in time to measure
exposure of the study subjects.
 The exposure/s is/are then compared among cases
and controls to determine if the exposure could
account for the health condition of the cases.
 Case-control study…
4

 In case-control study design

 First detect a number of people with disease under
study; cases.
 Then select a number of people who are free of the
disease; the controls
 The cases and controls are then investigated to see
which risk factors differ between them
 Case-control study…
5

 The clear difference between a case-control and

cohort study is:
 In case-control, we select by disease status and
look back to see what, in past, might have caused
the disease.
 By contrast, in cohort, we select by exposure status
and wait to see whether disease develops.
 Case-control study design
6
 When is it desirable to conduct a case-
control study?
7

 When exposure data are expensive or difficult to

obtain.
 When disease has long induction and latent period
Ex: Cancer, cardiovascular disease
 When the disease is rare
Ex: Studying risk factors for birth defects
 When little is known about the disease
Ex. Early studies of AIDS
 When underlying population is dynamic
Ex: Studying breast cancer on Cape Cod
 Why case-control design for study of
rare diseases?
8

 Consider some rare disease say some cancer

(leukemia)
 Crude Annual Incidence = 3.4/100,000 (< 15 years)
 Cohort Study: A year of observation on a million
children to identify 34 cases
 Sample of 34 cases: Sub-divided in 2 or more exposure
categories
 What about conducting case-control design?
 Issues in the design of case-control studies
9

 How to run a case-control study

 Decide on the research question to be answered.
 Formulate a hypothesis and then decide what will
be measured and how.
 Specify the characteristics of the study group and
decide how to construct a valid control group.
 Then compare the “exposure” of the two groups to
each variable.
 Issues in the design of case-control studies
10

A. Define the research question

 To what extent does use of insecticide-treated bed
nets (ITN) reduce the risk of malaria?
 Is OC use associated with MI in women?
 Is current IUD use associated with PID?
 Is OC use associated with the risk of breast cancer?
 Is age at first coitus associated with cervical cancer?
 Issues in the design…
11

B. Identify the study base

 Is the population in which the cases arise; could be
geographically defined population, patients of a
certain clinic, an occupational group, etc.
 Should be clearly defined
 Issues in the design …
12

C. Case definition
 It is essential that the case definition is clearly
defined at the outset of the investigation to ensure
that all cases included in the study are based on the
same diagnostic criteria.
 Precise definition of cases(clinical, laboratory and
other criteria)
 Case definition should avoid misclassification
 Standard/consensus definitions if available, must
be used.
 Issues in the design …
13

 For example;
 Conceptual definition, Obesity defined as body fat
percentage > 33%
 Operational definition, Body Mass Index > 30
 Issues in the design …
14

D. Inclusion and exclusion criteria

 Cases should be selected to improve validity(ex. By
excluding cases with coexisting disease)
 Cases should be restricted to limited time period and
geographic area, age range etc…
 Issues in the design …
15

E. Source of cases
 The source of cases needs to be clearly defined.
 Cases may be recruited from a number of sources;
 For example: they may be recruited from a
hospital, clinic, GP registers or may be population
bases.
 Population based case control studies are generally
more expensive and difficult to conduct.
 Issues in the design …
16

F. Selection of cases
 Should be all cases arising in the study base during
a period of time, or a representative sample of all
cases.
 Case-control studies may use incident or prevalent
cases.
 Incident cases comprise cases newly diagnosed
during a defined time period.
 Prevalent cases comprise individuals who have had
the outcome under investigation for some time.
 Issues in the design …
17

 Incident or prevalent cases?

 To the extent possible avoid prevalent cases. Why?
 Chicken-egg dilemma on cause and effect
 Non-representative cases since long-time survivors
are selected
 Prevalent cases with long disease duration may not
accurately recall antecedent events
 Issues in the design …
18

 Difficult to distinguish prognostic factor and cause

 The interpretation of results based on prevalent
cases may prove more problematic, as it may be
more difficult to distinguish prognostic factor and
cause.
 For example: individuals may modify their
exposure following the onset of disease.
 Issues in the design …
19

G. Selection of controls
 A particular problem inherent in case-control studies
is the selection of a comparable control group.
 Controls are used to estimate the prevalence of
exposure in the population which gave rise to the
cases.
 Ideal control group would comprise a random sample
from the general population that gave rise to the
cases.
 Issues in the design …
20

 The goal is to select individuals in whom the

distribution of exposure status would be the same as
that of the cases in the absence of an exposure
disease association.
 That is, if there is no true association between
exposure and disease, the cases and controls should
have the same distribution of exposure.
 Issues in the design …
21

 The controls should come from the population at risk

of the disease
 Men can not be controls for a gynecological
condition
 The controls should be “eligible for the exposure”

 The controls should have same exposure rate as that

of the population from where the cases are drawn
 Issues in the design …
22

 If the controls are to be used as a comparison group

to make inferences;
 Then comparability of the controls to cases becomes
another important dimension to consider in our
selection process.
 In selecting controls, we need to address both
Comparability and Generalizability concerns.
 Besides, practicability and economic impact should
be also considered.
 A high level of comparability assures the validity of
the findings.
 Issues in the design …
23

 Principles of control group selection

1. Study base: cases and controls must be drawn from

same population (controls and cases should came
from the same source population)
2. Comparable information: aimed at avoiding
information bias (recall bias and observer bias)
3. De-confounding: requires measurement of potential
confounders or matching.
 Issues in the design …
24

H. source of controls
 The source of controls is dependent on the source of
cases.
 In order to minimize bias, controls should be
selected to be a representative sample of the
population which produced the cases.
 For example, if cases are selected from a defined
population such as a GP register, then controls
should comprise a sample from the same register.
 Issues in the design …
25

 In case-control studies, controls can be recruit from

 General population
 Neighborhood
 Friends/relatives
 Hospital or clinic-based
 Issues in the design …
26

 Where to select controls from?

 Balance the pros and cons
 Analyze the situation for bias being introduced
 If possible,
 Select different sources of controls and compare
with each other
 Compare the inferences drawn
 Issues in the design …
27

General Population Controls:

 Population defined by geographic boundaries (or
specific characteristics)
 Cases may include all cases, or a random sample of
all cases.
 Controls should be a random sample of non-diseased
individuals eligible to be cases.
 Issues in the design …
28

 Advantages:
 Generalizability is possible
 Good when cases are selected to represent affected
individuals in a defined population.
 For example, if cases to that particular hospital are
coming from a geographically defined area selection
of controls from the entire population could be
possible.
 Issues in the design …
29

 Disadvantages:
 Costly and time-consuming
 Recall bias - controls may not recall exposures to the
same level of accuracy since they may not be
seriously concerned about their illness.
 People might be less motivated to participate for the
same reason given above, which increases non-
response rate, i.e., selection bias.
 Issues in the design …
30

 Hospital/Clinic-Based Controls:
 Source population refers to people who are served by
hospital or clinic.
 Are a convenient and cheap source of controls,
especially in situations when a clinical procedure,
such as a blood sample is required to measure the
risk factor.
 When cases are hospital based, it is common to
recruit controls from the hospital population.
 Issues in the design …
31

 The choice of controls from a hospital setting should

not include individuals with an outcome related to
the exposure(s) being studied.
 For example, in a C-C study of the association
between smoking and lung cancer the inclusion of
controls being treated for a condition related to
smoking (e.g. chronic bronchitis) may result in an
underestimate of the strength of the association
between exposure (smoking) and outcome.
 Issues in the design …
32

 Advantages:
 Easily identified and readily available in sufficient
number with reduced cost.
 More likely than healthy individuals to be aware of
antecedent exposures or events (minimize recall
bias).
 Are also likely to have been subject to the same
intangible selection factors that influence cases to come
to this particular physician or hospital(minimize selection
bias)
 Issues in the design …
33

 More likely to be cooperative because they anticipate

benefit from their involvement or might think that its
related with their illness(reduce non-response bias).
 Their medical data are likely to be of comparable quality
to those from the cases and may have been collected
prior to classification as controls (remove the possibility
of observer bias).
 In short, convenience, low-cost to identify and interview,
comparable information quality as cases, motivation
to participate, and comparable health-care seeking
behaviour.
 Issues in the design …
34

 Disadvantages:
 Because they are ill they are different from healthy
individuals in many ways.
 There is danger of altering the direction of
association or masking a true association between
exposure and outcome of interest.
 The risk factor for the study disease may be a risk
factor for the condition that a particular control has,
w/c is the cause of his or her hospitalization.
 Issues in the design …
35

 Special controls
 Special controls are individuals which are related to
the cases in some way.
 These are friends or household members
(siblings...).
 Issues in the design …
36

 Advantages:
 They are healthy
 More likely to be cooperative than members of the
general population, because of their interest in the
cases
 Offer a degree of control over some confounding
factors, such as ethnicity, socioeconomic status, or
environment
 Issues in the design …
37

 Disadvantage:
 If the study factor is likely to be similar to the cases,
an underestimate of the true effect of the exposure of
interest may result.
 E.g. If the study factor is diet, it will be similar for
both cases and controls, if controls are siblings.
 Issues in the design …
38

 Ratio of control to case

Statistical consideration
 When the number of subjects available in one group

(cases) is limited, an increase in the other group

increases the study power
 Gain in power is till the ratio of 4:1

 Thereafter, the gain is not substantial but cost

increases
 Issues in the design …
39

 Number of control groups

 Validity of inferences
 Even when there is no statistical need, more than one control
may be recruited per case
 Enrolling two or more types of controls is a way of checking
for biases introduced by choice of control group
 If the measure of effect is similar when comparing cases with
each control group
 Probably – no biases (no surety)
 If different measure of effect, then the bias is there
 Issues in the design …
40

I. Measuring exposure status

 Exposure status is measured to assess the
presence or level of exposure for each individual
for the period of time prior to the onset of the
disease or condition under investigation when the
exposure would have acted as a causal factor.
 Note that in case-control studies the measurement
of exposure is established after the development
of disease and as a result is prone to both recall
and observer bias.
 Issues in the design …
41

 Various methods can be used to ascertain exposure

status. These include:
 Standardized questionnaires
 Biological samples
 Interviews with the subject
 Interviews with spouse or other family members
 Medical records
 Employment records etc.
 The procedures used for the collection of exposure
data should be the same for cases and controls.
 Issues in the design …
42

 How far back should exposure be assessed?

 Define a part of the person’s exposure history
considered relevant to the etiology of disease (e.g.
the “empirical induction” period).
 Code the exposure data in an etiologically-relevant
manner (e.g. magnitude of exposure, years of
exposure, ever exposed, etc.).
 Analysis of case-control studies
43

 The odds ratio (OR) is used in case-control studies

to estimate the strength of the association between
exposure and outcome.

 Note that it is not possible to estimate the incidence

of disease from a case control study unless the
study is population based and all cases in a defined
population are obtained.
 Analysis …
44

 The results of a case-control study can be presented

in a 2x2 table as follow:

 The odds ratio is a measure of the ratio of the odds

that the cases were exposed to the odds that the
controls were exposed as:
 Analysis …
45

 An overall summary measure, adjusted for the

effects of confounding, and a statistical test of
significance can also be calculated.
 In addition, confidence intervals for the odds ratio
would also be presented.
 Bias in case-control studies
46

 Sampling bias
 The patients with the disease may be a biased
sample (for example, patients referred to a teaching
hospital) or the controls may be biased (for
example, volunteers, different ages, sex or
socioeconomic group).

 Selection bias: Systematic error due to differences

in characteristics between those selected for a study
and those not selected
 Bias …
47

 Observation and recall bias

 As the study assesses predictor variables
retrospectively there is great potential for a biased
assessment of their presence and significance by
the patient or the investigator, or both.

 Recall bias: Systematic error due to differences in

accuracy or completeness of reporting of past
events or experiences
 Overcoming selection bias
48

 Ideally the cases studied should be a random

sample of all the patients with the disease.
 This is not only very difficult but in many instances
is impossible because many cases may not have
been diagnosed or have been misdiagnosed.
 For example, many cases of non-insulin dependent
diabetes will not have sought medical attention and
therefore be undiagnosed.
 Selecting the controls is often amore difficult
problem.
 Overcoming selection bias …
49

 To enable the controls to represent the same

population as the cases, one of four techniques may
be used.
1. A convenience sample-
 Sampled in the same way as the cases, e.g.
attending the same outpatient department.
 While this is certainly convenient it may reduce the
external validity of the study.
 Overcoming selection bias …
50

2. Using a population based sample for both cases

and controls.
 It is possible to take a random sample of all the
patients with a particular disease from specific
registers.
 The control group can then be constructed by
selecting randomly from the same population as the
area covered by the disease register.
 Overcoming selection bias …
51

3. Matching-the controls may be a matched or

unmatched random sample from the unaffected
population.

 Again the problems of controlling for unknown

influences is present but if the controls are too
closely matched they may not be representative of
the general population.
 Overcoming selection bias …
52

4. Using two or more control groups.

 If the study demonstrates a significant difference
between the patients with the outcome of interest
and those without, even when the latter have been
sampled in a number of different ways (for
example, outpatients, in patients, GP patients)
then the conclusion is more robust.
 Overcoming information Bias
53

 Overcoming retrospective recall bias can be

achieved by using data recorded, for other
purposes, before the outcome had occurred and
therefore before the study had started.
 The success of this strategy is limited by the
availability and reliability of the data collected.
 Overcoming information Bias
54

 Another technique is blinding where neither the

subject nor the observer know if they are a case or
control subject.
 Nor are they aware of the study hypothesis.

 Observers can also be easily blinded to the case or

control status of the patient where the relevant
observation is not of the patient themselves but a
laboratory test or radiograph.
 Strengths and weaknesses of case-control
studies
55

 Strengths
 Cost effective relative to other analytical studies such as
cohort studies.
 Case-control studies are retrospective, and cases are identified
at the beginning of the study; therefore there is no long follow
up period (as compared to cohort studies).
 Efficient for the study of diseases with long latency periods.
 Efficient for the study of rare diseases.
 Good for examining multiple exposures.
 Strengths and weaknesses …
56

 Weaknesses
 Particularly prone to bias; especially selection,
recall and observer bias.
 Case-control studies are limited to examining one
outcome.
 Unable to estimate incidence rates of disease
(unless study is population based).
 Poor choice for the study of rare exposures.
 The temporal sequence between exposure and
disease may be difficult to determine.
 Sample Size calculation
57

 Power :  probability of detecting a real effect

 Alpha level : probability of detecting a false effect
 P0 :  probability of exposure in controls
 P1 :  probability of exposure in case subjects
 R  : odds ratio of exposures between cases and controls
 m : number of control subjects per case subject
 Sample Size Calculation
58
 The estimated sample size is:
 software for Sample Size Calculation
59
60

The end