CASE-CONTROL STUDY DESIGNS

Objective

• To describe the design of case control study

• To discuss potential biases in case control study

2
 Outline
• Analytic observational study
• Design of case control study
• Selection of cases and controls
• Sample size
• Assessment of exposure
• Analysis
• Application of case control study design
• Variants of case control study design
• Biases in case control study
3
 Analytic Observational Designs
• Purpose: identify causes of health related conditions
• Reasons for high or low magnitude in specific groups

• Importance
– For identifying causal factors
• In disease control and prevention
– In epidemic control
– For identifying prognostic factors
• In patient management
10/11/2012 4
 Method of identifying causes
• Analytic studies help in identifying causes via

– Comparing groups for presence of association

• Exposed vs non-exposed

• Outcome vs no outcome

• However, presence of association alone doesn’t establish

causation
– Sources of error should be ruled out and other evidences
considered
 Types of Analytic Observational Designs
• Analytic cross-sectional studies

• Cohort designs

• Case-Control designs

6
 Case-Control Studies
• Compare occurrence of exposure in people with the
disease or other condition (cases) and those without the
condition (controls)
• Data is collected on disease occurrence at one point in
time and exposures at a previous point in time

7
 Design of case-control study

• The starting point is selection of cases and controls

• Then exposure history of cases and controls is assessed
•The direction of inquiry is always backwards in time

Fig - Design of case-control study

8
 Types
• Depending on temporal relationship between initiation
of the study/current date/ and occurrence of the disease
– Prospective – uses incident cases

– Retrospective – uses prevalent cases

9
Case-Control studies based on prevalent cases

Exposure Disease Study

occurred occurred takes place
Selection based
ill on disease status
+ -
&
+
exp Retrospective
- assessment
of exposure

10
 Case-Control Studies based on incident cases

Exposure Study Disease Selection based

occurred begins occurs on disease status
Follow study
population
ill
+ -
+ Retrospective
exp assessment
- of exposure

11
12
13
 Design and conduct

Key activities in the design and conduct

of case-control studies
• Selection of cases and controls

• Sample size determination

• Exposure assessment

• Analysis of data

14
 Selection of cases
• Who? Establish criteria for selection of cases
– In order to have no doubt about the types of cases and
stages of disease to be included
– Preferably select all newly diagnosed cases during a
specific period of time
• Incident rather than prevalent cases are preferred
because prevalent cases do not include patients with
short course of disease which may bias the results
15
 Selection of cases

• Where? Identify source for cases– general

population, hospital records, records of employers

– Cases should represent all cases from

a specified population

16
 Selection of controls
• Who?
– Use criteria

– Controls should be representative of controls in the source

population
– Probability of selection should not be affected by exposure

• should have had the same opportunity to be exposed as

the cases have

17
 Selection of Controls
• Where? Sources:
– Controls should be selected from the same source
population of cases
• Random sample of general population usually
best
• Neighborhood (riskier)
• Friends (also problematic)
• Dead controls
• Hospital patients (base ill-defined) 18
How many controls

19
 Sample size
• Formula –
p 1 p  mp
z 1 mp1 2


  1 1 o
n
• Parameters:

 p  z  p1  p1 p
2 o

 and 80% respectively
– Confidence level and power – usually 95% o

– P1 ,P0–estimated magnitude of exposure in the case and controls

respectively
• P’ – derived from P1 ,P0, m and odds ratio (Measure of variability,
similar to standard deviation)
– OR : odds ratio of exposures between cases and controls
– m : number of control subjects per case subject
20
 Assessment of exposure status
• For each of the sampled cases and controls, their
past exposure status is assessed
– In order to assess exposure status, what constitutes an
exposure should be defined first
– All necessary data about exposure can then be obtained
through interview and/or review of medical records
• De-confounding: requires measurement of
potential confounders
21
 Assessment of exposure status

• The method used to measure exposure must be identical

in both cases and controls to avoid the risk of bias
• As much as possible the data collectors should be blinded
to outcome status of participants

– In order to minimize bias that could result from

intensively looking for history of exposure among cases

22
 Analysis
• Presence and strength of association in a sample

– OR: odds of exposure in cases divided by odds of

exposure in controls
• Statistical significance of association

– p-value, CIs

– Statistical analysis – chi-square, logistic

regression

23
 Merits of case-control studies

• Relatively small size is needed as compared to

cohort studies

– Important for etiologic studies of rare disease

• Results are obtainable relatively quickly

• Less expensive in cost

• More than one risk factor can be identified

24
 Demerits of case-control studies

• Affected by inaccuracy and incompleteness of records

• Affected by recall bias

• Problem of bias and confounding is higher

• Difficulty in selecting controls

• OR exaggerates risk

25
 Applications of case-control studies
• Used to screen factors for cohort study
• Not much dependable for identifying factors
to be tackled
• Efficient design for rare diseases or conditions
• Preferred design when resources are limited

– Less expensive and results are obtainable

relatively quickly
• To identify multiple risk factors
26
27
28
29
ODDs ratio as a good estimate of Relative Risk

30
 Biases in case control study

– Confounding

– Selection bias

– Ascertainment bias

• Interviewer bias

• Recall bias

31
 Confounding
• Exposure of interest may be confounded by a factor that is
associated with the exposure and the disease; is an
independent factor for the disease
– A risk factor for disease in the unexposed
– Associated with exposure in the population from which
cases arose
– Not an intermediate step in the causal pathway between
exposure and disease
32
 How to manage confounding

• At the design
– Randomization
– Restriction
– Matching
• At the analysis phase
– Age adjustment
– Stratification
– Multivariable Adjusted

33
 Matching

• Instead of selecting controls at random, controls are often

selected in such a way that they share certain (potentially
confounding) risk factors – with the case.
• Such as age - time

• The main reason is to improve study efficiency

• Matching can be done either on an individual basis or on a

group basis
34
 Matching…
A. Individual matching One or several controls are
matched to individual cases
– This is often done in a non-random manner, especially where the
'nearest neighbor' control is selected
– Selection of non-random control can be accepted if there really is
no risk of selection bias, but usually a random element is essential
– For example, in a case-control study on sleeping sickness,
selection of the nearest village for the control was done in a
randomly chosen direction from the village where the case lived.
35
 Matching…
B. Group or frequency matching Controls are selected so
that the overall make up of the control group is similar to
that of the cases
– If for example 55% of the cases are male, and 45% are
female, then controls will be selected so that the sexes
are in the same relative numbers
– Within that constraint, however selection would still be
random
36
 Matching…
• Advantage: improve study efficiency

• Disadvantage: needs a statistical analysis

• You can no longer use the study to assess the effect of
the matching variables
• May be difficult or impossible to find an exact match to a
case which adds to the cost of the study
• Overmatching may occur;
– The matching factor is not a true confounder
– The matching factor is highly correlated
with other matching variables. Lastly, effects are
irreversible 37
38
Biases in selection of cases and controls

39
 Variants of case control studies

• Nested case control study

• Case cohort

• Case crossover study

40
 Nested Case-Control Studies
• The population within which the study is conducted is a
fully enumerated cohort
• Conduct
– Start a cohort study, defining the main exposures
– Follow participants over time
– Each time a case develops, select a matched control
(time- matching)
– When enough cases have developed, analyze as a
matched
case-control study 41
 Advantages of Nested Case-Control Studies

• Reduces risk of biases

• Helps to get results earlier from the cohort study

• Reduces cost of obtaining further data

– Data is collected on selected cases, and

from only a fraction of the remaining cohort
members (the controls)

42
individual matching

43
 Frequency matching

O M
  [a  d i i
R H

i i
n] i
i
[b  c
44
 Case cohort study design

• A case-cohort study is similar to a nested case-control

study in that the cases and non-cases are within a parent
cohort
• Cases and non-cases are identified at time t1, after baseline

• In a case-cohort study, the cohort members were assessed

for risk factors at any time prior to t1

• Non-cases are randomly selected from the parent cohort

• No matching 45
 Case cross over study
• The case serves as his/her own control, the study is self-
matched
• Used to investigate the transient(Temporary) effects of an
intermittent exposure on the onset of acute outcomes
• For example, cell phone use or sleep disturbances are
transitory occurrences
• For each person, there is a 'case window', the period of time
during which the person was a case, and a 'control window', a
period time associated with not being a case
• Risk exposure during the case window is compared to risk
exposure during the control window

46