CASE CONTROL STUDY DESIGN
SULTANA AL SABAHI
MSC. IN EPIDEMIOLOGY AND MEDICAL STATISTICS
OUTLINE
1. Introduction
2. How to select the most appropriate study design?
3. When case control study is desirable?
4. Selection of cases
5. Selection of controls
6. Source of exposure information
7. Analysis of case control studies
8. Strengths and limitations of case control studies
9. Case cross over designe
INTRODUCTION
Epidemiologic
design
Formulate hypothesis
Observational
study
Descriptive
studies
Case report
and case series
Ecological
study
Interventional
study
Analytic studies
Cross sectional
survey
Case control
study
Randomized
control trail
Non
Randomized
control trail
Cohort study
Test hypothesis
HOW TO SELECT THE MOST APPROPRIATE STUDY DESIGN?
Clinical trail
Goal:
to harvest
valid and
precise
information
on
association
between
exposure
and disease
using a
minimum of
resources
Case control
Prospective
* Research question
involves prevention or
treatment
* Small effect expected
* Ethical and feasible
* Money is available
* Little known about disease
* Evaluate many exposure
* Disease is rare
* Disease has long induction
and latent period
* Disease has short
induction and latent
period
* Current exposure
* Want high quality data
* Research question involves
prevention, treatment or
causal factor
* Moderate or large effect
expected
* Trail not ethical or feasible
* Trail too expensive
* Little known about
exposure
* Evaluate many effect of an
exposure
* Exposure is rare
* Underlying population is
fixed
* Disease has long
induction and latent
period
* Historical exposure
* Want to save time and
money
Observational study
Cohort
Retrospective
FEATURE OF CASE-CONTROL STUDIES
Case-control studies are used to
identify
factors
that
may
contribute to a medical condition
by comparing subjects who have
that condition (the ‘cases’) with
subjects who do not have the
condition
1. Directionality: Outcome to
exposed
Disease population
exposed
exposure
2. Timing: Retrospective for
exposure, but case-ascertainment
can be either retrospective or
prospective
non-exposed
Disease +
non-exposed
SELECTION OF CASES
2. Source of the cases
1. Case definition
• Population based cases
• Standard case definition
• Hospital based cases
• Standard diagnostic criteria
* It better to be restrictive
than inclusive
* Subdivisions usually
improve the accuracy of case
definition
Both reduce classification
errors
* Population cases
more Representative.
* The risk factor
could be unique to
the selected hospital.
3.
Type of the cases
• Incident cases
• Prevalent cases
Incident cases are
preferable to the
prevalent
SELECTION OF CONTROLS
 Comparability is more important than representativeness in the selection of controls
 The control should be at risk of the disease
 Men can not be controls for a gynecological condition
 The controls should be “eligible for the exposure
 The control should resemble the case in all respects except for the presence of disease
(and any as yet undiscovered risk factors for disease)
TYPES OF CONTROLS
Source of controls
Advantage
Disadvantage
Peer or case-nominated
(friend/neighbor) control
Individual willing to participate
Cases unable to nominate
Case unwilling to nominate
Control share the habits of the cases
Dead control
Ensure comparable data collection
procedure
Not representative for source
population
Hospital or clinic control
Easy identification
Good participation rates
Less expensive
More comparable to the cases
Difficulty in determining the
appropriate illness for inclusion
Population controls
come from the same based
population as the cases
Time consuming and expensive
Cooperation is difficult
Recall bias
SOURCE OF EXPOSURE INFORMATION
Source
Type
Key characteristics
Questionnaires
Face-to-face,
telephone, selfadministered
Can obtain information on many exposures with relative ease and flexibility
Must be carefully designed and administered to elicit accurate information
Expensive
Pre-existing records
Administrative,
medical, regulatory
May be the only available exposure source
Avoids bias
May be incomplete
May lack uniformity and details
Inexpensive
Biomarkers
Level in blood,
urine, bone,
toenails
Can estimate internal dose
Infrequently used because of difficulty identifying valid and reliable markers
of exposure to non-infectious agents
Expensive
SUMMARY
Cases
Exposed
Not Exposed
Exposed
Population
Controls
Not Exposed
ANALYSIS OF CASE CONTROL STUDIES
Data is expressed in a four-fold table, and an odds ratio is calculated
OR= ad/bc
STRENGTHS AND LIMITATIONS OF CASE CONTROL STUDIES
Strength
Weaknesses
Efficient rare disease
Efficient for disease with long latency
period
Inefficient for rare exposure
May have poor information on exposure
Conducted within short period of time
Low cost
Allows to study all possible factors
associated with the disease
Vulnerable to bias
Difficult to infer temporal relationship
BIAS AND CONFOUNDERS
• case control studies
are potential sources of
many biases
• should be carefully
designed, analyzed,
and interpreted.
CASE-CROSSOVER STUDY: ANEW TYPE OF
CASE CONTROL STUDY
Definition of case cross-over
• The case-crossover design was developed to study the effects of
transient, short-term exposures on the risk of acute events.
Terminology
• The term crossover mainly used to describe experiment in which all individuals
pass through both the treatment and placebo phase.
Case cross-over
Case cross over studies are the case control version of crossover studies. In which
all subjects are cases and exposure is measured in two different periods of time.
Principle
• We can consider the case cross-over design as a scientific way to ask and
answer the following question
was the patient doing anything different just before the onset of the disease?
By making comparison within the individual
Illness
Did I do anything unusual right
before the illness
In comparison to my
usual routine
Component
Hazard period
Control period
The time period right before the Specified time period other than the hazard
disease or event onset.
period.
When the event occur.
When the event dose not occur.
The treatment phase.
The placebo phase.
Comparison
Illness
Hazard period
Control period
Relation with matched case control
• Since there is hazard period and control period and each individual provides
information for both hazard and control period, case crossover design can be
viewed as a matched case control study design where it involve cases only
and each individuals serve as his/her own control.
Hazard period
• The length of hazard period is very important and it is obtained from:
o Usually is the same as the length of exposure effect period.
o Or decided according to the investigator past experience.
o Overestimation…..many false exposure will become exposure.
o Underestimation….some of the true exposure will be excluded.
o When either of this occur, the association between the event and the exposure
can not be evaluated correctly.
Example
• Triggering of MI by physical exertion
 Event: Myocardial infraction (MI).
 Exposure: Heavy physical exertion.
 Length of the exposure effect: 1 hour .
 Hazard period: 1 hour before MI onset.
Hazard period (1 hr)
Physical exertion?
Control period (1 hr)
MI
9 pm May 28
9 pm May 27
Analysis
• The analysis will be the same as that for a matched case-control study. However,
instead of case and control we will have the hazard period and control period.
Control period
Hazard
period
Exposed
Unexposed
Exposed
A
B
Unexposed
C
D
OR = b/c
Limitations
• The design can only be applied when the time lag between the exposure
and outcome is brief and the exposure must have little carryover effect.
• The result of the analysis are short-term risk than cumulative risks and
relative risk but not absolute risk.
• Recall bias my be occur during data collection.