CASE CONTROL

Mzumbe University
MSc HME 602
Dec, 2016
Bunini (MD)
 Case control design
Definitions
Case?
Control?
 CC Definition
• Is an observational epidemiological
investigation, characterized by the measure
of past exposure to risk factors, of two
separate groups, named “cases” and
“controls”. (Breslow and Day, 1980;
Kleinbaum et al., 1982).
• Also known as “retrospective study” or
“research in reverse” designed to determine
the relationship between a particular
outcome of interest and a potential cause.
 Case control studies
• Start with a disease & work backwards to
find associations
• between exposures and disease
• Compare people with and without disease
to determine how they may differ in what
they were exposed to
 Checklist key components
in cc design
• Study base-all people who would be in the study if
they developed the disease
• Disease (case definition)-clear definition, newly
diagnosed
• Controls-random sample of those without disease
in the study base
• Exposure
• Confounders-what can confuse/distort the results
• Analysis-how the information is handled
• Selection bias-were the populations similar?
 Cases
• Selection is based on disease status- cases
(have disease)
• Selected only after developing clear criteria
for defining the disease of interest
• Should be representative of all cases
• May be found from clinics, hospitals, disease
registries
• Newly diagnosed disease is preferred
• Maybe limited to cases occurring within a
specific period of time
 Controls
• Selected based on disease status- controls
(not have disease of interest) but are at risk of
the disease in his/her future
• May be matched by age, weight, sex & other
factors
• Selection can be problematic
• Should be representative of population of
interest in general
• Can have multiple control groups
 Sources
Population based e.g.
• Schools
• Selected service
• Insurance
• Companies
• Neighbors
• Friends
• Spouse/siblings
Hospital based
• Clinics
• In-patient
• Disease registry
• Death registry
• Special surveys, and reporting systems such as cancer or birth defects
registries
 Indications for case-control studies

• Rare diseases
e.g. idiopathic pulmonary fibrosis and most cancers
• Little known about disease
• Evaluate many exposures
• Disease has long induction and latent period
• Exposure data are expensive
• Underlying population is dynamic

9
 Formation of groups in CC design
• Define a study population (source of cases and controls)
• Controls must have as similar background as possible to
the cases
• Control do not have the outcome
• Their selection should be independent of the exposures
of interest
• The criteria for diagnosis of a case should be defined as
well as the eligibility criteria used for selection.
• The diagnostic criteria should be sensitive and specific
(i.e. strict)
•

10
 Matching
• The process of selecting the controls so that they are
similar to cases in certain characteristics (i.e. age,
race, sex, socioeconomic status & occupation)
Types:
• Group/ frequency
• Individual/ matched pairs
Problems:
• Matching on many variables may make it difficult to
find appropriate controls
• Cannot explore possible association of disease with
variables that are matched
 Use of multiple controls
Same type
• Increases power of the study (1:1 but can
go up to 1 case: 4 controls)
Different types
• for exploring different hypotheses
• For accounting for potential biases such
as recall bias
 Exposure

• Examples: smoking, behaviors, food

habits, genetics
• Exposure occurred before disease
developed
• Subject to recall bias and reporting bias
• Confounders must be addressed; attempt
to control with appropriate design &
statistical techniques
 Data collection
Sources of information on exposure and
potential confounding factors:
• Existing records (i.e. hospital records,
study records)
• Questionnaires
• Interviews
• Biological specimens
• Tissue banks
 Bias
Focus is to minimize
• Selection bias
• Detection bias
• Information bias (recall bias)
• Confounding
– can be controlled
– by statistical analyses but we
– cannot do anything about bias
– after data collection.
 Distinguishing features of case-control
study
• multiple exposures in relation to a disease can be
examined and exposure histories are compared
• The study groups are defined by outcome
• Provides an OR
• important scientific findings with relatively little time,
money, and effort compared with other study
designs
• Usually, proceeds from documented disease and
investigates possible causes of the disease
• The starting point is subjects with the disease or
condition under study (cases).

16
 Analysis
The investigator compares the frequency of exposure of the two groups
Odds ratio
• is the measure of risk used to look at odds of exposure between
cases & controls
Interpretation:
• OR=1 - equal odds of exposure
• OR>1 - positively related
• OR<1 - negatively related
• If cases have lower odds of exposure than controls, exposure may
reduce the risk of disease (i.e. protective effect)
• If cases have higher odds of exposure than controls, exposure may
increase risk of disease
• If cases & controls have equal odds of exposure, exposure is likely
not related to disease development
 Exposure Rate

Exposure Yes disease No disease total

Yes A B A+B

No C D C+D

total A+C B+D A+B+C+D

 Measures of effects in cc design
Odds Ratio (OR) 2X2 table

Disease + Disease - Total

Exposed a b a+b

Non exposed c d c+d

Total a+c b+d a + b +c + d

Exposure odds among cases = a/c

Exposure Odds among controls = b/d
Disease Odds among Exposed = a/b
Disease Odds among non exposed = c/d

19
 Measures of effects…
Attributable risk in a case-control study
• Generally not possible to estimate
incidence rates. Therefore, it is not
possible to directly estimate attributable
risk (AR) in case control studies
• However, it is possible to estimate
attributable risk percent AR (%) using the
following alternative expression for AR%
• AR (% ) = (OR – 1)/(OR)x 100
20
Assessment for statistical significance in
case-control studies

• Case-control studies can not yield incidence

rates instead, they provide an odds ratio (OR)
• When the incidence rate of a particular outcome
in the population of interest is low (under 5% in
both the exposed and unexposed suffices) the
odds ratio from a case-control study is a good
estimate of relative risk

21
Odds Ratio
 Answer
OR
=Exposed cases/unexposed cases divided by
exposed controls/unexposed controls
=40/10 divided by 15/25
=9.33
Odds Ratio
 Answer
Cases Controls

Exposed 40 15

Not exposed 10 35
 A case control study of factors
associated with oropharyngeal cancer
Smoking ≥ 20 packs per yr Cancer No cancer

Exposed 34 39

Not exposed 44 119

 Answers
1. Exposure rate
Cases=a/a+c=34/34+44=43.6%
Controls=b/b+d=39/39+119=24.7%
2. Odds ratio
ad/bc=34*119/39*44
4046/1716=2.4
 Advantages

• Efficient- saves time & money

• Used for rare diseases, smaller sample sizes
• Multiple associations with a disease
• Can be used to examine multiple exposures
• No risk to participants
• No attrition problems
• Minimal ethical issues
 Disadvantages

• Susceptible to bias-recall
• Selection of controls can be difficult
• Concern about representativeness of cases &
controls
• Cannot measure incidence
• Cannot associate between causes &
associated factors
• Inefficient for examining rare exposures
• Not good for evaluating diagnostic tests
(cases have condition and controls do not)
The concept map of design of a case-control study

30
 Read the shared article using case
control design
• Understand the application odds ratio as a
measure of effect in the article
• In group of threes, download and discuss any
simple abstract from internet for presentation
in class
• Describe during presentation (objective,
hypothesis, methods, results and conclusion)
• Be ready to respond to reactions from the
floor members in the session
 Readings
1. Rothmans K. Modern Epidemiology
3rd ed.
2. Woodward M. Study Design and
Data Analysis, 2nd ed.
3. Journal of Medical Screening 2004 Volume 11
Number 4