2. INTRODUCTION
Unit of observation & analysis – individual, but conclusion is
generalized for the population.
Most frequently undertaken
To identify risk factors for chronic & rare diseases
Best suited to study diseases for which medical care is sought,
like cancers, cirrhosis, SLE & CCF.
2
3. FEATURES
Both exposure & outcome have occurred before start of study
RETROSPECTIVE study – from effect to cause.
CAUSE EFFECT
STUDY
Uses comparison group to support or refute an inference.
3
4. STEPS
1. Selection of cases and controls
2. Matching – making two groups comparable
3. Measurement of exposure
4. Analysis and interpretation
4
5. DISEASED (CASES)
EXPOSED
NOT EXPOSED
TARGET POPULATION
EXPOSED
NOT EXPOSED
NOT DISEASED (CONTROL)
5
6. SELECTION
DIAGNOSTIC CRITERIA
Establish diagnostic criteria and definition of disease & stage
of disease.
Once established, diagnostic criteria should not be changed.
Study cases should be representative of all cases.
Sources of cases can be hospitals or general population.
6
7. SELECTION
ELIGIBILITY CRITERIA
Incident cases are preferred to prevalent cases to reduce-
Recall bias
Over-representation of cases of long duration.
Most desirable way is to include all incident cases in the
population in a specific time period.
7
8. SELECTION
SELECTING CONTROLS
From same population at risk for the disease as the cases.
Should be representative of population
Help to estimate exposure rate to be exposed.
Sources can be hospitals, relatives, neighbours or general
population.
Best possible ratio for number of cases to controls is 1 : 1
Maximum permissible option is 1 : 4 (for rare diseases)
8
9. MATCHING
Can be group matching or pair matching
Example: If there are 10 people in cases of age group
15-20, even controls are matched such that they too
contain 10 people of age group 15-20 (Group
matching). If each person is matched with respect to
the factors, then it is pair matching (i.e., there is one
person in control for each person in case with similar
features except for the disease under study.)
Confounders should be avoided.
9
10. MEASUREMENT
Is an estimate unless past measurements are available.
It has to be assumed that exposure incurred at the time the
disease process began (this may not be valid)
Subjected to recall & interviewer bias
Potential confounders need to be assessed.
10
11. ANALYSIS
ENTERING DATA
CASES CONTROLS
EXPOSED a b
NOT EXPOSED c d
TOTAL a+c b+d
PROPORTION 𝑎 𝑏
EXPOSED 𝑎+ 𝑐 𝑏+ 𝑑
11
12. ANALYSIS
ODDS RATIO
Odds of exposure for cases compared to controls
𝑎𝑑
OR =
𝑏𝑐
INTERPRETATION OF OR
If OR = n, then the cases are n times more likely to be exposed than
the controls.
12
13. ADVANTAGES
Quick and easy to complete
Cost effective
Efficient for rare diseases
Small sample size sufficient
13
14. DISADVANTAGES
Temporality of association
Inability to provide a direct estimate of risk
Not efficient for studying rare exposures
Subject to biases.
14
15. THANK YOU
THANK YOU for reading my presentation.
If you have any doubts or interesting cases in any subject of
medicine, I will be delighted if you share at
keshavapavan533@gmail.com
15