Epidemiology
1)Observational – Descriptive – Formulation of
Hypothesis
ii) Analytical – Case Control, Cohort – Tests
Hypothesis by calculating strength of
Association
2) Experimental (Interventional)
CASE CONTROL STUDIES
A type of analytical study
• Tests hypothesis unlike descriptive studies which
only formulates hypothesis
- Determines whether or not statistical association
exists between a disease and suspected factor
- If it exists , measures strength of association –
Odds Ratio
- Called Retrospective study
• Both exposure & outcome(disease) has
occurred before the start of study
• Study proceeds backwards from effect to
cause
• It uses CONTROL , comparison group to
support or refute inference
4 steps in conducting case control study
1. Selection of cases and control
2. Matching
3. Measurement of exposure
4. Analysis & Interpretation
SELECTION OF CASES
I. Diagnostic criteria – eg stage of any disease
- Must be specified before the study is carried out
and should not be altered till the study is over
II. Eligibility criteria – eg inclusion of new cases
diagnosed recently or diagnosed within specified
period
• Sources of cases – Hospital, General population
SELECTION OF CONTROLS
• Controls must be free from the disease under
study and similar to cases in other aspects
• They can be selected from hospitals,
relatives , neighborhood or general population
• For 1 case there may be up to 4 controls
MATCHING
• A process by which controls are selected in such a
way that they are similar to cases in certain
selected variables that influences disease outcome
• Matching done to ensure comparability between
cases and controls
Controls should not differ from cases in age,
gender, occupation, social class
Etiological factor under study should not be matched
Confounding factors
• One which is associated with both exposure
and disease
• It is itself a risk factor for disease
• Example – In a study of role of alcohol in
esophageal cancer , smoking is confounding
factor
• Measurement of exposure - Information
obtained about the exposure by interviews,
questionnaires or by studying past records
from hospitals , employment records
• Analysis – Calculation of ODDS RATIO (Cross
product ratio) which measures strength of
association
Case control study
Cases(with lung cancer) Controls( no lung cancer)
Smokers 40 (a) 20 (b)
Non Smokers 10 (c) 30 (d)
50 50
Exposure Rates
Cases – 40/50 – 80%, Controls – 20/50- 40%
Odds ratio measures strength of association
Odds Ratio – a/c
b/d
ad/bc = (40 x 30)/ (10 x 20) = 6
Interpretation – Smokers are 6 times at higher risk of
developing lung cancer than non smokers
Bias in case control study
Bias – Systematic error
• Bias due to confounding
• Recall/ Memory bias
• Selection bias
• Interviewer’s bias
Advantages – Relatively cheap, rapid & easy to
carry out (compared to cohort)
• Suitable to investigate rare disease
• Require fewer subjects
• No risk to subjects
• Ethical problems minimal
• Risk factors can be identified
• No attrition problems, follow up of patients
not required
Disadvantages
Recall bias
Difficulty in selection of cases and controls
Association between maternal Diethylstilbestrol(DES) therapy and
adenocarcinoma of vagina in the female offspring (1971 case
control study)
DES given to pregnant women in 1st trimester of
pregnancy to prevent miscarriage
- Unusual occurrence of this tumor in 8 young
women(15 to 22 years of age) born in one
Boston hospital between 1966 and 1969
- This unusual time clustering of cases -7 occurring
within 4 years of a single hospital led to inquiry
- The possible causes of tumor was investigated by
a case control study in 1971
• Female birth occurring closest in time to each
patient was selected as controls
• 7 of 8 women had been exposed to DES in
their fetal life
• Case control method played a critical role in
revealing exposure to DES in utero as cause of
vaginal adenocarcinoma
Does use of oral contraceptive pills(OCP) increase the
risk of developing venous thrombosis
Cases(with Controls( no
venous venous
thrombosis) thrombosis)
OCP users 60 (a) 70 (b)
Non users 30 (c) 110 (d)
90 180
Odds ratio = ?
• In a study , when 20 patients diagnosed Coronary
Artery Disease(CAD) were asked about their dietary
habits in the past it was found that 15 were consuming
high fat diet. Out of 40 healthy subjects included in this
study only 10 were consuming high fat.
- What type of study is this and why?
- Draw 2 X 2 table for this study
- Measure the strength of association between the
exposure & outcome & interpret your findings
- Mention 2 advantages & 2 disadvantages of this type
of study
- What type of bias can occur in this type of study?
• Thank You
