2. OVERVIEW
๏ Introduction.
๏ Design of a Case-Control Study.
๏ Hallmark.
๏ Selection of cases.
๏ Selection of controls.
๏ Problems in control selection.
๏ Multiple control.
๏ Matching.
๏ When warranted?
๏ Case-Control Study based in a defined cohort.
๏ Measurement of exposure.
๏ Analysis & Interpretation.
๏ Bias
๏ Adv & disadv.
๏ Examples
4. DESIGN OF A CASE-CONTROL
STUDY
First Select
CASES CONTROLS
(with disease) (without ds)
THEN MEASURE PAST EXPOSURE
Were Exposed a b
Were Not exposed c d
Total a+c b+d
Proportions Exposed a b
a+c b+d
5. HALLMARK
๏ It begins with people with the disease(CASES) &
COMPARES them to people without the
ds(CONTROLS).
6. SELECTION OF CASES
๏ Definition of a CASE:
i)diagnostic criteria & stage of the ds ,if any.
ii)eligibility criteria-preferably incident cases.
๏ Sources:
1.hospital-based
2.population-based
๏ Case verification
๏ Exclusion criteria
7. SELECTION OF CONTROLS
๏ Who is a CONTROL
๏ Sources:
1.hospital-based
2.population-based
i)school rosters
ii)selective service lists
iii)insurance company lists
iv)neighbourhood
v)best friend
vi)spouse/ sibling
vii)dead
8. USE OF MULTIPLE CONTROLS
๏ Same types:
-increases the power of the study(only upto
1case:4control)
๏ Different types:
-for exploring alternate hypotheses
-for taking account possible potential biases like recall
bias.
9. MATCHING
๏ Def: The process of selecting the controls so that they are similar
to cases in certain characteristics ,such as age,race,sex,socio-eco.
status & occupation.
๏ Types:
1.Group/ Frequency
2.Individual/ Matched pairs.
๏ Problems:
a)practical-difficulty in finding a control
b)conceptual-cannot study that characteristic
c)unplanned matching
d)overmatching
10. WHEN A CASE-CONTROL STUDY
IS WARRANTED
๏ Often the 1st step in determining whether an exposure
is linked to an increased risk of ds.
๏ Rare ds.
11. CASE-CONTROL STUDIES BASED IN
A DEFINED COHORT
Defined Initial Data &/or Sr.,Urine or
Cohort Other specimens
YEARS
Develop
Disease Have Not developed
disease
Sub-gr
CASES selected as
CONTROL
12. ๏ Nested Case-Control Studies:
-controls are a sample of individuals who are at risk for the
ds at the time of each case of the ds develops.
-cases & controls are matched on calendar time & length
of follow-up.
๏ Case-Cohort Studies:
-controls are randomly chosen.
-Adv: possible to study different diseases.
13. ADVANTAGES OF EMBEDDING A
CASE-CONTROL STUDY IN A
DEFINED COHORT
๏ Recall bias-eliminated.
๏ Abnormalities in biologic characteristics like lab
values โrisk factors.
๏ More economical.
๏ Less lab burden.
๏ Greater comparability b/w cases & controls.
14. MEASUREMENT OF EXPOSURE
( IN PRICESLY THE SAME MANNER BOTH FOR
CASES & CONTROLS .)
๏ Interviews
๏ Questionnaires
๏ Past records of cases
-hospital records
-employment records etc.
15. ANALYSIS
๏ Exposure rates among cases & controls to suspected
factor.
๏ Estimation of ds risk asso. With exposure( ODDS
RATIO).
16. EXPOSURE RATES:
๏ Direct estimation.
A Case-Control Study of Smoking &Lung cancer.
Cases Controls
(with Lung cancer) (without lung cancer)
Smokers(<5 cigarettes/d) 33 (a) 55 (b)
Non-smokers 2 (c) 27 (d)
Total 35 (a+c) 82 (b+d)
Exposure rates:
. Cases= a/a+c= 33/35 =94.2%
. Controls= b/b+d =55/82 =67%.
17. ODDS RATIO
(RELATIVE ODDS/ CROSS-PRODUCTS RATIO.)
๏ A measure of the Strength of the association b/w risk
factor & ds.
๏ Def:
OR= Odds that a case was exposed
Odds that a control was exposed
=
a/c
=
b/d
ad
bc
=
Product of 2 cells that SUPPORT the hypothesis
Product of 2 cells that NEGATE the hypothesis
18. ๏ Interpretation:
OR=1 (exposure is not related to the ds)
>1 (+ly related)
<1 (- ly related).
๏ OR is a good approximation of RR when:
-cases studied are representative of those with the ds.
-controls studied are representative of those without the
ds.
-ds being studied does not occur frequently.
19. ๏ CALCULATING โORโ IN AN UNMATCHED CASE-
CONTROL STUDY:
OR= ad
bc
๏ CALCULATING โORโ IN A MATCHED PAIRS CASE-
CONTROL STUDY:
CONTROL
Exposed Not exposed
C Exposed p q
A
S
Not exposed r s
E
OR=Ratio of discordant pairs=q/r.
20. BIAS
๏ Bias due to confounding.
๏ Recall bias.
๏ Selection bias.
๏ Berkesonian bias.
๏ Interviewerโs bias.
21. ADVANTAGES
๏ Relatively easy.
๏ Rapid & inexpensive.
๏ Rare ds.
๏ No risk to subjects.
๏ Allows study of several different aetiological factors.
๏ Risk factors can be identified-----๏
rational prevention & control programme
๏ No attrition problems.
๏ Minimal ethical problem.
22. DISVANTAGES
๏ Recall bias.
๏ Selection of appropriate control gr. โmay be difficult.
๏ Cannot measure incidence.
๏ Cannot distinguish b/w causes & asso. Factors.
๏ Not suited to evaluation of Tt/ Px of ds.
๏ Concern about representativeness of cases & controls.
23. EXAMPLES
MATERNAL DES THERAPY & ADENOCARCINOMA
OF VAGINA IN FEMALE OFFSPRINGS:
CASES CONTROLS SIGNIFICANCE
(8) (32) LEVEL
Maternal age 26.1 29.3 n.s.
Maternal smoking 7 21 n.s.
Antenatal radiology 1 4 n.s.
Oestrogen exposure 7 -- P< 0.00001
24. OCP & THROMBOEMBOLIC DISEASE:
CASES(Venous CONTROLS
thrombosis &
pul.embolism)
(84) (168)
% who used OCP 50 14
THALIDOMIDE TRAGEDY:
CASES(46) CONTROLS(300)
Thalidomide exposure in 41 --
early pregnancy