ANALYTICAL

EPIDEMIOLOGY
CASE CONTROL STUDY
EPIDEMIOLOGY: The study of the distribution and
determinants of health-related states or events in
specified populations, and the application of this study for
the control of health problems.

A wide variety of definitions by various epidemiologist

are available but all of them have 3 components in
common.
They are;
1) disease frequency,
2) distribution of disease,
3) determinants of disease.
ANALYTICAL EPIDEMIOLOGY:

2nd major type of epidemiological studies after

Descriptive epidemiology.

Analytical epidemiology is ‘testing a specific

hypothesis – a relationship of a disease to a
supposed cause, by conducting an epidemiological
study that relates the exposure of interest to
the disease of interest’.

All analytical studies must begin with a clearly

formulated hypothesis. It must be quantitative
and specific and must predict the relationship of a
risk factor to a disease.
Here the subject of interest is the individual
with in the population but not the entire
population. But at the end inference is to the
whole population from which the subjects are
selected.

Analytical studies consists of 2 types :-

a) case control study b) cohort study.

From each of these one can determine,

a) whether or not a statistical association exists
between a disease and a suspected risk factor.
b) if exists, the strength of the association.
 Case control study:
(Retrospective study)
A case control study is “an enquiry in which
groups of individuals are selected in terms of
whether they do (cases) or do not (controls) have
the disease of which the etiology is to be
studied and the groups are then compared with
respect to existing or past characteristics,
judged to be of possible relevance to the etiology
of the disease”.

For e.g., testing the connection between lung

cancer and prior cigarette smoking or between
congenital malformations and maternal rubella
during pregnancy.
CASES

CONTROLS
The case control methods has 3 distinct features,
1) both exposure and disease has occurred before the
start of the study.
2) the study proceeds backwards from effect to cause
3) it uses a control or comparison group to support or
refute an inference.
A case control study involves 2 populations-
cases and controls;
and the unit of study is the individual rather than
the group.
Case control studies are basically comparison studies.
Here the cases are compared with the controls with
respect to known confounding factors like age, sex,
occupation etc.
 It is illustrated in a 2 2 table, ex in a study to test

+
the hypothesis ‘cigarette smoking causes lung cancer’,
the investigation begins by assembling a group of lung
cancer cases (a+c), and a group of suitably matched
controls (b+d).

Suspected or Cases control

risk factor (Disease present) (Disease absent)

Present a b
absent c d
a+c b+d
If the frequency of smoking in cases i.e a/(a+c) is higher
than in controls i.e b/(b+d), an association is said to exist
between smoking and lung cancer.
BASIC STEPS:
1) Selection of cases and controls,

2) Matching,

3) Measurement of exposure,

4) Analysis and interpretation.

1) SELECTION OF CASES AND CONTROLS:
SELECTION OF CASES: Criteria:
Diagnostic criteria: criteria which provide clear
and reproducible applications of definition and those
which provide manifestationally more homogeneous groups
of cases are preferable. e.g., the diagnostic criteria and
stage of the disease have to be included in the study.
Eligibility criteria: newly diagnosed cases with
in a specified period of time only are eligible.
Sources of cases:
 Hospitals: most convenient,
Single or network of hospitals
over a specified period of time
and area.
More common, easy, inexpensive.
 General population: cases occurring
within a defined geographic area during
a specified period of time.

It is laborious and involves special efforts to locate and

obtain the data from all individuals from a chosen
population.
Advantages like,
1) It avoids the bias arising from the selective factors
that guide affected individuals to a particular medical
care or a physician.
2) It allows computation of rates of the disease in the
total Population related to the etiologic factor under
study.
 SELECTION OF CONTROLS:
Criteria: 1) Controls must be free from the disease under
Study ,similar to the case in all other aspects (age, sex,
occupation, social status etc).
2) Information on study factor must be obtained from the
controls in a manner similar to that obtained from cases.
3) Make sure to select or match the controls similar to
the cases with respect to certain confounding factors.
4) Controls and cases should be selected from the same
population (so that confounding factors like residence,
ethnic group, socioeconomic status can be over come).
5) Practical and economic considerations: greatest amount
of information would be obtained if data were obtained
for all possible controls in the selected source, but
economic considerations usually limit these, therefore
they have to be balanced.
SOURCES OF CONTROLS:

Hospitals: other patients in or attending

the same hospital.
e.g., in a study of cancer cervix, the control
group may comprise patients with cancer
breast or stomach or other non-cancerous lesions.
Commonly used, easy, inexpensive.

Relatives: spouse and siblings:Advantage

is that they are similar in ethnic and social
Background and will eliminate certain spurious
associations.
But sibling controls are not suitable
where genetic conditions are studied.
Neighborhood: persons living in same locality,

working in same factory, children

attending the same school.

General population: controls are obtained from

individuals of a defined geographic area, randomly. It is

of use only if cases are selected in similar manner, and is
also expensive and time consuming.
MATCHING:
It is the selection of controls in such a way
that they are similar to the cases in all aspects
except for the presence of disease.
This is because certain pertinent factors may
influence the outcome of disease and if not
properly matched may confound the result.
A CONFOUNDING FACTOR is the one which is
associated with both exposure and disease and is
distributed unequally in the case and control
groups. Examples, age, sex, occupation, social status.
And therefore it must be seen that the control group
has the same distribution as the cases with respect to
these confounding variables
 In some cases the confounding factor though associated
with the exposure, it can independently become a risk
factor for the disease under study.

Therefore MATCHING is defined as “ the process by

which we select controls in such a way that they are
similar to cases with regard to certain pertinent selected
variables (age) which are known to influence the outcome
of disease which, if not adequately matched for
comparability, could distort or confound the results”.
Certain factors to be considered:
1)confounding factors must be taken account of: these
are variables associated with both disease and study
variable, like in studying relation between breast cancer
and lactation, parity is the confounding factor as it is
related with both breast cancer and lactation. Cases
and controls of same parity and different frequency
and duration of lactation should be selected.
2)the unusual distribution of cases with respect to a
particular confounding factor, the less overlap, the less
efficient in controlling the confounding factors. Eg in
studying about a disease like stroke, controls under the
age of 50 will be waste and there would be only few
controls over the age 70. To over come this controls of
other sex should also be taken .
3)the cost of study must considered, if less multiple
controls for a single case should be selected and
difference in confounding variables should be taken into
account. And if the cost is high equal number of cases
and controls should be selected which are of perfect
match .
 OVER MATCHING:

Matching may sometimes be harmful. There are 2 such

situations where they have to be avoided:
1)variables intermediate in causal pathway between the
study factor and the disease.eg, if smoking altered blood
cholesterol which in turn was causally associated with
cardiovascular disease, smoking should be considered as
the cause. If cases and controls are matched on
cholesterol level, then no association with smoking should
be there.
2)factors should not be matched that are related to
suspected factors alone but not to the disease.eg, if
contraceptive use were related to religion which is not
associated to the disease, it would be inappropriate to
match on religion.
E.g.,
AGE could be a confounding factor,
in studying the relationship between the usage
of oral contraceptives and breast cancer, if
women taking OC are younger than those in
comparison group, they would be definitely at a
lesser risk, as the disease risk also increases
with increasing age.
So this confounding effect of age should be
neutralized by matching both the groups such
that they have equal proportion of each age
group.
MEASUREMENT OF EXPOSURE:
Definitions and criteria about exposure are as
important as those used to define cases and
Controls.
Information about exposure should be collected
in the same manner for both cases and controls.
This is done by interviews, questionnaires,
studying past records ( hospital, employment
records).
When case control studies are being used to
test associations, it is most important to rule
out any question of bias or systemic errors.
 ANALYSIS:
It is used to find out,
1) exposure rates
2) estimation of risk.

Exposure rates:
Here a direct estimation of the exposure rates
or the frequency of exposure to a suspected
factor in disease and non-disease groups is made.
E.g., a case control study of smoking and
lung cancer
 CASES CONTROLS
with lung cancer without lung
cancer
SMOKERS
>5/day 33(a) 55(b)
NON
SMOKERS 2(c) 27(d)

TOTAL 35(a+c) 82(b+d)

EXPOSURE RATES:
a. Cases = a / a+c = 33/35 = 94.2%

b. Controls= b / b+d = 55/82 = 76.0%

 Estimation of risk:
Estimation of disease risk associated with
exposure.
It is obtained by an index called
ODD’S RATIO ( cross product ratio):
Measure of the strength of the association between risk
factor and outcome
 It is closely related to relative risk and its
derivation is based on 3 assumptions;
a. rare disease

b. the case must be representative of those

with the disease, and

c. the control of those without the disease.

disease
yes no Odd’s ratio = ad
Exposed a b bc
c d 33 * 27
Not exposed = = 8.1
55 * 2
Here, smokers have a risk of developing lung cancer 8.1
Times more than that of non-smokers.
BIAS:
Systemic non random deviation of result and
inferences from the truth or process leading to
such deviation. or
Any systemic error in the determination of the
association between the exposure and disease.
The relative risk estimation may increase or
decrease as a result of the bias.
Bias reflects non-comparability between study and
control groups.
Many varieties are present, the common ones are:-
Bias due to confounding
Memory or recall bias
Selection bias
Berkesonian bias
Interviewers bias
1) Bias due to confounding:
2) Memory or recall bias: cases are more likely to recall
the existence of certain events or factors compared to
the controls. E.g., persons with MI are more likely to
recall certain habits or events than those who have not.
3) Selection bias: arises when cases and controls are
selected from hospital where they may not represent the
general population. It is best controlled by its prevention.
4) Berkesonian bias: it arises because of the different
rates of admission to hospitals for people with different
diseases.
5) Interviewer’s bias: occurs when the interviewer knows
the hypothesis and also knows who the cases are. This
eads him to question the cases more thoroughly than
controls regarding the positive history of the suspected
causal factor.
 ADVANTAGES DISADVANTAGES
1) easy to carry out. 1) problems of bias.
2) rapid, less expensive. 2) selection of control
3) require few subjects. group may be difficult.
4) no risk to subjects. 3) incidence cannot be
5) allows the study of measured.
several different 4) not useful for
etiological factors. evaluation of therapy &
6) risk factors can be prophylaxis.
identified. 5) cannot assess effects
7) no attrition problems. of matching variables.
8) ethical problems
minimal.
9) finds its major use in
chronic diseases and
rare diseases.
 Few examples of case control studies:
 Adenocarcinoma of vagina: the unusual occurrence
of adenocarcinoma of vagina in 7 young women between
the age of 15-22 years born in one Boston hospital
between the years 1966 and 1969 and a 8th case in
another Boston hospital, lead to suspicion and is
investigated using case control studies. Matched
controls were identified from birth records of the same
hospital (controls born closest in time to the cases
are selected).

Information was collected by interviews regarding-

maternal age, maternal smoking , antenatal
radiology, diethyl stilbestrol exposure in fetal life.
7 out of 8 cases were exposed to DES in fetal life.
DES was given to mothers in the 1st trimester to
prevent miscarriages. None of the mothers of the control
group received DES.
 Oral contraceptives and Thromboembolic disease: in
1965 British committee on safety of drugs had received
249 reports of adverse reactions and 16 reports of
death in women taking OC’s.
Vassey & Doll reported their findings in 1969. They
interviewed women admitted to hospitals with venous
thrombosis or pulmonary embolism with out medical
cause and compared to other women admitted due to
other reasons.
No. % using OC
50% of cases used OC, compared Cases 84 50
to 14% 0f controls.
So it is confirmed that taking controls 168 14
OC pills has a high frequency
of developing thromboembolism Relative risk= 6.3
than others.