CASE CONTROL AND COHORT

STUDY
 CONTENT
 DEFINITION
 TYPES OF STUDY
 ANALYTICAL STUDIES
 CASE CONTROL STUDY
 VARIANTS OF CASE CONTROL STUDY
 SUMMARY
 COHORT STUDY
 DIFFERENCE
 SUMMARY
 REFERENCE
 The most conventional definition
of epidemiology is "the study of the
distribution and determinants of
health-related states or events in
specified populations, and the
application of this study to control
health problems." (John M.Last,1988)
 Experimental Observational

RCT Non RCT

Analytical Descriptive

Ecological Cross-sectional Case-control Cohort

 In analytical studies , the subject of interest is the
individual within the population.
 The object is not to formulate but to test the
hypothesis.
 To evaluate an association between exposure and
disease.
 Analytical studies focuses on the magnitude of the
association between the exposure and the health
problem under the study.
 Unit of Study: Cases/Control(Individuals)
 Study Question : What had happened 
 Direction of Inquiry: O
E
 Study Design:

Exposed Exposed

Control
Not
 Cases
Exposed Not
Exposed
 A case–control study is an observational study in which subjects
are sampled based upon presence or absence of disease and
then their prior exposure status is determined.

 Distinct feature:

a.Both exposure and outcome (disease) have occurred

before the start of the study.

b.The study proceeds backwards from effect to cause.

c.It uses a control or comparison group to support or

refute an inference.
 DEFINITION
Case : A person in the population or study group
identified as having the particular disease,
health disorder or condition under
investigation.
Control: Person or persons in a comparison
group that differs, in disease experience (or
other health related outcome) in not having
the outcome being studied.
RISK CASES CONTORLS
FACTOR (Disease (Disease
S Absent)
Present
)

PRESENT
a b
ABSENT
c d
Total
a+c b+d
 Selection of cases and controls.
 Matching.
 Measurement of exposure and
 Analysis and interpretation.
 Study begins with cases, i.e. the patients in whom
the disease has already occurred.
 Patients with the disease in question (cases) were
enquired for all the details of their exposure to the
suspected cause.
 The new cases, which are similar clinically,
histologically, pathologically and in their duration
of exposure (stage) will be chosen to avoid any
error and for better comparison.
Definition of case: it involve two specifications-
(i)Diagnostic criteria :Enunciate clear cut
diagnostic criteria for the disease of interest. As far
as possible use criteria given by expert bodies.
(ii)Eligibility criteria : It is always advisable to take
the incident cases since the prevalent cases might
have changed their exposure status due to medical
advice etc.
Sources of Cases
 Hospitals.
 General population:
 CONTROL SELECTION
• Should the controls be similar to the cases in all
respects other than having the disease? i.e.
COMPARABLE
• Should the controls be representative of all non
diseased people in the population from which
the cases are selected? i.e. REPRESENTATIVE
Population-Based Hospital-Based
Source population is better defined Subjects are more accessible

Easier to make certain that cases Subjects tend to be more cooperative

and controls derive from the same
source population

Exposure histories of controls Easier to collect exposure

more likely to reflect those information from medical records
of persons without the and biological specimens
disease of interest
 Defined as one which is associated both with
exposure and disease and is distributed
unequally in study and control groups.

Confou nder
(i)Associated with the exposure of
interest.
(ii) Related to the outcome of the
interest.
(iii) It should not be in the direct chain
or link between the exposure and
Exposure outcome outcome
Hypothesis:Whether consumption of alcohol is a risk factor for oral CA.

F1ni 0d0ni cagsmsaoyf obreaflaCslAea:Dnude1t0o0 thheeal“thhiyddsuebnj”eectfse

f wcteroef
atoskbeadccroeguasredibnegcause
History of
the history of alcohol pacaeuonosncpesu
Oral Cc erelofmow
OprahratilolcoanCdrndani cunke
rincargecl,prowahehT
sottolh
1tae5arerylone
aelasros.often
Alcohol Present Absent
Dissecting
drinks or not. hypothetical data into two strata Odds ratio
the ones who
= (a x d / b x c)
Tobacco Users also use tobacco; and tobacco use is
Non-Tobacco Usersitself a direct
= (80 x 80)
Present 80 20 100 (20 x 20)
= 16

Absent 20 80 100

Stratum OR=60x5/20x15 =1 Stratum OR= 5x 60/15 x 20=1 Total

100 100 200

Conclusion :Both the strata OR falls to 1 i.e. there is no risk of

Risk of cancer
getting from
oral cancer
alcoholisafter
16 times higher
adjusting forif the
a person
effectdrinks alcohol
of tobacco
.
 Randomisation: If a group of subjects is divided
into two , using “random allocation” (syn.
Randomization) the 2 groups will be similar
to each other in all respect.
 Restriction: the subjects having the particular
confounding variable(s) are not taken up at all.
 Matching
 Defined as the process by
which we select controls in
such a way that they are
similar to cases with regards
to certain pertinent selected
variables (e g. age, sex,
occupation, social status etc. )
which are known to
influence the outcome of the
disease.
Advantages Disadvantages
May increase the precision of case- May be time-consuming and
control comparisons and thus allow expensive to perform.
a smaller study.

The sampling process is easy to Some potential cases and controls

understand and explain. may be excluded because
matches cannot be made.

If analyzed correctly, provides The matched variables cannot be

reassurance that matched evaluated as risk factors in the
variables cannot explain case- study population.
control differences in the risk
factor of interest.
 Information about the exposure should be obtained
in precisely the same manner for both cases and
controls.
 This may be obtained by the interviews, by
questionnaires, or by studying past records
of cases such as hospital records,
employment records.
The final step is Analysis:

 Exposure rate among cases and controls to

suspected factors.
 Estimation of the Disease risk associated with
exposure (Odds ratio).
 CASES CONTROLS
(WITH LUNG CANCER (WITHOUT LUNG
CANCER)

SMOKERS 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 = 67.0%
 This shows frequency rate of lung cancer is
definitely higher among smokers than among
non- smokers.
 The chance of something happening can
be expressed as a risk and/or as an odds

Risk = the chances of something

happening the chances of all things
happening

Odds = the chances of something happening

the chances of it not happening
Example-1: If we choose a student randomly from your
class of say 9, how likely is it that you will be chosen?

Risk (probability) = 1/9 = .111

Odds = 1/8 = .125

 Example-2: Among 100 people at baseline, 20

develop
influenza over a year.

The risk is 1 in 5 (i.e. 20 among 100) = .2

The odds is 1 to 4 (i.e. 20 compared to 80) = .25
 Measure of strength of association between risk factors
and outcome.
 Odds ratio= P/1-P, P= Probability
 The odds ratio is also known as the cross-products
ratio
 Based on 3 assumption:

1.Disease being investigated must be relatively rare. In fact

majority of the chronic disease have a low incidence in the
general population.
2.The cases must be representative of those with the
disease.
3.The controls must be representative of those without the
disease.
Cohort study Case control study
 Odds Ratio : ad/ bc
33 X 27/55 X 2 = 8.1
 Odds ratio is a Key Parameter in the analysis of case
control studies.
 It interprets that odds of cases being exposed are
so many times higher compared to the odds of
controls being exposed.
 In our example risk of lung cancer due to smoking is
8.1 as compared to non smoking.
 Selection Biases
 Berksonian Bias : The probability of admission to hospital or detection of
the outcome (disease) may be more among the cases simply because of
the exposure.
 Selection of inappropriate Cases or Controls : Cases or controls who do
not have adequate chance of exposure.
 Self selection Bias : Patients who are admitted to a particular hospital and
hence taken as cases may be systematically very different from most of the
patients with the disease but who are not admitted to that hospital, as
regards the exposure status.
 Survivorship Bias : Case control study generally takes the patients who
are living. Cases who have died are generally not taken and these may be
systematically very different from living case as regards the exposure status
 Selection of wrong control group : Controls who are not from the same
source population from where the cases have come; selection of close
friends of cases - since they would in general have the same behavioural
factors as cases (birds of a feather flock together ), example of condom use
and STDs.
 Information (measurement) Biases
 Recall bias : Cases who are suffering from a disease are likely to recall
much more as regards their exposure (example on congenital malformation
and exposure to X - rays).
 Observer bias : If observer is aware of the case - control status, he/she may
subconsciously tend to ask much more from cases.
 Combines the advantages of a cohort and a case
control study.
 Firstly , the study becomes inexpensive and take
care of the logistics.
 Secondly, we can calculate the incidence of the
disease which would not have been possible in
a usual case control study.
 Thirdly, the problem of recall bias and that the
controls may be from a different source population
than cases (which occur in case control study)
have been prevented.
Hypothesis : High serum lithium levels are a cause of subsequent mental illness.

Take a cohort of say 1000 persons

who are free of mental disease, collect their
blood sample, preserve them in cold
storage

Watch for 15 - 20 years

20 randomly selected
samples of those
20 cases of mental who Rest of the cohort is
disease(cases) continously folowed
have not developed
mental illness
(controls)

analyse these 40
samples for serum
lithium and Rest of the cohort ris
make comparisons continously folowed
between the two
groups
Advantages:
 Recall bias is eliminated.
 If abnormalities in biologic characteristics such as
laboratory values are found, because the specimens were
obtained years before the development of clinical disease, it
is more likely that these findings represent risk factors or
other premorbid characteristics than a manifestation of
early, subclinical disease. When such abnormalities are
found in the traditional case-control study, we do not know
whether they preceded the disease or were a result of the
disease.
 More economical to conduct.
 It is possible to study different diseases (different sets
of cases) in the same case-cohort study using the same
cohort for controls.
Advantages
Efficient for the study of rare
diseases
Efficient for the study of
chronic diseases
Tend to require a smaller sample
size than other designs
Less expensive than alternative
designs
May be completed more rapidly
than alternative designs
Disadvantages
Risk of disease cannot be
estimated directly
Not efficient for the study of
rare exposure
More susceptible to selection
bias than alternative designs
Information on exposure may be
less accurate than that available
in alternative designs
 • Well suited for diseases which have a long latent
period(e.g. cancers, AIDS, MI, CVA etc.)
• Well suited for an outcome which is ‘rare’
• Well suited for conditions in which medical
care is usually sought
• Helps in examining multiple etiologic factors - once we
have the cases of the disease, we can take history of all
the factors that we feel may be risk factors
• Reasonably good for diseases that have a “relatively
Review of research question and confirm th at case -
rapid onset” and are usually hospitalised (e.g. most of the
control study is the right design. acute infections; injuries etc.)

Specify the total population and actual (study)

population.

Specify the major study variables

(exposure,outcome,confounding factors) and their
‘scales’ of measurement(dichotomous etc)

Calculate the sample size.

Specify the selection criteria of cases

Specify the selection procedure for controls

Specify the procedures of measurement and

specially take care to ensure validity and
reliability

Do a pilot study on 5 to 10 cases and

controls

Conduct the study

Analysis of data
 Forward looking ,incidence , longitudinal,
prospective study or follow up study
 Cohort = Group of people who share a common
characteristic or experience within a defined
time period(age, occupation ,exposure etc).
 Cohort study: Cohort studies are observational
studies in which the investigator determines the
exposure status of subjects and then follows them for
subsequent outcomes
 Quantified with relative risk/incidence
rates/attributable risk
 Cohorts are identified prior to the appearance of the
disease under investigation.
  In cohort study the exposure has occurred , but
the disease has not.
Cohort With Without Total
disease disease exposure
Exposure a b a+b

(etiologic
factor)
Non- Exposure c d c+ d

a/(a + b) - Incidence of disease in exposed

c/( c + d)- Incidence of disease in non exposed
if a/(a + b )> c/ (c + d) It would suggest that the disease and
suspected
 Cohorts must be free from the disease under study.
 Study and control group must be easily susceptible
to the disease under study.
 Both the groups must be comparable in respect to
all the possible variables which may influence the
frequency of the disease.
 The diagnostic and eligibility criteria of the disease
must be defined before hand.
 Groups are then followed , under the
same identical conditions, over a period of
time to determine the outcome of the
exposure.
 Retros
Prospective
pective
2000 define population 1987

Non randomization

2010 1997
exposed Non exposed

2007
Not diseased diseased Not diseased
2020 diseased

1987 2007 2017

combined
 SELECTION OF STUDY SUBJECT
 OBTAINING THE DATA ON THE EXPOSURE.
 SELECTION OF THE COMPARISION GROUP.
 FOLLOW UP
 ANALYSIS
 Special Exposure Groups (e.g. radiologists for
studies on effect of radiation; ANC cases having
PIH for studying the outcome of pregnancy, etc.)
 Cohort defined on basis of geographical or
administrative boundaries (e.g. people living in a
given state or district like Framingham heart
study). The special advantage of such cohort is that
the same group will give an exposed as well as
unexposed (comparison) cohort.
 Groups offering special resources (e.g. all
registered doctors can be followed up for
development of IHD after recording their
physical activity levels.
DATA External Sources Internal Sources

Exposure Hospital records Questionnaires,

physical examinations,
and/or blood tests,
other diagnostic tests

Event Disease registries, Questionnaires,

death certificates, physical examinations,
physician and hospital and/or blood tests,
records other diagnostic tests

Confounder Hospital records Questionnaires,

registries physical examinations
 
Exposed
Internal and
Control non-
Group  General Population: If none of the
exposed in the same above comparison is available than
the mortality experience of the
Study population
exposed group is compared with
(Framingham study) the mortality experience of the
 Minimise the differences between

exposed and non- exposed

general population in the same
geographic area as the exposed
•
 External Control Grouppeople.
•  When information  E.g. comparison of frequency of
on degree of exposure cancer among uranium mine
is not available chose workers with the rate in general
another group, another population in same geographic
cohort (smokers and area.
 One of the problem in cohort Death.studies is the regular
Procedures: Change of residence.
follow up of the participants.
Migration.
the start ofWithdrawal
 Therefore , at Periodic the
etc.
study, from occupation
methods
should be devised depending
medical
examination of
upon the outcome to
Routine
surveillance of
be determinedeach (morbidity
member of
the cohort.
or Death)
death records.to obtain the

data assessing the outcome.

Mailed
questionnaires, Review
telephone calls, physician and
periodic home hospital records
visits.
 Absolute comparison
 Riskdifference
I exposed - I unexposed
Measures public health problem caused by
the
exposure

 Relative comparison
 Relative
Risk
 Odds Ratio

RR=I exposed / I unexposed

Measures strength of an association
 DATA ARE ANALYSED IN TERMS OF
a) Incidence rates of outcome among exposed
and non- exposed.
b) Estimation of risk.
(i) relative risk
(ii) attributable risk
Cigarette Develop Did not Total Incidence
smoking CHD develop
CHD

Yes 70 6930 7000 70/7000

(a) (b) (a + b ) =10 per
1000

No 3 2997 3000 3/3000

(c) (d) (c +d) =1 per
1000
 R. R = incidence of disease
(or Death) among exposed
incidence of disease (or
Cigarette
Death) among non-
Develop
CHD
Did not
develop
Total Incidence

smoking CHD
Yes 70
exposed
6930 7000 70/7000
(a) (b) (a + b )
0.01
No 3 2997 3000 3/3000
(c) (d) (c+d) .001

RR= a/a+b = 70/7000 = 10

c/c+d 3/3000
 RR=1 = No association between
exposure
incidence andbetween
rates are identical disease
groups
 RR=> 1 = Positive association
 exposedgroup has higher incidence than non-
exposed group
 RR=< 1 = Negative association or protective effect.
 non-exposed group has higher incidence than exposed
or exposed group has lower incidence than non-
exposed e.g. RR 10% / 20% = 0.5 it would
indicate that if one smokes, the risk of getting IHD is
10%; on the other hand if one does not smokes, the
risk is 20%.
Smoking thus reduces the risk of getting IHD by half.
 Risk difference =I exposed- I non exposed
 Attributable risk percent
 Population attributable risk percent.
Incidence

Iexposed – Iunexposed

Exposed Unexposed

= ( Iexposed-I unexposed)x 100

I = Incidence
Iexposed
Ciga
rett
Develop Did not Tota Incidence
e CHD develop l
smo
Ye 70
CHD
6930 7000
king
s (a) (b) 70/7000
(a + b ) 0.01
No 3 2997 3000
The limitation of AR% is that it tells us
(c) 3/3000
the quantum of reduction
(d) ( c +would
in the disease that d ) be achieved
.001
in the “exposed” group if
Attributable risk in our example:A “exposure” was given up by them.
 It indicates to what extent dise However, it does not tell us
R=(
exposure. If smoking is given u about .01-.001/.01)x
the 100=90%
reduction that will occur in
CHD among smokers. the “total population”
ase under study can be attributed to
p then there will be 90% reduction in
  Population attributable risk percent
 Proportion of disease in the study
population that could
be eliminated if exposure is removed

Incidence in total population – Incidence in

unexposed incidence in total population

{(73/10,000)-(3/3000)}/73/10,000=.86 PAR%=86%
 Measurement (Ascertainment) bias : For obviating this, inform all
subjects of both groups well in advance of the dates and timings of
medical examination and ensure that both the groups are examined
by observers who have similar type of training and using similar
type of instruments and techniques.
 Observer bias : This occurs because the investigator is aware about
the fact as to which subject is ‘exposed’ and who is not exposed. For
obviating this, if possible, ‘blind’ the observer to the exposure status,
the details of exposure being known only to another co - worker who
is, himself, not making any observation regarding ascertainment of
outcome.
 Cross over bias : This may happen because those having the
exposure (e.g. smokers) may cross over to the non exposed group (i.e.
become non smokers) and vice versa. Periodic evaluation of both the
groups as regards level of exposure, making record entries and
subsequent adjustments in the data analysis can help overcoming this
problem.
 ‘Loss to follow up’ bias : Some subjects in any case are likely to be
lost to follow up / drop out.
 • Large No. of population.
 Incidence can • Very lengthy- takes very long
be calculated time to complete.
 Several possible • Certain administrative.
outcomes related • Loss of experience staff.
to exposure can • Loss of funding.
be studied
• Extensive record keeping.
simultaneously.
 Cohort studies provide  Selection of comparison group-
limiting factor
a direct estimate of
R.R  There may be changes in study
 Dose – response ratio methods or Diagnostic Criteria
of the Disease over the
can also be calculated. prolonged period.
 Cohort studies are expensive.
 The study may itself alter the
patients Behavior.
 Best-known cohort studies is the Framingham Study of
cardiovascular disease.
 Started in 1948.
 Framingham is a town in Massachusetts, about 20
miles from Boston.
 Residents between 30 and 62 years of age were
considered eligible for study.
 1971 enrolled a second generation of
participants.
 In April 2002, a third generation was enrolled in the
core study.
  Hypothesis:
: contd..
Incidence of CHD increases with age
 Hypertension develop CHD
 Elevated cholestrol is associated with ed CHD
Tobacco smoking and habitual use of alcohol
increased CHD
 Increased physical activity a/w with decreased
incidence of
• CHD
 Increased Body weight inceases incidence of
 

Resul contd...


ts:
1960s: Cigarette smoking Increased cholesterol and
elevated blood pressure obesity increases risk of
heart disease. Exercise decreases risk of heart
disease.
 1970s: Elevated blood pressure increases risk of
stroke. Postmenopausal women risk of heart disease is
increased compared with who are premenopausal.
 1980s High levels of HDL cholesterol reduce risk of
heart disease.
 1990s: Elevated blood pressure can progress to heart
failure. At 40 years of age, the lifetime risk for CHD
is 50% for men and 33% for women.
 contd...
 2000s “High normal blood pressure" increases risk of
cardiovascular disease (high normal blood pressure is
called prehypertension in medicine; it is defined as a
systolic pressure of 120–139 mm Hg and/or a diastolic
pressure of 80–89 mm Hg). Lifetime risk of developing
elevated blood pressure is 90%. Serum aldosterone
levels predict risk of elevated blood pressure. Lifetime
risk for obesity is approximately 50%.
 •
Where there is good evidence of association
between exposure and disease, as derived
from clinical observation and supported by
descriptive and case –control studies.
• When exposure is rare, but the incidence of
disease is high among exposed.
Specify the research question, objectives and
• When attrition of study population can be
background significance, confirm cohort s tudyminimized
is e. g. follow up is easy , cohort is
to be done stable.
• When ample funds are available.

Specify the variables of interest and their scales

of Measurement (Exposure variable, Outcome
variable, confounders)

Specify the exclusion criteria ( e.g.

like to restrict the study to males)

Calculate the sample size

Select the study cohort(Special Exposure Groups ,

on basis of geographical or
administrative boundaries)
Select the study cohort

Select the comparison cohort (Ext. group,Int.

group)

Specify the sampling procedure ( simple

random
or by systematic random sampling
method).

Exclude the disease or outcome of

interest in
both the exposed and unexposed
cohort groups
Obtain Data on all Potential
confounding factors

Consider matching (matching is not

important , if eligible then
frequency matching )

Follow up and ascertainment of

‘outcome’ of interest

Analysis
 Text book of PSM 19th ed by K. Park
 Lange Medical Epidemiology 4th by Raymonds
S Greenberg , Stephen R Daniels ,John William
Elley
 Epidemiology by Leon Gordis.
 Textbook of Public Health and community
medicine by Rajvir Bhalwar ,Rajesh Vaidya, Reena
Tilak
 http://en.wikipedia.org/wiki/Cohort_(statistics)