Epidemiological Study Designs And

Measures Of Risks
1
 Objectives of the Lecture
• To describe observational study designs

• To describe experimental study designs and

clinical trials

• Calculation and interpretation of measures of risk

STUDY DESIGNS
 Case stud.

Descriptive stud.

Surveys
Non intervention
stud.
(Observational)
Ecological
(correlational)

Analytical stud.
Cross-sectional
(Prevalence)

Study designs

Case con. Stud.

Randomized control (case-reference)
Trials (Clinical trials)

Intervention stud. Cohort stud.

(Experimental) (Follow-up)
Non-randomized trials
(Quasi-experimental
study)
 Observational Studies

• It allows nature to take its own course.

• We measure but not intervene.

 DESCRIPTIVE STUDY
DESIGNS
– Uses:
• Describes magnitude of the disease load in
terms of morbidity and mortality rates.
• Provides clues to disease etiology (formulation
of etiological hypothesis).
• Provides data for planning, organizing and
evaluating health services.
• Contribute to research by describing variation
in disease occurrence by time, place and
person.
A. Case studies
 Case studies

A case study describes in depth the

characteristics of one or limited number of
cases in its natural environment.
or
A detailed report by a physician of an unusual
disease in a single person.
 Case studies

• A case may be a patient, a health centre, a

village etc…
• Can provide useful insight into the problem
e.g. a new disease
• Common in clinical medicine, social
sciences, management and administration
etc..
 Case Studies Features:

• Should be well planned and data will be

collected thorough predetermined questions
• Should be flexible to deal with unexpected
situations
• No comparison group
• Population: unknown
• Select patient: (case report)
• or patients (case series) with disease of
interest
• Assessment: Describe clinical findings
• Analysis: Radiographs, lab reports, etc
• Interpretation: Special features of this
disease
• Example: “Normal plasma cholesterol in an
88-year-old man who eats 25 eggs a day”
[Kern J, NEJM 1991; 324:896–899]12
 Case studies

• Advantage:
• It permits a holistic approach to the problem
under investigation

• Disadvantage:
• Not representative
B. SURVEYS
 SURVEYS

• Use :
• To collect information on demographic
characteristics. Age, sex, education etc…
• To study characteristics on health related
variables. E.g. incidence rate, etc….
• To study attitudes, opinions and beliefs
 SURVEYS
Surveys answer the following questions:
WHEN IS THE DISEASE OCCURRING?
(TIME DISTRIBUTION)

WHERE IS THE DISEASE OCCURRING?

(PLACE DISTRIBUTION)

WHO IS AFFECTED?
(PERSON DISTRIBUTION)
 PROCEDURE

1. Define the problem under study.

2. Define the population under the study.
3. Describe the disease by TIME, PERSON and
PLACE.
4. Measurement of the disease.
5. Comparing with known indices.
6. Formulation of an etiological hypothesis.
 Analytical Studies

• Analysis of the relationship between health

status and other variables.
• It is to test hypothesis.
• Interested in individual and inference is to
population.
 Analytical Epidemiology
1. Ecological or correlation
2. Cross-sectional or prevalence
3. Case-control or case-reference
4. Cohort or follow-up
 A) Ecological Studies
• They look for associations between the occurrence
of disease and exposure to known or suspected
causes.
• The unit of observation is the population or
community.
• Often the information about disease and exposure
is abstracted from published statistics and
therefore does not require expensive or time
consuming data collection.
 Ecologic Studies
• Aggregates of individuals.
• Aggregates often defined by units:
geographic region, school, health care
facility.
• Does the overall occurrence disease in a
population correlate with occurrence of the
exposure.
• No individual data
• Examples of aggregate data:
• Disease rates (incidence, mortality)
• Birth rates
• “Exposure” data: smoking rates, geographic
residence,
• air pollution data, mean income, per capita
• consumption of saturated fats
 Ecologic Fallacy
• Grouped data do not necessarily represent
individual level data
Example: Fat intake and breast cancer rates
with countries as the unit of measurement
have consistently been found to be highly
correlated.
• But studies of individuals (cohort, case
control studies) have not found any
association with fat intake.
 Why?
• Possible reasons–countries with high fat intake are
more likely to have other risk factors associated
with breast cancer (i.e. late age at first pregnancy)
• Or-- within population variability is low, but inter-
population variability is high.
• i.e. Extreme example– if everyone in a country had
high fat intake, we would not be able to detect any
excess because there would not be any population to
compare them to with low fat intake
 Examples

• Ecological studies are useful for generation of

hypotheses, supporting hypotheses, or for intervening
at the population level.
• Rates of stomach cancer declined dramatically after
the advent of refrigeration in the 1930s–
• Supports studies showing risk of stomach cancer
increases with consumption of nitrates in preserved
foods (sausage, lunch meat etc)
• Smoking and lung cancer
• Oral cancer and snuff use in the KPK
 CROSS SECTIONAL STUDY
• Data collected at a single point in time

• Describes associations A “Snapshot”

• Prevalence

• Burden of Disease

• It is based on single examination of

cross-section of population at one point
of time.
 Cross-Sectional Study: Definition

• Exposure status and disease status are

measured at one point in time or over a
short period of time.
• No Follow-up.
• Prevalence studies. Comparison of
prevalence among exposed and non-
exposed.
 Cross-sectional Studies

• Select sample of individual subjects and

report disease prevalence (%)

• Can also simultaneously classify subjects

according to exposure and disease status to
draw inferences
– Describe association between exposure and
disease prevalence.

27
 Cross-sectional studies

• They are more useful for chronic illnesses,

e.g. hypertension.
• Cross-sectional studies save on time and
resources, but provide very little
information about natural history of disease
and incidence of illness.
• If the sampling methodology is accurate,
results can be projected to the entire
population.
 Cross-Sectional Studies
• Exposure and outcome status are determined at
the same time
• Examples include:
– Behavioral Risk Factor Surveillance System
(BRFSS) - http://www.cdc.gov/brfss/
– National Health and Nutrition Surveys (NHANES)
- http://www.cdc.gov/nchs/nhanes.htm
• Also include most opinion and political polls
 Cross-sectional studies

• Prevalence measured by conducting a survey of the

population of interest e.g.,
– Interview of clinic patients
– Random-digit-dialing telephone survey

• Mainstay of descriptive epidemiology

– patterns of occurrence by time, place and person
– estimate disease frequency (prevalence) and time trends

• Useful for:
– program planning
– resource allocation
– generate hypotheses

30
 Examples

– Prevalence of Asthma in School-aged Children

in Lahore

– Trends and changing epidemiology of hepatitis

in Pakistan

– Characteristics of teenage smokers in Multan

– Prevalence of stroke in Gujranwala

31
 Cross-sectional: Advantages

• Usually use population-based samples,

instead of convenient samples.
• Conducted over short period of time
(Quick)
• Relatively inexpensive
 Cross-sectional: Disadvantages
• Difficult to separate cause from effect, because
measurement of exposure and disease is conducted at
the same time.
uncertain temporal relationships
• A persons exposure status at the time of the study may
have little to do with their exposure status at the time
the disease began.
– survivor effect
– low prevalence due to
• rare disease
• short duration
 Case Control Studies
They are comparison studies
To determine
• Whether or not a statistical
association exist
• And its strength/Estimate of risk
 Time

Direction of inquiry

Start with
Exposed

cases
Not exposed

population
Exposed

controls
Not exposed
 Distinct Features
1. Both exposure & outcome have occurred
before the start of the study
2. The study proceeds backwards from effect
to cause/retrospective studies
3. Uses controls to support or refute an
inference
• Two populations (cases & controls)
• Select subjects based on their disease status.
A group of individuals that are disease
positive (the "case" group) is compared with
a group of disease negative individuals (the
"control" group). Both groups should come
from same populations.
• The unit is individual
• The focus is on the disease
• Because they are comparison
studies, cases and controls must be
comparable with respect to known
confounding factors (age, sex, social
status, occupation….etc)
 The Basic Design Is 2x2
Contingency Table
Suspected or Cases Controls
risk factor disease present disease absent
Present a b

Absent c d

(a+c) (b+d)
• If the frequency of risk factor (smoking)
a/(a+c) is higher in cases (lung cancer) than
in controls b/(b+d), an association is said to
exist between smoking and lung cancer.
 Basic Steps
1. Selection of case & controls
2. Matching
3. Measurement of exposure
4. Analysis & interpretation
 1. Selection of Cases &
Controls
• Proper selection is crucial
• Avoid selection bias
• Conducting of more than one study in
different geographical areas increases
the validity of the inferences
 (A) Selection of Cases
1. Definition of cases
(i) diagnostic criteria
(ii) eligibility criteria
2. Sources of cases
Hospitals - general population
 (B) Selection of Controls
• More difficult ( sub-clinical form)
• Free from the disease under study
• Similar to cases as possible
• Sources of controls
Hospitals
same hospital
different illness
selection bias is common
Relatives
unsuitable in genetic conditions
Neighborhood
same locality
factory
school
General population
from defined geographic area
must reflect the population free
from the disease
 How many controls
• One to one in large no. of
cases

• 2, 3 or 4 to one study subject

in small no.of cases (< 50)
 2. Matching
It is a 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)
which are known to influence the
outcome of disease & which,if not
adequately matched for comparability,
could distort or confound the results
Types of Matching Procedure
• They are many
1. Grouping matching

2. Pairs
 3. Measurement of Exposure
• Definition & criteria are important
• By :
interviews
questionnaire
study past records
 4. Analysis
• Exposure rate

• Odd ratio
 Exposure rate
(frequency of exposure)
Cases Controls
(Lung cancer) (no lung cancer)
Smoker 33 55
(< 5/d) (a) (b)
Non-smoker 2 27
(c) (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%
Lung cancer is higher among
smokers than non-smokers
• If the exposure rate among the cases is more
than the controls.
• We must see association is there in
exposure and outcome. This is done by
using the chi-square test
• It is significant if p is less than 0.05.
 Odds ratio (cross-product ratio)
• So odd ratio is calculated from a case control
study.
• It is the ratio of the odds of exposure among
the cases to the odds in favour of exposure
among the controls.
• It is a measure of the strength of the
association between risk factor and outcome
• The derivation of the odds ratio is based on
three assumptions:
● The disease being investigated is relatively
rare
● The cases must be representative of those
with the disease
● The controls must be representative of those
without the disease.
• It is the cross product of the entries of the
table above
a/b c/d
= ad/bc
= 33X27/55X2 = 8.1
• So we can say smokers of less than 5
cigarettes per day showed a risk of having
lung cancer 8.1 times that of non-smokers.
 Odds ratio
• OR=1 Exposure does not affect odds of outcome means
the exposure and disease are not likely associated.

• OR>1 Exposure associated with higher odds of outcome

mean "those with the disease are more likely to have
been exposed,"

• OR <1Exposure associated with lower odds of outcome

means 1-exposure is a protective factor in the causation
of the disease.
• Case control studies are usually faster and
more cost effective
• Sensitive to bias (selection bias).
• The main challenge is to identify the
appropriate control group;
• The distribution of exposure among the
control group should be representative of
the distribution in the population that gave
rise to the cases.
Thanks