Descriptive Epidemiology

study designs
Zziwa Swaibu(BEH,MPH MUK)

1
 Goals of Research
• Evaluate new therapies and diagnostics
• Evaluate screening procedures
• Determine cause effect relationships
• Describe a disease, prevalence and natural
history
• Determine a prognosis
• Medical Review
 Types of Epidemiological Studies
I. Observational Studies
A. Descriptive Studies
• Case reports/case series and ecological studies
• Cross-sectional studies
B. Analytical Studies
• Case-Control studies
• Cohort studies
II. Intervention Studies
A. Clinical Trials
B. Community Trials

3
Hierarchy of research designs

4
 Describe
• describe [dɪˈskraɪb]vb (tr)1. to give an account
or representation of in words
• de·scrip·tive  (d-skrptv)adj. Involving or
characterized by description; serving to
describe.

5
 Research Study Design

Method of ascertaining subjects and

measuring states or events in them to
answer specific scientific questions.
 The starting point
Descriptive Epidemiology is the starting point in
the journey to
• Hypothesis formulation
• Analytical studies
• Experimental studies
It is concerned with amount / frequency of
disease events and their distribution patterns
by person, place and time characteristics
7
Select Study Design to Match the
Research Goals
Objective Design
Description of disease or spectrum Case series or report
Cross-sectional study
Determine operating characteristics Cross-sectional
of a new diagnostic test
Describe prognosis Cohort study
Determine cause-effect Cohort study
Case-control study
Compare new interventions Randomized clinical trial
Summarize literature Meta-analysis
 Introduces the design and conduct of
Epidemiological studies

• Descriptive study designs include:

– Case reports
– Case series
– Ecological studies
– Cross sectional studies

9
 Case reports
• A case report is a report describing a case in
detail and comprehensively.
• The case may be a patient (most commonly)
or may be an organisation.
• If it is a patient, the report should include the
address, the history of the illness, clinical
findings, outcome of laboratory investigations,
diagnosis, diet, behaviour, habits, treatments
+ outcome of a follow up period.

10
 Time frame of case reports

• Most of the content is cross sectional, but

• Information from the past (retrospective data)
may be required
• It commonly contains a follow up period
(prospective period) as may be found
necessary

11
 Case series
• Case series are studies of similar cases. They
are cases with the same diagnosis.
• Examples include cases of malnutrition ,
diarrhoea, hypertension, diabetes etc.
• They may be compared with the general
population with the aim of investigating
causality or prognosis.

12
 Cross-sectional studies
• Definition
• A study design that is conducted at a single
point in time or over a short period
• NO follow-up
• It looks at the population at a snapshot
•

13
 Cross-sectional studies
• In cross-sectional studies all the information
required from study subjects is obtained at
one point in time.
• It is also called a prevalence study
• It is difficult to ascertain cause and effect
relationships
• Many surveys are cross-sectional, but a repeat
survey in the same population may provide
longitudinal data
14
 Cross-Sectional Study
• Determines simultaneously both the exposure
and outcome in study participants

• One can determine whether there is an

association between exposure and outcome

• If outcome is disease, then this design relies on

prevalent cases of disease
– Design is sometimes called Prevalence Study
 Cross-Sectional Study
• The outcome of interest and the factors
that may be associated with that outcome
are assessed at the same point in time
• We can compare prevalence of ‘disease’ in
exposed & non-exposed
• We can also compare the prevalence of
exposure in those with ‘disease’ and those
without

16
 Cross-Sectional Study

Defined Study Population

Exposed Exposed Not Exposed Not Exposed

Disease No Disease Disease No Disease

 Cross-Sectional
No
Study
Disease Disease

Exposed A B A+B

Unexposed C D C+D

A+C B+D N

Ratio of Disease Prevalence: [A/A+B]/[C/C+D]

Ratio of Exposure: [A/A+C]/[B/B+D]

 Descriptive Information

• Cross-sectional studies mainly generate

descriptive information
• The primary objective is to describe
• We mainly report on:
– Means, modes, medians for numerical data
– Frequencies and proportions for categorical data
• – We then formulate hypotheses that can be tested
in analytical studies

19
 Strengths
Strengths
• Association studies
– Fast
– Inexpensive
– May answer question of interest
• Preliminary analysis
– Determine prevalence of outcome and exposure
– May be used to design future studies
• Baseline Data
– Start of cohort study
– May be used to select cohort
– May even be used for repeat cross-sectional analyses
 Weaknesses
Weaknesses
• Prevalent cases
– May not be representative of all cases
– Survival bias, the measure of association would determine the
association with disease plus survival to diagnosis as opposed
to disease incidence.
• Temporality
– Exposure and outcome are measured at the same time
– Cannot determine whether exposure preceded outcome
• Outcome may precede exposure
– Unable to establish causal relations – need for longitudinal
study designs
 Difference bet’n case-control, cohort and
crossectional study:

• In a case-control, starting point is the pple with disease and

those without disease
– We then look work backwards to assess whether they were
exposed or not
• In a cohort, starting point is the pple who were exposed
– We follow them forward in time to see if they develop a
disease or not
• In a crossectional study, the starting point is the population
at a point in time
– We study both the suspected exposure and the disease at
the same time

22
 Case-control study
• In case-control studies the relationship between
cases of a particular disease with a possible exposure
is investigated
• The investigator goes through the following steps:
– Identify persons with the disease of interest i.e. the cases
– Identify a comparable group of people without the Disease
i.e. the controls
– Establish history of previous exposure to the suspected
agent(s) in both cases and controls
– Compare frequency of exposure in the 2 groups

23
 Illustration:

Were Were not Were Were not

exposed exposed exposed exposed

Diseased Not Diseased

(Cases) (Controls)

24
The findings can be represented in
a 2X2 table as follows:

25
 The time frame
• The process in case-control studies is
retrospective because it looks back in time
• It begins with identifying cases of a disease
• Proceeds with inquiries about past exposure
to the suspected cause
• Similar inquiries are done for the controls and
the two groups then get compared

26
 Advantages of Case-controls
• They can be conducted quickly
• Can provide quick answers to explain current
problems
• Usually cheap compared to other types of
studies
• Since there is no follow up period, study
subjects cannot be lost
• Suited to investigating rare conditions
• We can study several possible causes of the
same outcome 27
 Disadvantages of Case-control
• Problems with recall
– Limitations in recall (forgetting),
– Recall bias (cases may be more keen to recall
historical exposures more than controls) or
• Reporting bias (deception)
• Selection of cases
• Selection of controls
• Interviewer bias
28
 Cohort Studies

• Synonyms:
– Follow-up studies
– Incidence studies
– Longitudinal studies
• Definition:
– “Cohort” means a group of persons who share a
common experience within a defined time period
• Examples
– Birth cohort, marriage cohort, class cohort, workers in
a mine

29
 Important Features

• At the beginning of the study, persons to be

studied have not yet developed the disease
• The exposure is known and can be measured
• Two groups of people from the same „mother‟
population are followed up:
– One group has had or is under the exposure factor
– The other group (the control group) is free of
exposure factor
– Both groups are followed up in time
– Both groups are initially disease free

30
 Important Features

• During follow-up, incident cases of disease are

noted and comparisons made through
calculations of disease risk in both groups
• Any suspected causal factor precedes the onset
of disease
• This type of design is quite useful in proving
causal relationship
• Cohort studies are observational: We record
observations in the natural setting without any
manipulation
• Cohort studies are analytical: The investigator
compares risk in two groups
31
Illustration

32
We can Summarize the outcomes
in a 2X2 Table

33
 Advantages of Cohort study
Designs

• The suspected causal factor levels are known on

each subject at the beginning so there is no
selection bias
• Most appropriate to study rare causes e.g. the
Hiroshima Atomic Bomb blast
• They are useful to investigate multiple effects of
an exposure over time.
• Very useful in measurements of time relationships
(Survival Analysis)
• It is easy to calculate the risk of developing disease
through the use of incidence rates
34
 Disadvantages of Cohort Study
Designs

• It is almost practically impossible to investigate

rare diseases ; it would mean big numbers
• Difficult to investigate slow developing diseases
it would mean following up for long
• Expensive: Money, time, personnel, etc
• Loss to follow-up of subjects
• Potential bias in assessing the outcomes
• Difficulty in identifying comparable cohorts

35
 When to Use a Cohort Study
Design

• When good evidence suggests association

betw‟n disease and a certain exposure
• When we are able to minimize loss to follow up,
a cohort study may be a good option
• When the interval between exposure to disease
is relatively short (prospective cohorts)
• When there is a complete recording system for
important events (historical cohorts)
• When comparable cohorts of such exposed and
unexposed persons are hard to come by

36
 Potential biases in cohort
studies

1.Biases in assessing outcome: The person

reading may be aware of the exposure status
2. Information Bias: Especially if the quality
and depth of information obtained is different for
exposed vs. non-exposed groups
3. Non-response: Non-response and loss to
follow up
4. Analytical bias: Strong preconceptions in
epidemiologist or statistician analyzing the data

37
 Types of Cohort Studies

• Broadly, there are three types of cohort

studies:
– Concurrent cohorts (or concurrent
prospective studies)
– Retrospective cohorts (or historical cohorts)
– A hybrid of the two designs (or
Ambidirectional cohorts)

38
 Ecologic Studies
• Like cross-sectional studies in that they
measure associations
• Unlike cross sectional studies, they are
conducted on group characteristics
 Ecological studies
• Ecological studies are based on group
characteristics or features.
• The group is typically a country, but it could
be a smaller (or larger) area
• Typically there is country-specific data on food
consumption e.g. alcohol intake, dietary fat
intake, sugar or salt intake, and other
consumables

40
 Analysis of ecological studies
• The common practice is to correlate disease
frequency with the per capita consumption of
the food item of interest per country
• Examples include alcohol consumption with
liver disease, fat consumption with either
heart disease or breast cancer

41
Heart disease with sugar consumption

42
 Measuring Exposures
• Aggregate measures
– Summarize characteristics of individuals, e.g. median
income
• Environmental measures
– Characteristics of location, e.g, air pollution intensity
• Global measures
– Group characteristics that cannot be reduced to the
individual level, eg. Political system
 Weaknesses of ecological studies
• It is not known or certain that the individuals
that develop the disease are the ones with the
higher dietary intake. Every one in the
specified country is assumed to be taking the
same average amount of the food of interest.

• This problem is called the “ecological fallacy”

44
 Ecologic Fallacy
• The bias that may occur because an
association observed between variables on an
aggregate level(group level) does not
necessarily represent the association that
exists on the individual level. (Last, Dictionary of Epidemiology)
 Strengths
• Provide valuable perspective on risk factors and
disease across populations
• Provide a natural comparison across groups of
people
• Data may be easily available
• Studies may be most appropriate for exposures with
population level effects
• Studies provide valuable insight for the development
of research questions
 Weaknesses
• Prone to ecologic fallacy
• Cannot demonstrate causal association