Analytic Epidemiology

BY: Fisaha.H (MPhil)

Analytic epidemiological studies
Purpose/ Aim
To search for cause and effect.
Why?? How??

To test hypothesis about causal relationship

Proof Vs Sufficient evidence

To quantify the association between exposure

and outcome
Measure of association
 Cont…
It focuses on determinants of disease by testing
hypothesis.
Basic features

Appropriate comparison group needed

• Exposed Vs Unexposed(experimental Vs control)

It is the use of comparison group that allows

testing of epidemiologic hypotheses
 Cont…

Two types of analytic study designs

Observational

Interventional designs
 Cont…
Difference lies in the role of the investigator.
Observational studies
the investigator simply observes the natural course of
an event
the investigator measures but does not intervene.
Interventional studies
the investigator assigns study subjects to exposed and
non-exposed, then follows to measure for disease
occurrence.
the investigator manipulates the intervention or
exposure.
 Observational studies
Information is obtained by simple observation of
the event. (Two basic types)

Case control study

Cohort studies
 Case control study
Cases (subjects having a specific disease)

controls (subjects not having the disease)

are compared for their exposure status.

It assess retrospectively on exposure status

Both the exposure and disease have already

occurred at the time of the study
Principles of the Case-Control approach

Persons with the disease are compared with

controls( free of disease), for the presence of the
factor(exposure) under investigation

Cases compared to controls with respect to

exposure frequency via the exposure odds ratio
 Steps of a Case-Control Study
Identify group of cases

Identify group of controls

Asks both groups for exposure

Measures frequency of exposure occurrence

in both groups

Compare frequency of exposure between

cases and controls
 Design of case control
Time
Direction of
inquiry
Exposed
Cases
(People with
disease)
Non-exposed
Population
Exposed Controls
(People without
disease)
Non-exposed

10
 Source population

Exposed
Unexposed
 Source population

Exposed
Cases
Unexposed
 Source population

Exposed
Cases
Unexposed

Controls
 Who is the Right Control?
As similar to a case as possible but without the disease
in question

Must have the same opportunity for exposure as a case

Must be subject to the same inclusion and exclusion

criteria

No control group is optimal for all situations

Scientific, economic and practical considerations

 Examples of Controls
Community

Neighbourhood

Hospitals

Friends
 Community Controls

Advantages
If all cases in general population known
Direct calculation of risk

Disadvantages
Cost
Sampling frame
 Neighborhood Controls
Advantages:
Inexpensive, efficient
Matched for potentially confounding variables

Disadvantages
Exposure related to neighborhood
Potential bias
 Hospital Controls
Advantages:
Convenient
Come from same catchment area

Disadvantages:
Control disease may be linked to exposure
Hospitalized controls differ from general population
 Friend Controls
Advantages
Convenient
Disadvantages

Bias
Friends may share same exposure
 Number of controls
Availability
Ratio controls / cases

Trade-off: cost vs. power

Decision based on power calculation

More than one control group?

Application of Case-Control studies

• Rare diseases or outcome

• Long latency between exposure and outcome

• Explore a wide range of potential exposures for

a single outcome
 Case-Control Studies
Advantages
Rare disease, e.g. cancer of a specific organ

Suitable for the evaluation of diseases with long latent

periods

Quick and inexpensive

Relatively efficient, small sample size

Little problem with attrition

Can examine multiple etiologic exposures

No ethical problems
 Case-Control Studies
Disadvantages

Inefficient for rare exposures

No calculation of rates and risks

In some situations, the temporal relationship between

exposure and disease may be difficult to establish
Temporal E – D uncertainty
Prone to selection and information bias

Selection of controls difficult

Cohort Studies
 Cohort study
Population
at risk

Exposed Disease among

exposed?

and Usually prospective

Disease among non-

Not Exposed exposed?
 What is a cohort?
A group of persons
sharing the same experience
followed for a specified period of time
Examples
birth cohort
workers at a chemical plant
graduating university class
Attendants of this course
 Cohort Studies
Disease free exposed and non-exposed people
are followed up and then outcome events are
picked up when they occur

Measure and compare the incidence of disease

in two or more study cohorts

Usually prospective or forward looking.

Are also called longitudinal studies.

 What is a cohort?
A group of persons
– sharing the same experience
– followed for a specified period of time
Examples
– birth cohort
– workers at a chemical plant
– graduating university class
– attendants of this course
 Cohort study

Population
at risk Disease among
exposed?
Exposed
Usually prospective

Not Exposed
Disease among
non-exposed?
 Types of Cohort Studies

• Based on the starting point of the study

– Prospective
– Retrospective (historical)
 Prospective Cohort Study

Exposure Study starts Disease

occurrence occurrence

ill
+ -
+ +
exp exp
- -

Selection based Prospective assessment

on exposure of disease
Time
 Prospective cohort …

Includes either an entire population or a representative

sample of the population.

Exposures are unknown until the first period of

observation when exposure information is collected
after administration of questionnaires, collection of
biologic samples, etc.

The cohort can be divided into exposed and non-

exposed group.
 Retrospective cohort studies

Disease
Exposure occurrence Study sarts

time

Disease outbreak following a gathering

Occupational exposure in mine workers

 Rétrospective Cohort Study

Exposure Disease Study

occurrence occurrence takes place
ill
+ -
+
exp
-

Rétrospective assessment
of exposure and disease
Limitations of retrospective cohort:
All relevant variables may not be available in
the original records

Difficult to ascertain that the study population

was free from the disease at the start

Loss of records, incomplete data

 Fixed or dynamic cohort
Fixed cohort
No additional participants are added to the cohort
once enrolled.

Dynamic cohort
The members in the cohort vary over follow up time
Some leave and new recruits are done
Use of person time
 Issues in Design of Cohort Studies

Selection of the Exposed Population

The choice of a particular group for enrollment depends
on:
Frequency of exposures of interest
Need complete and accurate exposure and follow-up
information
For common exposures – sufficient numbers can be
identified from many sources
For rare exposures or special exposure groups
(occupational) – need to get enough exposed persons
in reasonable time
 Issues in Design of Cohort Studies…

Selection of the Comparison Group

Comparison groups are as similar as possible except for
determinant under study.
Similar information available as for exposed group
For single general cohort – members can be categorized
into different exposure groups (internal comparison group)
Single, well defined reference population with no risk
factors under consideration
 Issues in Design …
Sources of Data
Exposure Information
Pre-existing records
– Availability for much of cohort
– Inexpensive
– Objective, bias-free categorization of exposure
status
– But – insufficient detail and no information on
potential confounders
 Sources of Data…

Information from study subjects

– Information on data not routinely collected
– Questionnaires/interviews
– Potential bias

Ascertainment of exposure must be comparable

for all
 Issues in Design …Sources of Data

Outcome Information
Obtain complete, comparable, unbiased information
Death certificates (potential bias when cause-specific
mortality)
Medical records, Medical Aid schemes, etc.
From study subjects
Periodic direct medical examinations

Apply equally to exposed and non-exposed

 Issues in Design …Follow-up
Failure to ascertain outcome data is the major source of
potential bias

Follow-up = time, cost, initiative

Length of follow-up period is related to latency period of
disease
The longer the follow-up period the more difficult to
ensure complete data
If lost to follow-up is large (eg, 30-40%) ? Validity ?
Loss to follow-up may be differential
 Different outcomes in cohort

No disease

Diseased

Lost to follow up
 Advantages of cohort studies
Directly measure relative risk or rate
Measures of effect have clear meaning and are
easily understandable
Temporal relationship between exposure &
disease is clear
 Advantages of cohort studies

Prospective cohort studies less susceptible to

selection bias because outcome not known
Well suited to rare exposures
Several outcomes can be examined in one study
 Disadvantages of cohort studies
Large sample size

Inefficient for disease with latency period

Loss to follow-up

Exposure can change over time

Multiple exposures = difficult

High cost

Time consuming
 Experimental
( Intervention studies)
Definition:
An experiment is a set of observations,
conducted under controlled circumstances, in
which the scientist manipulates the
conditions to ascertain what effect such
manipulation has on the observations.

Investigator has control over who gets exposure

and who don't. The key is that the investigator
assign into either group, whether it is done
randomly or not.
Key Features of Experimental Design

1) Investigator manipulates the condition under

study

2) Always prospective
• Gold standard for epidemiologic research

• Blinded

• Randomized

Placebo control

• Only interventional study fulfills this

Blinding
 Blinding…
Relevant groups who may/may not have
knowledge of treatment assignments
Participants
Investigators/clinicians
administering intervention
Investigators assessing outcomes
Data analyst(s)
 Single, double, triple, and beyond
Open/Unblind = all know which intervention a
patient is receiving

Single-blind

The participants (usually) or investigators

assessing out(alternately) do not know the
assignments
 Blinding…
Double-blind
Two groups do not know usually it is the
participant and the outcome
assessors/investigators
 Blinding…
Triple or quadruple blinding

Three or four of the relevant groups (prior slide) are

not aware of the treatment assignment
 Why use blinding?
Use blinding to eliminate the effect of
knowledge of treatment - which introduces
bias

If a patient is aware of assignment, knowledge of

therapy may affect outcome (use a placebo)

Observer may differentially recognize and record

information regarding the patient based on the
treatment group of the subject
 What is random allocation?
Random allocation means that all
Participants have a defined probability of
Assignment to a particular intervention
Allocation is not determined by the
Investigator, clinicians, or participants
Allocation is not predictable based on a pattern
 What purpose is served by random
allocation?

Covariates are distributed equally across the

groups at baseline
Not always (especially if N is small)
Affects both measured and, more importantly,
unmeasured variables
The risk of imbalance remains even after properly
executed randomization
 What purpose…

If randomization has been performed

correctly, chance is the only explanation for
any observed difference between groups, in
which case statistical tests are considered
superfluous
 What elements of a trial can be
randomized?
Most common unit is individual patient Sometimes
groups are randomized=cluster randomization

Examples: families, schools, towns, hospitals,

communities
Worry about contamination in cluster randomization

Special statistical techniques needed to cope with the

loss of independence of the individual units
 Placebo
A biologically inert intervention which is used to elicit
any non-specific psychological effects of the test
intervention.

Designed to resemble the test intervention as closely

as possible, except for the presumably active
component.

May not be justified if an intervention of known

effectiveness is already in general use
 Purpose of intervention trial
• Proof of concept trail: designed solely to produce
knowledge about cause and effect, does not test
the efficacy of the intervention in actual practice.

• Prevention trail: interventions are to prevent

disease and study participants are persons without
disease.

• Clinical trail: interventions are treatment based on

drugs and study participants are persons with
disease.
 Classification
• Based on population
 Clinical trial - usually performed in clinical
setting and the subjects are patients

Field trial- used in testing medicine for

preventive purpose and the subjects are
healthy people.

Community trial - a field trial in which the

unit of the study is group of people/community
 Classification …
Based on design:
 Uncontrolled trial - no control group. control
will be past experience (history).
Non-randomized controlled Trial - there is
control group but allocation into either group
is not randomized.
 Randomized controlled - there is control group
and allocation into either group is randomized.
 Classification…
Based on Trial Objective:
• Phase I
– Clinical pharmacology, drug safety & Toxicity
– Scale: 20-80 healthy individuals
• Phase II
– Initial assessment of clinical effect
– Scale: 100-200 patients
• Phase III
– Randomized Controlled Trial(RCT): for test of
effectiveness
– Scale: determined according to minimum clinically
significant effect
• Phase IV
– Long term prospective assessment of effects & side-
effects
 Major Problems Related to
Intervention Studies
• Feasibility/practical problems related to
Compliance

• Cost

• Ethics
 Feasibility/ Practical Issues

• Subject recruitment
 getting adequate individuals to enroll into a difficult
study is not easy.
 Field trials particularly require greater number of
subjects since the risk of contracting a given disease for
the first time is small.
• Loss to follow-up
 Select population who are both interested and reliable.
 Arrange frequent contacts with individuals
 Use incentives, such as providing medical information
 Cost
 Large number of subjects

 Expenses for visits by study subjects to

designated sites (especially if they do not come
to clinics)

 Preventive trials are limited either extremely

common or extremely serious diseases.
 Ethical Issues

 Ethical regulations are more stringent for intervention

studies
 End benefit assurance: chances to offer the benefit of the
trial
for study subjects/community.
 No participating subject could be treated better than the
alternative treatment possibilities that the study protocol
provides.
 This implies that any exposure or treatments given to
subjects should be limited to potential preventives of
disease or disease consequences- serious limitation to
etiologic research.
 Ethical Considerations
Risks vs benefits
Comparison: Standard care vs placebo
Ethical approval
Informed consent & confidentiality
Freedom to withdraw
Duty of care
Stopping/Monitoring
Reporting findings
Quality: ‘Poor’ quality research is unethical!
 Practical Considerations
Supplies of drugs/placebos
Compliance of clinicians/nurses & patients:
– Keep data required to a minimum
– Keep forms simple & logical
– Keep people informed of progress, e.g. newsletters, e-
mails, websites
Patient Information Leaflet
– Clear & easily understood
– Voluntary nature, free to withdraw
– Aim of RCT
– Interventions, incl. Side effects
Consent form
Summary of study designs