Association and Causation
Association and Causation
Association and Causation
&
Causation
a) Spurious Association
b) Indirect Association
c) Direct (causal) Association
i. One-to-one Causal Association
ii. Multi factorial Causation
Spurious Association
Spurious Association
Sometime an observed association
between a disease and suspected factor
may not be real.
E.g.
A study was conducted between births at home and
births in hospital.
Apparently Perinatal Mortality was higher in
hospital births than in home birth.
It may be concluded that home deliveries are safer
than hospital deliveries.
Such a conclusion is spurious because in general,
hospitals attract women at high risk for delivery
because of their special equipment and expertise.
Indirect Association
Indirect Association
Many associations which at first
appeared to be causal have been found
on further study to be due to indirect
association.
Indirect association is a statistical association
between a characteristic of interest and a
disease due to presence of a third common
(confounding) factor, known or unknown.
C
Cigarette
Smoking
A
CHD
Direct Association
Direct Association
One-to-one causal relationship
Multifactorial Causation
One-to-One Causal
Relationship
One-to-one Causal Relationship
Two variable are stated to be causally
related (AB) if a change in A is followed
by a change in B.
This is known as one-to-one causal
relationship.
Suggesting that when the factor A is
present, the disease B must result.
E.g Measles
One-to-One Causal
Relationship
RNA Paramyxovirus
Measles
Multi Factorial Causation
Especially important in non-
communicable disease i.e CHD, Lung
Cancer
Multi Factorial Causation
Each of the factor below can act
independently to produce a disease e.g
air pollution, smoking , asbestos
exposure Lung cancer
Factor 1
Reaction At Disease
Factor 2 Cellular Level
Factor 3
Multi Factorial Causation
Each factor can cumulatively produce a
disease and can have a synergistic effect.
Factor 1
+
Factor 2 Reaction At
Disease
+ Cellular Level
Factor 3
Causal Association
Causal Association
In the absence of controlled
experimental evidence to incriminate
the cause, certain additional criteria
have been evolved for deciding when
an association maybe considered a
causal association
Causal Association
The following concepts are used by
epidemiologists in making a causal
inference:
a) Strength of association
b) Consistency of the observed
association
c) Specificity of the association
d) Temporal sequence of events
e) Dose-response relationship
f) Biological plausibility of the
observed association
g) Coherence of the association
h) Experimental evidence
i) Reversibility
Strength of Association
Strength of Association
The strength of association is measured by;
Relative risk-------is it large?
Ratio of disease incidence among exposed and
non-exposed
E.g. RR of lung cancer = 10: 1
31
Large the relative risk, greater the
likelihood of a causal association.
Causal relationship is strengthened if there
is a biological gradient or dose-response
relation.
Increasing exposure to risk factors increase
rise in disease incidence.
Consistency of the
Observed Association
Consistency of the
Observed Association
Confirmation by repeated findings of an
association in case-control and cohort studies
in different population groups and different
settings strengthens the inference of a causal
connection.
Many cohort and case-control studies have
shown an increased risk of cardiovascular
disease associated with oral contraceptive use.
34
Specificity of the Association
Specificity of the Association
39
Dose-response relationship
Dose-response relationship
If a factor is of causal importance in the
occurrence of a disease, then the risk of
developing the disease should be related to
the degree of exposure to the factor, i.e., a
dose-response relationship should exist.
43
Cigarette smoking and lung cancer hypothesis is
biological plausible.
46
Experimental evidence
Experimental evidence
The randomized clinical trial (RCT) is the
closest approximation in epidemiology to an
experiment, and a well-run trial may
confirm a causal relationship between an
exposure and an outcome.
48
Reversibility
Reversibility
When the removal of a possible cause results
in a reduced disease risk, the likelihood of
the association being causal is strengthened.
50
Observed Association
Could it be due to selection or
Measurement bias?
NO
NO
PROBABLY NOT
Could it be causal?
54
If the association between two variables
given by a researcher is simulated by
other researchers too then it has:
a) Specificity
b) Consistency
c) Coherence
d) Temporal Sequence
e) Biological Plausibility
b. Consistency
56
If a virus does not results in any other
disease other than chickenpox, this
association exhibits:
a) Specificity
b) Strength of Association
c) Coherence
d) Temporal Sequence
a. Specificity
58
Any Questions ?
Thank you