ASSOCIATION AND CAUSATION
ASSOCIATION
Association is defined as the occurrence of two variables more often than
would be expected by chance.
Association does not necessarily imply causation.
Classification of association
1. Spurious association
Not a real association
Occurs due to selection bias, when like is not compared with like.
Example – perinatal deaths in hospital births compared with the home
deliveries. This comparison shows that perinatal death is higher in hospital
birth as compared to home births leading to a conclusion that home is a
safer place for delivery, whereas is not true. The higher perinatal mortality
in hospital is due to complicated cases reporting to the hospitals.
Therefore it is important to have a proper selection of study objects.
2. Indirect association
It is a statistical association between a characteristic and a disease due to
the presence of another factor (called confounding factor), known or
unknown that is common to both the characteristic and the disease.
Example- endemic goitre and high altitude -shows association. Iodine is
the cofounding factor since goitre occurs due to iodine deficiency and
iodine deficiency exists in high altitude.
3. Direct association
I) One to one causal relationship
Two variables are stated to be causally related, if change in one
is followed by change in another. It means when a factor say A is
present, disease say B must result. Conversely, when disease B
is present, factor A must also occur.
Although ideal in disease aetiology, sometimes a one to one
causal relationship does not always explain the situation as –
1. The disease may not occur in spite of harbouring the organism.
Eg- TB
2. The same organism can lead to multiple diseases. Eg-
haemolytic streptococci can lead to streptococcal tonsillitis,
scarlet fever and erysipelas.
ii) multifactorial association
More than one factors, can act independently or cumulatively to produce
the same disease. Eg- smoking, air pollution, occupational exposure to
asbestos can lead to lung cancer.
Additional criteria for judging causality
Hill’s criteria
1. Temporal association
2. Strength of association
3. Specificity of the association
4.
Consistency of the association
5. Biological plausibility
6. Coherence of the association
