Lecture 12 causality

Question 1

Why is establishing causality important?

Answer 1

So that the data collected by epidemiological studies can be used for evidence based measures

Question 2

What groups do epidemiological studies determine the cause of disease

Answer 2

In populations, not in individuals

Question 3

Can preventative measures be applied before knowing about the causality of a disease?

Answer 3

Yes

Question 4

Can causality be proven in human studies? Why?

Answer 4

No because it would not be ethical or practical

Question 5

What is the nature of most epidemiological studies?

Answer 5

They are conducted in noisy environments and non experimental

Question 6

What do epidemiological studies determine?

Answer 6

Most epidemiological studies determine the association or relationship between a exposure and disease outcome, a statistical association can also be determined

Question 7

What should be look for when looking for links within exposure and outcome?

Answer 7

See if there are lots of studies conducted in diverse settings and adequately limiting confounding, non random errors and random errors,
Judge findings against a framework

Question 8

What are the components of the bradford hill framework?

Answer 8

temporality, specificity of association, consistency of association, strength of association, biological gradient, biological plausibility, reversiblity

Question 9

What is temporality?

Answer 9

The exposure occurs before outcome, this is an essential criteria to establish causality

Question 10

What is reversibility?

Answer 10

If the exposure is removed, the outcome is reversed in some way

Question 11

What is specificity of association?

Answer 11

A cause leads to a single effect and an effect has a single cause

Question 12

What is the strength of association?

Answer 12

The stronger the association the more likely to be causal in absence of known biases, e.g high RR or RD

Question 13

What is the biological gradient?

Answer 13

Incremental change in exposure correspond with incremental increases of the disease rates

Question 14

What is the biological plausibility?

Answer 14

Associations makes sense biologically

Question 15

What is the consistency of association?

Answer 15

Replication of the findings by different investigators, at different times, in different places, with different methods

Question 16

Does cause always result in effect?

Answer 16

No, causal phenomena are usually complex and exposure:outcome relationships are not usually 1:1, they are an interplay of the environment, host and agent

Question 17

What is a sufficient cause?

Answer 17

It is the whole pie: made up of multiple slices, it is the minimum set of conditions required for a outcome to occur, a disease may have several sufficient causes

Question 18

What is a component cause?

Answer 18

Each slice is a component cause, it contributes towards disease but is not enough to cause the disease on its own, it interacts with other component causes to contribute towards disease

Question 19

What is a necessary cause?

Answer 19

A component cause that must be present for the disease to occur

Question 20

What does almost every causal mechanism have?

Answer 20

An environmental factor

Question 21

Knowing the causal pie, how can we prevent disease

Answer 21

Knowledge of all components are not required, but removing any slices of the pie will result in prevention of some cases, can intervene at any part of the pie, knowledge of complete pathway is not a prerequisite

Question 22

What are problems with the causal pie model?

Answer 22

It follows a deterministic model, A causes B so whenever A occurs B will occur

Question 23

What is a probabilistic concept of causation?

Answer 23

A cause increases the chance that it’s effect will occur. A sufficient cause raises probability to 1, necessary cause raises it from 0, component cause contributes from 0-1

Question 24

What is the counterfactual definition of causation?

Answer 24

The presence or absence of the cause makes a difference in the outcome, and the sufficient, necessary and component causes clarify what difference it makes, is consistent with both deterministic and probabilistic models.