Cause and Association

Question 1

Define cause

Answer 1

A factor that is associated with the disease so that if the intensity or prevalence of the factor in the population changed, the incidence of the disease also changes.

Question 2

Define association

Answer 2

The level of variable differs according to the level of another variable

Question 3

Define causal association

Answer 3

Level of one variable changes according to the level of another directly

Question 4

Define Non-Causal association

Answer 4

The results are due to chance, bias or confounding

Question 5

What usually means association is likely to be causal?

Answer 5

 Have considered (a) chance, (b) bias and c (Confounding)
 Considered the criteria of causality
 Consider of the scope for preventative action

Question 6

Describe chance of association (to be cause)

Answer 6

• Test P-value to ensure it is <0.05
• Test confidence interval to ensure it doesn’t contain 1 if looking at RR or odds ratio, or doesn’t contain 0 if looking at difference between means or proportions
• Have a large RR or odds ratio to infer a greater association
• Increasing the size of the study to narrow t he confidence interval and a more significant P-value

Question 7

Describe bias of association (to be cause)

Answer 7

• Ensure there is no bias at all, such as selection bias or information bias etc
• Bias can obscure/exaggerate associations, and even if you increase the size of the study, the bias will not be removed!!!

Question 8

Describe confounding of association (to be cause)

Answer 8

• Confounding can be addressed in the study design (RCT, matching on confounders in a case-control study)
• Confounding can also be removed by standardisation

Question 9

Describe the 8 criteria for causality

Answer 9

1. Strength of association is high, ie (high RR, low CI etc)
2. Association is independent of known confounding variables, or is adjusted
3. The findings are specific for one disease per exposure
4. Results are consistent with other studies
5. Exposure preceeds outcome, exclude reverse causality
6. Identify a dose-response relationship
7. Finding are biologically plausible with understood mechanisms
8. See if the relationship is reversible, which can pave the way for intervention studies

Question 10

Preventative action of associatio n

Answer 10

• Can be primary, secondary or tertiary prevention (mostly primary)
• Ensure intervention is effected in reducing incidence
• Does not result in unacceptable levels of adverse side effects
• Cost effective (opportunity cost) in a sense that resources used here would not generator greater benefits if used elsewhere in the health services of a community

Cannot make reliable associations based on numerator data (ie how many cases of H.Pylori), you need to know something about the prevalence in undiseased subjects from the same community.
Ie – 80% of H.pylori infections drink tap water, however 80% of normal population also drink tap water.

Question 11

Equation for case-fatality

Answer 11

Relates deaths to prevalent cases of disease

Case fatality = Deaths in cases per time/ Total cases in study per time