Measure of Association

Question 1

What is an exposure?

Answer 1

Independent variable in a study which refers to factors postulated to be associated with the outcome

Question 2

What is an outcome?

Answer 2

Event (disease) resulting from the exposure.

The dependent variable.

Question 3

What does the risk difference measure?

Answer 3

Absolute effect associated with exposure. Rexposed - Rnon-exposed

Null value is 0 because it means no difference

Question 4

What does the relative risk measure?

Answer 4

Rexposed/Rnon-exposed

gives a proportion of people that have higher risk

Null value is 1 because it means exposed and non-exposed are the same

Question 5

Can relative risk be measured from case control studies?

Answer 5

No because it is not possible to measure incidence of the disease.

Question 6

How is level of the effect measured in case control studies?

Answer 6

Using the ‘odds ratio’

Question 7

What is the odds ratio (OR) a measure of?

Answer 7

OR= Odds that a case was exposed / Odds that a control was exposed

Question 8

What does Odds Ratio tell us?

Answer 8

If OR = 1 there is no evidence of association between exposure and disease/condition

If OR > 1 positive association

If OR < 1 Negative association

Question 9

What do confidence intervals tell us?

Answer 9

Reflects precision of the estimate.

Provides indication of statistical significance.

If we took repeated samples and calculated measure of association then 95% of the time the estimate would be contained within the interval.