Measures of Association

Question 1

What are the two components to measures of association?

Answer 1

1. Logic of analytic epidemiology
   - Importance of comparison groups
   - PECOT and GATE
2. Measures of association
   - relative risk (relative)
   - risk difference (absolute)
   - Calculation and interpretation. What does it tell us?

Question 2

What are the guts of analytic epidemiology?

Answer 2

Is the exposure associated with the outcome?
Does the exposure increase or decrease the occurrence of the outcome?

Question 3

What type of analytic study designs can we use to compare groups?

Answer 3

Cross sectional and ecological
Cohort
Case-control
Randomised controlled trial

PECOT and GATE can be used for each of these

Question 4

What does PECOT stand for?

Answer 4

Population - the group of people in the study
Exposure - what the potential determinant is
Comparison - what the potential determinant is being compared to
Outcome - the health outcome being assessed
Time - how long the people are being followed up

Question 5

Source vs. sample population

Answer 5

Source: population the sample is recruited from (the triangle in the GATE frame)

Sample: population included in the study

Question 6

What is relative risk?

Answer 6

The ratio of incidences:
Exposed/Comparison = Relative risk

This quantifies the degree to which an exposure increases or decreases the occurrence of the outcome

Question 7

What do the different values of relative risk mean?

Answer 7

RR above 1: greater occurrence of outcome in exposed group
- if outcome is bad, exposure is potentially a risk factor for the outcome

RR of 1: is the null value so this means that the groups are the exact same and there is a null effect of the exposure.
- Exposure doesn’t change the occurrence of the outcome, so no association between exposure and outcome

RR below 1: greater occurrence of outcome in comparison group
- if outcome is bad, exposure is potentially a protective factor for the outcome (prevents you from getting the outcome)

Question 8

How do you interpret relative risk (RR)?

Answer 8

The ‘exposed group’ were ‘value’ as likely to develop ‘outcome’ compared to ‘comparison group’.

(Same interpretation if using incidence proportion or incidence rate)

Question 9

What is risk difference/attributed risk?

Answer 9

The differences in incidence (incidence of exposure - incidence of comparison)

This tells you how many extra/fewer cases of the outcome in the exposed group are attributable to the exposure

Question 10

What do the different values of risk different/attributed risk mean?

Answer 10

Exposure incidence > comparison incidence means risk factor

Null value = 0 means no association

Exposure incidence < comparison incidence means protective factor

Question 11

How to interpret/report the risk difference?

Answer 11

Report differently for incidence proportion and incidence rate!! Use the value you calculated

There were ‘value’ extra/fewer cases of ‘outcome’ in ‘exposed group’ compared to ‘comparison group’

IP eg. Value: 25 extra cases per 100 people over one year
IR eg. Value: 25 extra cases per person-year

Question 12

Relative risk vs. risk difference

Answer 12

Relative risk:
- clues to aetiology
- strength of association

Risk difference:
- impact of exposure
- impact of removing exposure

Both are important