Week 3 - Measures of Association

Question 1

What is exposure?

Answer 1

• An event/state/behaviour (i.e. any variable) to which a group of people is
  ‘exposed’

Question 2

What is outcome?

Answer 2

• An event/state/behaviour (i.e. any variable), which we hypothesize has
  resulted from exposure to some other event/state/behaviour (i.e. an
  exposure)

Question 3

What does outcome measure?

Answer 3

Frequency

Question 4

Measure of association summary

Answer 4

Question 5

What does “Measures of Association” mean?

Answer 5

*Assess the presence and the magnitude of an
association between an exposure (IV) and
an outcome (DV)
*Basically, assess the change in outcome for a given
change in exposure
*Apply to both numeric and categorical exposures
and outcomes

Question 6

What is relative risk?

Answer 6

• Relative Risk refers to the comparison between two risks
  (incidences); or better refers to a risk in relation to another
  (reference) risk
• Relative Risk is a measure of association and as such it
  assesses the probability of an event occurring (i.e. the risk
  of outcome) in an exposed group of people relative to the
  probability of the same event occurring in an unexposed
  group of people (reference)
• Thus both exposure and outcome variables are categorical

Question 7

What two measures of association is used for relative risk?

Answer 7

Risk Ratio is a ratio of two risks (cumulative
incidences)
Rate Ratio is a ratio of two rates (incidence rates) (Hazard Ratio)

Question 8

What are the exposure categories in relative risk?

Answer 8

The two groups of individuals compared in a
Relative Risk are the exposed category and the
unexposed (reference) category

Question 9

What is important to remember when calculating relative risk?

Answer 9

• When calculating a Relative Risk, the exposed
  category is always compared to the unexposed
  (reference) category, not vice versa!
• Both the Risk Ratio and Rate Ratio are applicable in
  situations where there is follow-up (i.e.
  prospective studies) and ascertainment of new
  (rather than existing) outcome cases

Question 10

Risk Ratio formula

Answer 10

Question 11

Rate Ratio formula

Answer 11

Question 12

Risk Ratio Example

Answer 12

Question 13

Important to remember!

Answer 13

1. Prevalence = percentage
2. Incidence rate = person-years
3. Relative Risk = -1 *100 (percentage)
4. OR Relative Risk = 2 DP (-fold)
5. Odds = divided by NON-CASES

Question 14

What are the three ways you can interpret relative risk?

Answer 14

> 1: the exposed have higher risk (or probability) of the
outcome (disease) than the unexposed
(i.e. the exposure is harmful for the outcome!)
<1: the exposed have lower risk (or probability) of the
outcome (disease) than the unexposed
(i.e. the exposure is protective for the outcome!)
=1: the exposed have exactly the same risk (or probability) of
the outcome (disease) with the unexposed
(i.e. the exposure has no association with the outcome!)

Question 15

Why is rate ratio more accurate than risk ratio?

Answer 15

Rate Ratios are considered
more accurate, since they take into account the real follow-up time.

Question 16

Odds formula

Answer 16

Question 17

What is odds?

Answer 17

• An alternative measure for the likelihood of having
  an outcome of interest (e.g. disease) which instead of
  using the whole population as the denominator,
  uses the non-cases (i.e. non-diseased)
• Odds measures the likelihood of having the outcome
  in relation to the probability of not having the outcome,
  i.e. odds of being affected (by disease) versus odds
  of not being affected.
• Odds is not commonly used on its own as a measure
  of disease frequency, but it is very commonly used for
  calculating the Odds Ratio (see next slides)

Question 18

When is Odds ratio used instead of Relative risk?

Answer 18

When it investigates association prospectively due to:
– lack of financial and other resources
– limited time availability
– outcome condition of interest is very rare in the
population