Measures of Association

Question 1

Define Attributable Differences

Answer 1

Subtracting frequencies of each group from one another

Ex. 65 surgeries in men and 25 surgeries in women. Attributable difference = 65-25

Question 2

Define Attributable Risk

Answer 2

aka absolute risk reduction (ARR) (proportion)

2 different risks subtracted to find the attributable risk between the 2 values

Ex. 40%-53% = l13%l (is a positive value because it measures difference between the two values)

Question 3

Relative difference in frequencies

Answer 3

Division (ratio) of frequencies

ex. 65 male surgeries divided by 25 surgeries in females = males had 2.6 times the number of surgeries

Question 4

Relative difference in proportions

Answer 4

Division (ratio) of proportions

ex. # of males with surgery / total number of males

DIVIDED BY

of females with surgery / total number of females

Question 5

Incidence Risk (IR)

Answer 5

with disease and exposed / total # exposed

OR

with disease and not exposed / total # not exposed

Used to measure probability of outcome in exposed and non exposed groups

Question 6

Risk Ratio (RR)

Answer 6

(aka relative risk) is the ratio of risk proportions from two different groups being compared

Risk in outcome of disease in exposed / risk in outcome of disease in non exposed

Question 7

Absolute Risk Reduction (ARR)

Answer 7

Compares two groups by subtracting their calculated risks of occurrence from one another

Question 8

Relative Risk Reduction (RRR)

Answer 8

ARR / Risk of unexposed group

Question 9

Odds

Answer 9

A Ratio that describes the frequency of an outcome occurring vs. not occurring

Question 10

Odds Ratios (OR)

Answer 10

odds of exposure in diseased / odds of exposure in non diseased

Can cross multiply in 2x2 tables to get this

A ratio of the odds from 2 different groups

Question 11

What does a Proportion describe?

Answer 11

The equality, or inequality, between 2 ratios

Question 12

Number Needed to Treat (NNT)

Answer 12

1 / ARR (in decimal format)

The answer must be rounded up because it represents a number of people (cant have half people

Question 13

What is counterfactual Theory?

Answer 13

The outcome if the exposure didn’t occur in the same group being studied (impossible to actually do)

Question 14

What is Risk?

Answer 14

a proportion (part over whole) that predicts the probability of an outcome

Question 15

What is Odds?

Answer 15

a Ratio that describes the frequency of an outcome occurring vs. not occurring

NOT A PERCENTAGE DAMN IT

Question 16

When interpreting Ratios, what 3 parts need to be included in the interpretation of the results?

Answer 16

1. Group comparison orientation (say groups in correct order)
2. Direction of words (increased/decreased)
3. Magnitude (80% is 1.8 times or 20% is 0.8 times)
   [basically describe the distance from zero]

Question 17

When interpreting a ratio that was adjusted to account for confounding, what MUST be added to the interpretation?

Answer 17

what variable the ratio was mathematically adjusted for.

ex. if it’s for a preexisting condition that was causing confounding, you must say “this ratio was adjusted for preexisting condition A morbidity”

Question 18

Define Exchangeability

Answer 18

comparability with respect to all other determinants of the outcome