Lecture 3: Measures Of Association Flashcards
Creating possible alternatives to events that already happened, for example someone got a disease instead of not having had it.
Counterfactual thinking/theory
Counter-to the-facts
What assumption does counterfactual theory require?
Assumption of exchangeability
What is always in the columns of a 2x2 chart?
Disease/Outcome
Absolute differences is found by doing what with frequencies?
Subtracting
Relative differences is found by doing what with frequencies?
Dividing
True or False:
Absolute differences are always larger than relative differences
False
Absolute is smaller
Define ‘risk’
The probability of an outcome in an individual group, exposed or not exposed
Is risk considered a proportion or ratio?
Proportion
It is a simple percentage
Absolute risk reduction formula
Subtracting the absolute risk of those exposed - AR not exposed
Relative risk reduction formula
ARR
——
Absolute Risk of the unexposed
What does NNT stand for
Number needed to treat
What does NNH stand for?
Number needed to harm
Doe we want NNH or NNT to be smaller?
We want NNT to be small
We want NNH to be huge
Give an example of NNT
It would take 8 patients on a drug for 1 person to get positive benefits
1 in 8 patients on this drug will get positive benefits
How to calculate NNT/NNH?
1 divided by Absolute risk reduction
1
——
ARR
how to calculate risk ratio?
risk of outcome in exposed//////risk of outcome non-exposed=placebo
The exposure or treatment group of interest, is always in the denominator when calculating risk ratio. T/F
false
always in numerator
if risk ratio is >1
the outcome is more likely to occur in the study than in the unexposed group
if risk ratio is <1
the outcome is less likely to occur in the study group than the unexposed group
if risk ratio = 1
outcomes are equally likely in both groups
other names for risk ratio
odds ratio
hazard ratio
interpret a RR = 1.53
53% more likely
interpret a OR = 2.35
outcome is 2.35 times more likely
interpret a HR = 0.75
outcome is 25% less likely to occur
when interpreting a RR/OR/HR, there are 3 things you must include in your interpretation…..
- groups you are comparing
- direction of your results
- magnitude
visual representation of ratio data, horizontal plot allowing the comparison of multiple ratios and their differences from 1.0
forest plot
OR INTERPRETATION: association between baseline characteristics of firearm and pedestrian motor vehicle injury hospitalizations. Reference: white, Hispanic OR = 1.43
hospitalization due to firearm injury or pedestrian motor vehicle injury, is 43% more likely to occur in Hispanic ethnicities than white.
Odds is a ……?
ratio
define ‘odds’
frequency of an outcome occurring vs. not occurring
or
frequency of exposure vs. frequency of not being exposed
or
occurring/not occurring within one group
relate odds calculation to a 2x2 table
A/C or B/D
odds ratio calculation formula
odds of exposure in non-diseased
how is odds ratio different from odds
odds is within same group
OR; ratio of odds from 2 different groups
what is the easiest way to calculate OR?
cross multiply
in a 2x2 table: odds of exposure formula
A/C
in a 2x2 table: odds of exposure in controls
B/D
in a 2x2 table: odds ratio formula
AD/BC
when reading a research study, how do you know which 2 groups are being compared?
one of the groups will be listed as a reference or their RR will be = 1.0
