Lectures 8-10 Flashcards
Relative differences
Dividing frequencies, proportions, or counts
Which will ALWAYS be larger? Absolute or relative differences?
Relative
Why do studies often use relative differences instead of absolute differences?
Relative differences are always larger. Saying something has a 50% relative difference compared to something is more eye catching than saying it has a 10% absolute difference.
(T/F) Relative differences are the more representative statistic when compared to absolute differences.
False. Absolute differences are more representative.
Risk, Incidence Risk (IR), Attack Rate
Probability of outcome in exposed and non exposed.
A)/(A+B
(C)/(C+D)
Risk Ratio (RR), Relative Risk
Risk in Exposed/Risk in Unexposed
[(A)/(A+B)] / [(C)/(C+D)]
Interpret: RR=1.0 (Group A/Group B)
No difference in risk between two groups
Interpret: RR=1.53 (Group A/Group B)
Group A has 53% greater risk than Group B.
Interpret: RR=6.18 (Group A/Group B)
Group A has 6x greater risk than Group B
Intepret: RR = 0.73 (Group A/Group B)
Group A has 27% lower risk than Group B.
If risk in Forest Plot is:
- Left of RR=1
- Right of RR=1
- Left is decrease from control
2. Right is increase from control
Absolute Risk Reduction (ARR) a.k.a. Attributable Risk (AR)
The difference in risk (subtraction) between the exposed and unexposed that can be attributed to to the actual exposure.
Exposed Risk = 14%
Unexposed Risk = 17.8%
What is ARR?
ARR = 3.8%
Relative Risk Reduction (RRR)
(ARR)/(R in unexposed)
Exposed Risk = 14%
Unexposed Risk = 17.8%
ARR = 3.8%
RRR = ?
RRR = (ARR)/Risk in unexposed RRR = 3.8/17.8 RRR = 21.3%
Which is always higher? ARR or RRR?
RRR
Number Needed to Treat (NNT)
1/ARR
ARR=3.8% or 0.038
NNT = ?
NNT = 1/0.038 = 26.32
Needed to treat 27 patients to experience the study outcome.
For every 27 patients, one will have outcome.
If good outcome do you want small or large NNT?
Small NNT; Treat less people to get good results; Higher ratio of good results
If bad outcome do you want small or large NNT?
Large NNT; Want to treat many people to get one bad outcome; Small ratio of bad outcomes
Odds:
Outcome Exposure Cases Controls Total Yes A B A+B No C D C+D Total A+C B+D A+B+C+D
Odds of exposure in cases and controls
A/C
B/D
Odds Ratio (OR)
Odds of exposure in cases vs. odds of exposure in controls
A/C)/(B/D) or (A x D)/(B x C
What 3 things could be incorrect in the wrong answer choices for RR/OR/HR?
- Direction of words (increase or decrease)
- Magnitude (correct numbers)
- Group Comparison
Absolute Differences
Subtracting frequencies or counts