3 - Biostatistics 2 - Survival Analysis

Question 1

Relative Risk

Answer 1

Easily calculated by clinicians as Event Rate Treatment/Event Rate control

Within each group, it is calculated as # of Events/#in group

Question 2

Relative Risk Limitations

Answer 2

Doesn’t adjust for differences in follow up length/completion or confounding circumstances

Question 3

Number Needed to Treat

Answer 3

1/Absolute Risk Reduction

Question 4

Absolute Risk Reduction

Answer 4

Event Rate (Control) - Event Rate (Treatment)

Question 5

Relative Risk Reduction

Answer 5

1 - RR

Question 6

Censoring

Answer 6

When someone drops out, drop their data from the denominator, since you just don’t know.

Question 7

Commonly Used Censoring Stats

Answer 7

Kaplan-Meier Curve
Log-Rank Statistic
Cox Proportional Hazards Regression

Question 8

Kaplan-Meier Curves

Answer 8

Graphic depiction of the cumulative probability of survival in each group

Question 9

Log-Rank Test

Answer 9

Compares ratio of expected to observed events between treatment group across time units.
Adjusts for censoring
Can’t adjust for confounders

Question 10

Hazard Ratio (from Cox Proportional Hazards Model)

Answer 10

Uses censoring to reflect differences in follow-up length.
Allows adjustment for confounders
Allows testing of interactions (differences in treatment effect depending on other factors about who you are like diabetes)

Question 11

Hazard Ratio

Answer 11

The ratio of hazards in each group

Question 12

HR > 1

Answer 12

Risk is increased

Question 13

HR

Answer 13

Risk is reduced

Question 14

HR = 1

Answer 14

Inconclusive