When to use KM curves w/ log-rank test
when predictor variable is:
categorical (placebo vs. drug) or,
only has a small # of values
What do KM curves w/ log-rank tests suck at?
Continuous variables (e.g. gene expression values, whatever the **** that is)
How do we do survival analysis on continuous variables? (e.g. white blood count, viral load, gene expression)
Cox proportional hazards regression analysis (Cox PH regression)
Cox can also do categorical data w/ dummy variables
And it "handles" censoring, whatever that means
What is a hazard function? h(t)
h(t) is function of the probability of an event in a certain time interval GIVEN that the individual has survived up to time t
What does it mean, proportional hazards?
hazards between group A and group B are proportional at all times