Displaying 1 result from an estimated 1 matches for "hrac".
Did you mean:
hmac
2018 Jan 18
1
Time-dependent coefficients in a Cox model with categorical variants
...ables for a regression matrix. The tt() call is a simple function, however, and ordinary multiplication and does not have those powers. In this case you need to do the setup by hand : create your own 0/1 dummy variables and work with them.
sqcc <- ifelse(dta$Histology == 'Sqcc', 0, 1)
hrac <- ifelse(dta$Histology == 'High risk AC', 0, 1)
fit <- coxph(Surv(time, status) ~ Sex + sqcc + hrac + tt(sqcc) + tt(hrac),
data = dta, tt = list(function(x,t, ...) x*log(t),
function(x, t, ...) x* log(t)))
Terry Therneau
PS I...