Cox regression using time varying covariates

[dac_js]

I have the following data:

data have;
input
ID Dose Days Event Age;
datalines;
1 1 10 1 35
2 1 40 0 35
2 2 55 0 25
2 3 23 1 67
3 1 20 0 34
3 2 35 0 34
3 3 15 1 25
4 1 40 0 53
4 2 55 1 53
;
run;

Here, dose variable changes with time. Days is the duration an individual stays alive (or die at the end of period) at a certain dose level. If the individual does not die, the dose is increased (1 to 2 or 2 to 3). Event variable shows when an event (death = 1) occurred. This can be at the end of any dosing period.

Is there any way to run a cox model where dose is the exposure of interest and age is a covariate?

[svh]

Have you looked into using the counting process when you have data with more than one row per subject? Look at p. 2595 of the documentation: http://www.math.wpi.edu/saspdf/stat/chap49.pdf

Using the counting process would involve change the structure of your data somewhat to be able to write the model statement (that is, you need to specify the beginning and end points of the interval).