Hi,
I have count-time survival data that contains the time variable to event,
the number of
failures at each time, and then a number of variables that describe the
various ways in
which a subject can be censored. For example:
suvival_time n_dead n_censor_1 n_censor_2 n_censor_3
3 2 1 3 0
5 0 3 0 1
The ctset syntax does not appear to permit a varlist for ncensor, so I
create a new
variable n_censor, followed by ctset & cttost:
egen n_censor = rsum(n_censor_1 n_censor_2 n_censor_3)
ctset survival_time n_dead n_censor
cttost, clear
list
+--------------------------------------------------------------------------
----+
surviv~e n_dead n_cens~1 n_cens~2 n_cens~3 w _st _d _t _t0
---------------------------------------------------------------------------
---
1. 3 0 1 3 0 4 1 0 3 0
2. 5 0 3 0 1 4 1 0 5 0
3. 3 1 1 3 0 2 1 1 3 0
+--------------------------------------------------------------------------
----+
1 - Why has cttost dropped the variable n_censor?
2 – One unintended consequence of this process is that I have treated the
censoring
variables as enduring group variables. And thus they have been expanded
by the cttost.
This problem is even worse when I expand on w in order to use stci. I
would like to later
perform analyses on censoring by group. Should I partition the data,
have one data set
for survival and another for censoring analyses? Or should I merge back
censoring
information to the first observation per each unit of time? Or finally
is there a way of
correctly incorporating the censoring values?
Thank you so much for your help.
Douglas Crawford
Ph.D. Candidate, Dept. of Biochemistry
UCSF
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/