st: multiple failure-time data (Prentice, William & Peterson)

["Kallimanis, Bellinda" <bkallimanis@fmhi.usf.edu>]

Hi!

I am trying to work out how to set up my data to run a condition risk
set model (time from entry) to analyse my multiple failure data. 

My time variable is discrete (measures taken approximately every 90
days) and there are at least three observations for each subject. I have
time varying covariates and interactions. 

I have read the FAQ "How do I analyze multiple failure-time data using
Stata?" And I'm not sure how to construct my data as the FAQ says care
needs to be taken when coding the data. 

Currently a sample of what I have is


resident  los   wander   cps_cat   age   n  
  8372      0        1         1    82   1  
  8372     92        1         1    83   2  
  8372    182        1         1    83   3  
  8372    283        0         2    83   4  
  8372    365        0         2    83   5  
  8611      0        1         1    76   1  
  8611     95        0         1    76   2  
  8611    204        0         1    77   3  
  8611    294        0         1    77   4  
  8611    384        0         1    77   5  
  8611    474        0         1    78   6  
  8611    533        0         1    78   7  
  8611    623        0         1    78   8  
  8674      0        0         1    68   1  
  8674     90        0         1    69   2  
  8674    179        1         1    69   3


LOS is calculated as date of entry into nursing home until discharge/end
of study.