Delayed entry due to missing X values in survival analysis?

[Bo Zhao]

Dear Statalisters,

I observe each subject’s event history since t=0 (when the subject is at risk of failure). However, I don’t have values for the explanatory variable X for the entire period. For example,

ID begin end fail X
1 0 1 0 .
1 1 2 0 .
1 2 3 0 10
1 3 4 1 15

Is this considered “delayed entry” in the survival analysis? Is it proper to use the following stset command?
stset end, id(id) fail(fail) enter (x !=.)

Thanks.

Bo

[Steve Samuels]

Stat will ignore the missing x's, but in case there gaps, use the time0() option.

clear
input id  begin end fail x
1 0 1 0 .
1 1 2 0 4
1 3 4 1 15
2 0 1 0  .
2 1 2 0 .
2 2 3 0 16
2 3 4 0 .
2 4 5 1 5
3 0 10 0 14
end
list, sepby(id)
stset end ,  fail(fail) time0(begin) id(id)
order id _t0 _t
list, sepby(id)
/*
  +----------------------------------------------------+
     | id   _t0   _t   begin   end   fail    x   _st   _d |
     |----------------------------------------------------|
  1. |  1     0    1       0     1      0    .     1    0 |
  2. |  1     1    2       1     2      0    4     1    0 |
  3. |  1     3    4       3     4      1   15     1    1 |
     |----------------------------------------------------|
  4. |  2     0    1       0     1      0    .     1    0 |
  5. |  2     1    2       1     2      0    .     1    0 |
  6. |  2     2    3       2     3      0   16     1    0 |
  7. |  2     3    4       3     4      0    .     1    0 |
  8. |  2     4    5       4     5      1    5     1    1 |
     |----------------------------------------------------|
  9. |  3     0   10       0    10      0   14     1    0 |
     +----------------------------------------------------+
*/
stcox x
stcox x if x!=.  //same result
/*
. stcox x if x!=.

         failure _d:  fail
   analysis time _t:  end
                 id:  id

Iteration 0:   log likelihood = -1.3862944
Iteration 1:   log likelihood = -.95478346
Iteration 2:   log likelihood = -.92165942
Iteration 3:   log likelihood = -.91941357
Iteration 4:   log likelihood = -.91939296
Refining estimates:
Iteration 0:   log likelihood = -.91939296

Cox regression -- no ties

No. of subjects =            3                  Number of obs    =           5
No. of failures =            2
Time at risk    =           14
                                                LR chi2(1)       =        0.93
Log likelihood  =   -.91939296                  Prob > chi2      =      0.3339

------------------------------------------------------------------------------
          _t | Haz. Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
           x |   .7419454   .3291334    -0.67   0.501      .311008    1.769996
------------------------------------------------------------------------------

[Bo Zhao]

Thanks a lot, Steve. Always appreciate it.