Hi all, I am trying to use stcox to create a multivariable cox proportional hazard model. I am running Stata 15.1 on macOS Big Sur (version 11.2). This is my data (with some changes for confidentiality purposes).

Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input str11 proc_type float(re_op postop_month_reop diagnosis severity)
"trab" 0  3.766666 . .
"tube" 1        2.7 . .
"trab" 0   41.03333 . 2
"tube" 0  176.16667 . 2
"tube" 1      109.9 . 3
"trab" 0   3.23333 . 3
"trab" 0   20.06667 . .
"tube" 0  26.766666 6 3
"trab" 1   95.96667 . .
"trab" 0  65.833336 . .
"tube" 0      178.1 . .
"tube" 0   13.633333 6 3
"tube" 0   3.766667 . .
"tube" 0       10.7 . 3
"trab" 0   3.76667 . 3
"tube" 0       23.1 6 .
"trab" 0  33.666668 . .
"trab" 0       14.6 . .
"trab" 0  16.766666 . .
"trab" 0       3.9 . .
"tube" 0       36.6 . .
"tube" 0  16.333334 2 3
"trab" 0       9.8 . .
"tube" 0       19.7 6 3
"trab" 0        80 . 1
"trab" 0   89.16666 . 3
"trab" 1   78.13333 . 3
"trab" 1       4.7 . 3
"trab" 0       15.9 . .
"trab" 0        9.5 . .
"tube" 0       30.9 . 3
"trab" 0   7.666667 2 3
"tube" 1   17.466667 1 .
"trab" 0       23.9 3 3
"tube" 0          4 2 3
"tube" 0  55.066666 . .
"tube" 1       16.8 . .
"trab" 0   33.46667 . .
"tube" 0  36.633335 . .
"trab" 0  121.63333 . 2
"tube" 0  27.366667 6 1
"tube" 0       27.6 6 1
"trab" 0       28.2 5 3
"tube" 0       64.7 5 .
"trab" 0  12.033334 4 3
"tube" 0  15.633333 . 3
"trab" 0   44.56667 . .
"trab" 0       56.3 . .
"trab" 0   .26667 . .
"trab" 0   54.56667 2 .
"trab" 0   53.33333 . 2
"trab" 0  22.233334 2 3
"tube" 0   50.26667 . .
"tube" 0   69.96667 6 .
"trab" 0  116.66666 . .
"trab" 0   41.96667 . 2
"tube" 0       77.5 . 3
"trab" 0  106.76667 . .
"tube" 0   9.466666 6 2
"trab" 0          7 2 3
"tube" 0       60.5 . .
"trab" 1   78.16666 . .
"tube" 0   5.666667 . .
"tube" 0  34.866665 . .
"trab" 0       72.1 . .
"trab" 0   73.03333 . .
"tube" 1  .26666668 . .
"trab" 0   16.566667 5 3
"trab" 1   18.46667 . .
"tube" 0   77.233333 . .
"tube" 1   2.76667 . .
"tube" 0         90 . .
"trab" 0   13.46667 2 .
"tube" 0  7.733334 1 .
"tube" 0   6.53333 1 .
"tube" 0  99.866667 6 3
"trab" 0  31.866667 . .
"trab" 0   9.066667 . .
"tube" 0  28.233334 . 3
"tube" 0  107.13333 . .
"tube" 0   62.76667 . .
"tube" 0       25.2 2 .
"trab" 0      141.1 . 3
"tube" 0   60.46667 . 3
"trab" 0         13 . .
"trab" 1       38.5 . .
"trab" 0        4.6 . 3
"trab" 1   52.73333 . .
"trab" 1  28.233334 . .
"trab" 0       42.9 3 1
"trab" 0       45.3 . 1
"tube" 1 .033333335 6 3
"trab" 0   53.46667 . 3
"trab" 0   66.86667 2 .
"tube" 0   3.63334 6 .
"trab" 0  40.13333 . 3
"tube" 0   115.53333 2 .
"tube" 0   31.06667 . .
"trab" 0  1.866667 . 3
"tube" 1          4 2 3
end
When I create a hazard model using stcox including both the severity and diagnosis variable I get the following output:

code:
stset postop_month, failure(re_op) exit(time 36)
generate procedure_type=1 if proc_type=="trab"
replace procedure_type=2 if proc_type=="tube"
stcox i.procedure_type i.diagnosis i.severity

Code:
stcox i.procedure_type i.diagnosis i.severity

         failure _d:  re_op
   analysis time _t:  postop_month_reop
  exit on or before:  time 36

Iteration 0:   log likelihood = -5.8348107
Iteration 1:   log likelihood = -4.3840398
Iteration 2:   log likelihood =  -4.075028
Iteration 3:   log likelihood =  -4.015207
Iteration 4:   log likelihood = -3.9966115
Iteration 5:   log likelihood = -3.9898231
Iteration 6:   log likelihood = -3.9873317
Iteration 7:   log likelihood =  -3.986416
Iteration 8:   log likelihood = -3.9860792
Iteration 9:   log likelihood = -3.9859553
Iteration 10:  log likelihood = -3.9859098
Iteration 11:  log likelihood =  -3.985893
Iteration 12:  log likelihood = -3.9858868
Iteration 13:  log likelihood = -3.9858846
Iteration 14:  log likelihood = -3.9858837
Iteration 15:  log likelihood = -3.9858834
Iteration 16:  log likelihood = -3.9858833
Iteration 17:  log likelihood = -3.9858833
Iteration 18:  log likelihood = -3.9858832
Iteration 19:  log likelihood = -3.9858832
Iteration 20:  log likelihood = -3.9858832
Iteration 21:  log likelihood = -3.9858832
Iteration 22:  log likelihood = -3.9858832
Iteration 23:  log likelihood = -3.9858832
Iteration 24:  log likelihood = -3.9858832
Iteration 25:  log likelihood = -3.9858832
Iteration 26:  log likelihood = -3.9858832
Iteration 27:  log likelihood = -3.9858832
Iteration 28:  log likelihood = -3.9858832
Iteration 29:  log likelihood = -3.9858832
Iteration 30:  log likelihood = -3.9858832
Iteration 31:  log likelihood = -3.9858832
Iteration 32:  log likelihood = -3.9858832
Iteration 33:  log likelihood = -3.9858832
Refining estimates:
Iteration 0:   log likelihood = -3.9858832
Iteration 1:   log likelihood = -3.9858832
Iteration 2:   log likelihood = -3.9858832

Cox regression -- no ties

No. of subjects =           19                  Number of obs    =          19
No. of failures =            2
Time at risk    =  338.5000034
                                                LR chi2(5)       =        3.70
Log likelihood  =   -3.9858832                  Prob > chi2      =      0.5937

----------------------------------------------------------------------------------
              _t | Haz. Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------+----------------------------------------------------------------
2.procedure_type |   2.49e+17   4.47e+25     0.00   1.000            0           .
                 |
       diagnosis |
              3  |   60.43064          .        .       .            .           .
              4  |    19.9061   5.25e+09     0.00   1.000            0           .
              5  |    19.9061   4.49e+09     0.00   1.000            0           .
              6  |   .6708204   .9501599    -0.28   0.778     .0417784    10.77111
                 |
        severity |
              2  |   23.28885   4.51e+09     0.00   1.000            0           .
              3  |   1.96e+17          .        .       .            .           .
----------------------------------------------------------------------------------
There is no output in the model for diagnosis==2. If I remove the "severity" variable from the model then the model creates an output for diagnosis==2.

Code:
stcox i.procedure_type i.diagnosis

         failure _d:  re_op
   analysis time _t:  postop_month_reop
  exit on or before:  time 36

Iteration 0:   log likelihood = -9.6344964
Iteration 1:   log likelihood = -8.5303892
Iteration 2:   log likelihood =   -7.64667
Iteration 3:   log likelihood = -7.5746907
Iteration 4:   log likelihood = -7.5582573
Iteration 5:   log likelihood = -7.5522466
Iteration 6:   log likelihood = -7.5500349
Iteration 7:   log likelihood = -7.5492213
Iteration 8:   log likelihood = -7.5489219
Iteration 9:   log likelihood = -7.5488118
Iteration 10:  log likelihood = -7.5487713
Iteration 11:  log likelihood = -7.5487564
Iteration 12:  log likelihood = -7.5487509
Iteration 13:  log likelihood = -7.5487489
Iteration 14:  log likelihood = -7.5487482
Iteration 15:  log likelihood = -7.5487479
Iteration 16:  log likelihood = -7.5487478
Iteration 17:  log likelihood = -7.5487477
Iteration 18:  log likelihood = -7.5487477
Iteration 19:  log likelihood = -7.5487477
Iteration 20:  log likelihood = -7.5487477
Iteration 21:  log likelihood = -7.5487477
Iteration 22:  log likelihood = -7.5487477
Iteration 23:  log likelihood = -7.5487477
Iteration 24:  log likelihood = -7.5487477
Iteration 25:  log likelihood = -7.5487477
Iteration 26:  log likelihood = -7.5487477
Iteration 27:  log likelihood = -7.5487477
Iteration 28:  log likelihood = -7.5487477
Iteration 29:  log likelihood = -7.5487477
Iteration 30:  log likelihood = -7.5487477
Iteration 31:  log likelihood = -7.5487477
Iteration 32:  log likelihood = -7.5487477
Iteration 33:  log likelihood = -7.5487477
Iteration 34:  log likelihood = -7.5487477
Iteration 35:  log likelihood = -7.5487477
Refining estimates:
Iteration 0:   log likelihood = -7.5487477
Iteration 1:   log likelihood = -7.5487477

Cox regression -- no ties

No. of subjects =           31                  Number of obs    =          31
No. of failures =            3
Time at risk    =  615.6333456
                                                LR chi2(6)       =        4.17
Log likelihood  =   -7.5487477                  Prob > chi2      =      0.6535

----------------------------------------------------------------------------------
              _t | Haz. Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------+----------------------------------------------------------------
2.procedure_type |   7.78e+15   5.49e+23     0.00   1.000            0           .
                 |
       diagnosis |
              2  |   .5697548   .8061791    -0.40   0.691     .0355855     9.12226
              3  |   .1304417   3.29e+07    -0.00   1.000            0           .
              4  |   .1337592   6.59e+07    -0.00   1.000            0           .
              5  |   7.48e-17   1.21e-08    -0.00   1.000            0           .
              6  |   .2325807   .3325712    -1.02   0.308     .0141067    3.834616
----------------------------------------------------------------------------------
Why is this happening? And will it be possible to include both variables in the model and get an output for diagnosis==2? Thank you!