Dear Stata experts,
I have a dataset with typical survival type data, and I am interested in determining time-dependent AUC's after competing risk regressions. The built in Stata function stcrreg does not appear to have a post-estimation AUC (ROC) function that I can identify (time-dependent, nor otherwise). There is an excellent 3rd party package (strocccurve) that can compute time-dependent AUC's, however, this package does not account for competing risk, and cannot estimate time-dependent AUC for variables that are categorical.
I am attaching the following dataset below.
the dataset contains the following:
Time Variable: timevar
Fail Variable: failvar: 0=no failure, 1=failure, 2= competing risk censor
Categorical Biomarker indicator variable: Categorical_var
Continuous Biomarker variable: Continuous_var
Is it possible to perform an AUC calculation after a competing risk survival regression on the categorical variable and continuous variables (separately) at a specified time point, like timevar==2.5?
Any help is greatly appreciated
Thanks, JT
I have a dataset with typical survival type data, and I am interested in determining time-dependent AUC's after competing risk regressions. The built in Stata function stcrreg does not appear to have a post-estimation AUC (ROC) function that I can identify (time-dependent, nor otherwise). There is an excellent 3rd party package (strocccurve) that can compute time-dependent AUC's, however, this package does not account for competing risk, and cannot estimate time-dependent AUC for variables that are categorical.
I am attaching the following dataset below.
the dataset contains the following:
Time Variable: timevar
Fail Variable: failvar: 0=no failure, 1=failure, 2= competing risk censor
Categorical Biomarker indicator variable: Categorical_var
Continuous Biomarker variable: Continuous_var
Is it possible to perform an AUC calculation after a competing risk survival regression on the categorical variable and continuous variables (separately) at a specified time point, like timevar==2.5?
Any help is greatly appreciated
Code:
stset timevar, failure(failvar==1)
Code:
* Example generated by -dataex-. For more info, type help dataex clear input float(Categorical_var Continuous_var failvar timevar) 0 .8280001 0 1.9575633 1 1.599 0 4.958248 1 2.121 0 2.0698152 2 2.673 0 1.4592744 1 2.34 0 4.0985627 1 1.503 0 2.510609 1 1.494 0 3.759069 1 2.349 0 5.234771 1 2.397 0 6.19028 2 .429 1 -.16427104 2 2.112 1 6.313484 0 .14400002 0 3.9890485 1 2.853 0 5.021218 1 1.836 2 .56399727 1 1.437 0 .18069816 3 4.308 0 3.978097 0 .999 0 7.394935 0 .37199995 0 1.6180698 0 1.455 2 .5968515 2 2.967 0 2.017796 1 .99 0 9.445585 0 1.104 0 4.405202 1 1.284 0 4.0958247 0 .714 0 .9007529 2 1.104 0 1.2073922 1 1.161 2 3.3483915 1 1.494 2 1.776865 0 1.341 0 4.1587954 0 .26699996 0 2.3408625 0 .78 0 4.657084 1 1.161 0 2.702259 1 .03900003 0 5.221081 2 3.897 1 .7474332 1 1.56 0 2.956879 1 2.349 0 6.510609 1 1.095 0 5.374401 1 1.0469999 0 5.295003 1 2.169 0 1.946612 2 2.073 0 .9801506 2 2.52 0 5.245722 2 1.713 0 6.617385 1 2.787 0 1.2539357 1 2.616 0 2.494182 1 .876 0 5.097878 1 .762 0 1.377139 1 2.283 0 4.25462 2 1.77 0 5.612594 1 1.494 0 1.9958932 2 3.756 1 .05749487 0 .885 0 3.671458 0 .78 0 5.804244 0 1.056 0 1.6208076 0 -.53099996 0 .7638603 0 .6 0 1.28679 3 2.91 0 1.522245 0 -.018000055 0 7.895962 0 .315 0 5.689254 1 1.608 0 2.0123203 2 1.2 0 7.904175 1 1.77 0 .3066393 1 1.665 0 2.740589 2 2.967 0 1.4592744 2 2.682 1 .09856263 0 .609 0 6.57358 0 .8939999 0 1.0431211 0 .837 0 .6078029 1 .486 2 . 1 1.836 0 3.3812456 1 1.941 0 . 1 .25800005 0 3.947981 1 1.446 0 3.3648186 0 1.104 0 .6926762 1 2.796 0 .008213553 0 -.588 0 5.174538 1 -.18900003 0 4.5667353 0 .714 0 1.2073922 1 1.494 0 1.3826146 0 .933 0 3.12115 0 .837 0 2.469541 1 1.722 0 2.255989 1 .705 0 6.321697 1 .54300004 0 4.1971254 0 .885 0 .7063655 2 4.0230002 0 1.1608487 3 2.055 0 1.560575 0 .09599997 0 6.811773 0 .5519999 0 3.455168 1 1.551 0 1.8261465 2 2.274 0 .17522246 0 .723 2 3.285421 1 2.454 0 1.779603 1 2.007 0 3.4058864 0 1.389 0 1.393566 0 .4518 0 10.182067 0 -.987 0 7.036276 1 .933 0 3.258042 0 .999 0 7.70705 0 .714 0 9.804244 2 2.559 2 9.281314 0 .942 0 1.544148 end label values Categorical_var Categorical_var label def Categorical_var 0 "Best", modify label def Categorical_var 1 "Middle", modify label def Categorical_var 2 "High", modify label def Categorical_var 3 "Worst", modify label values failvar failvar label def failvar 0 "no fail", modify label def failvar 1 "metastasis", modify label def failvar 2 "Non Cancer Death", modify
Comment