I would appreciate if you could let me know whether it is correct to use parametric panel survival regression for my data. If so, could you please let me know how to test proportionality hazard assumption? My data set is as follows:
I used the following commands to do survival analysis:
. xtset id time, yearly
. stset time, failure(event==1)
. xtstreg x1 x2 x3 x4 x5, dist(exponential)
. estat ic
The estimated regression:
Random-effects exponential regression Number of obs = 1,554
Group variable: id Number of groups = 152
Obs per group:
min = 2
avg = 10.2
max = 12
Integration method: mvaghermite Integration pts. = 12
Wald chi2(5) = 84.85
Log likelihood = -209.83151 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
_t | Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
x1 | .7474386 .5523192 -0.39 0.694 .1756228 3.181048
x2 | .0269099 .0935812 -1.04 0.299 .0000295 24.54876
x3 | 814.9069 1289.601 4.24 0.000 36.64971 18119.46
x4 | 1.032383 .4342369 0.08 0.940 .4527012 2.354344
x5 | .0005696 .0019129 -2.22 0.026 7.89e-07 .4112869
_cons | .0000404 .0000722 -5.66 0.000 1.22e-06 .0013406
-------------+----------------------------------------------------------------
/sigma2_u | .4575335 .352393 .1011175 2.070235
------------------------------------------------------------------------------
LR test vs. exponential model: chibar2(01) = 2.69 Prob >= chibar2 = 0.0504
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 1,554 . -209.8315 7 433.663 471.1031
-----------------------------------------------------------------------------
Note: N=Obs used in calculating BIC; see [R] BIC note.
Best regards,
ID | TIME | EVENT | x1 | x2 | x3 | x4 | x5 |
1 | 1 | 0 | 1.281024 | 0.022934 | 0.874538 | 1.216367 | 0.06094 |
1 | 2 | 0 | 1.270012 | 0.006454 | 0.820673 | 1.004861 | -0.01408 |
1 | 3 | 0 | 1.05289 | -0.05938 | 0.922422 | 0.729264 | 0.020476 |
1 | 4 | 0 | 1.113115 | -0.01523 | 0.858878 | 0.809738 | 0.076037 |
1 | 5 | 1 | 1.219644 | -0.05868 | 0.887088 | 0.484342 | 0.009778 |
2 | 1 | 0 | 1.062264 | 0.107021 | 0.814602 | 0.835928 | 0.19996 |
2 | 2 | 0 | 1.055786 | 0.081916 | 0.879486 | 0.686728 | 0.142627 |
2 | 3 | 0 | 0.970588 | 0.076064 | 0.906775 | 0.809796 | 0.165915 |
2 | 4 | 0 | 1.058996 | 0.130203 | 0.818112 | 0.875989 | 0.234452 |
2 | 5 | 0 | 1.124525 | 0.147841 | 0.75871 | 1.079925 | 0.276444 |
2 | 6 | 0 | 1.599781 | 0.262461 | 0.546151 | 1.312741 | 0.369479 |
2 | 7 | 0 | 1.575608 | 0.262096 | 0.564481 | 1.156476 | 0.348624 |
2 | 8 | 0 | 1.544273 | 0.240911 | 0.590729 | 1.07697 | 0.325612 |
2 | 9 | 0 | 1.721708 | 0.215246 | 0.552291 | 0.841011 | 0.2935 |
2 | 10 | 0 | 1.723163 | 0.20863 | 0.533981 | 0.786512 | 0.293033 |
2 | 11 | 0 | 1.630677 | 0.186235 | 0.547719 | 0.728193 | 0.273577 |
2 | 12 | 0 | 2.172313 | 0.319455 | 0.441393 | 0.946985 | 0.427395 |
3 | 1 | 0 | 0.874395 | -0.03468 | 0.793502 | 0.609515 | -0.00263 |
3 | 2 | 1 | 0.825239 | -0.14194 | 0.952213 | 0.57288 | -0.01915 |
. xtset id time, yearly
. stset time, failure(event==1)
. xtstreg x1 x2 x3 x4 x5, dist(exponential)
. estat ic
The estimated regression:
Random-effects exponential regression Number of obs = 1,554
Group variable: id Number of groups = 152
Obs per group:
min = 2
avg = 10.2
max = 12
Integration method: mvaghermite Integration pts. = 12
Wald chi2(5) = 84.85
Log likelihood = -209.83151 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
_t | Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
x1 | .7474386 .5523192 -0.39 0.694 .1756228 3.181048
x2 | .0269099 .0935812 -1.04 0.299 .0000295 24.54876
x3 | 814.9069 1289.601 4.24 0.000 36.64971 18119.46
x4 | 1.032383 .4342369 0.08 0.940 .4527012 2.354344
x5 | .0005696 .0019129 -2.22 0.026 7.89e-07 .4112869
_cons | .0000404 .0000722 -5.66 0.000 1.22e-06 .0013406
-------------+----------------------------------------------------------------
/sigma2_u | .4575335 .352393 .1011175 2.070235
------------------------------------------------------------------------------
LR test vs. exponential model: chibar2(01) = 2.69 Prob >= chibar2 = 0.0504
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | Obs ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 1,554 . -209.8315 7 433.663 471.1031
-----------------------------------------------------------------------------
Note: N=Obs used in calculating BIC; see [R] BIC note.
Best regards,
Comment