I have used stpm3 to make a flexible survival model and would like to know if there's a natural way to interpret the HR for those terms with accompanying tvc's (I understand that the usual way to present results when splines are involved is to plot the results)
For example, for the output below, does the HR of .333409 for live birth (which also has a tvc specified) indicate the HR at the start of observation? Can the commensurate tvc value be used to calculate the HR for live birth at other times?
Just trying to get a handle on this. I have made a plot of how the HR for live birth varies over time but I suspect reviewers will want to know how to interpret the accompanying regression output.
Thanks.
For example, for the output below, does the HR of .333409 for live birth (which also has a tvc specified) indicate the HR at the start of observation? Can the commensurate tvc value be used to calculate the HR for live birth at other times?
Just trying to get a handle on this. I have made a plot of how the HR for live birth varies over time but I suspect reviewers will want to know how to interpret the accompanying regression output.
Thanks.
Code:
. stpm3 i.live_birth2 c.age_start##c.disease_years c.base_EDSS i.High_eff_DMT c.ARR i.sym_period i.co
> untry, scale(lncumhazard) tvc(i.live_birth2 c.age_start c.base_EDSS c.ARR) dftvc(1) df(1) eform
Iteration 0: Log likelihood = -1497.1973
Iteration 1: Log likelihood = -1459.1511
Iteration 2: Log likelihood = -1455.1277
Iteration 3: Log likelihood = -1455.1259
Iteration 4: Log likelihood = -1455.1259
Number of obs = 9,646
Wald chi2(12) = 244.43
Log likelihood = -1455.1259 Prob > chi2 = 0.0000
---------------------------------------------------------------------------------------------
| exp(b) Std. err. z P>|z| [95% conf. interval]
----------------------------+----------------------------------------------------------------
xb |
1.live_birth2 | .333409 .1316618 -2.78 0.005 .1537593 .7229583
age_start | 1.038147 .0160587 2.42 0.016 1.007145 1.070103
disease_years | 1.24285 .0647789 4.17 0.000 1.122156 1.376526
|
c.age_start#c.disease_years | .9942693 .0013643 -4.19 0.000 .991599 .9969468
|
base_EDSS | 4.940149 .904781 8.72 0.000 3.450196 7.073533
1.High_eff_DMT | .8193135 .1327575 -1.23 0.219 .5963835 1.125575
ARR | .71923 .0375318 -6.32 0.000 .6493059 .7966844
|
sym_period |
1995 - 2004 | 1.052205 .159497 0.34 0.737 .781758 1.416213
2005 - 2009 | .7154194 .1411554 -1.70 0.090 .4859761 1.05319
≥ 2010 | .4522372 .1079951 -3.32 0.001 .283203 .7221622
|
country |
CZ | 1.288201 .1411014 2.31 0.021 1.039318 1.596683
KW | .6773885 .2370614 -1.11 0.266 .3411522 1.345016
----------------------------+----------------------------------------------------------------
time |
_ns1 | 1.318926 .2663271 4.95 0.000 .7969347 1.840918
|
live_birth2#c._ns_tvc1 |
1 | .3429797 .1563769 2.19 0.028 .0364865 .6494729
|
c.age_start#c._ns_tvc1 | -.0004062 .0058406 -0.07 0.945 -.0118535 .0110411
|
c.base_EDSS#c._ns_tvc1 | -.2662749 .0730666 -3.64 0.000 -.4094829 -.123067
|
c.ARR#c._ns_tvc1 | .1144284 .0193021 5.93 0.000 .076597 .1522597
|
_cons | -6.755101 .7284361 -9.27 0.000 -8.182809 -5.327393
---------------------------------------------------------------------------------------------
