s-news
[Top] [All Lists]

yearly hazard with std error

To: "S-news" <s-news@lists.biostat.wustl.edu>
Subject: yearly hazard with std error
From: "Fredrik Lundgren" <fredrik.lundgren@norrkoping.mail.telia.com>
Date: Tue, 23 Dec 2003 09:13:03 +0100
Hello,

I'm working with a large database with aneurysm operations and want yearly 
hazards post op for different age groups.

I get cummulative survival (excluding the first month) with:

First model

basal.op.surv <- coxph(Surv(I(time-30), event) ~ age, data = el.aaa[el.aaa$sex 
== 'male' & el.aaa$time > 30,])
summary(survfit(basal.op.surv, newdata = data.frame(age = 66:80)), times = 
365.25 * 1:8)

output

Call: survfit.coxph(object = basal.op.surv, newdata = data.frame(age = 66:80))

 time n.risk n.event survival1 survival2 survival3 survival4 survival5 
survival6 survival7 survival8          -    etc
  365   2663     169     0.958     0.956     0.954     0.951     0.949     
0.946     0.943     0.940          -
  730   2187     132     0.920     0.916     0.911     0.907     0.902     
0.897     0.892     0.886          -
  1096   1714     134     0.873     0.866     0.860     0.853     0.845     
0.838     0.829     0.821        -
 1461   1339     106     0.827     0.818     0.809     0.800     0.790     
0.780     0.770     0.759         -
 1826    927     106     -              -            -            -            
-                -            -            -               -
  -          -          -         -             -            -            -     
        -               -            -             -               -
  etc

Second model

basal.op.surv <- coxph(Surv(I(time-30), event) ~ age, data = el.aaa[el.aaa$sex 
== 'male' & el.aaa$time > 30,])
for (i in 65:66) print(summary(survfit(basal.op.surv, newdata = data.frame(age 
= i)), times = 365.25 * 1:5))

output

Call: survfit.coxph(object = basal.op.surv, newdata = data.frame(age = i))

 time n.risk n.event survival std.err lower 95% CI upper 95% CI 
  365   2663     169    0.961 0.00338        0.954        0.967
  730   2187     132    0.924 0.00517        0.914        0.934
 1096   1714     134    0.879 0.00713        0.865        0.893
 1461   1339     106    0.835 0.00891        0.817        0.852
 1826    927     106    0.778 0.01104        0.756        0.800
Call: survfit.coxph(object = basal.op.surv, newdata = data.frame(age = i))

 time n.risk n.event survival std.err lower 95% CI upper 95% CI 
  365   2663     169    0.958 0.00346        0.952        0.965
  730   2187     132    0.920 0.00521        0.910        0.930
 1096   1714     134    0.873 0.00709        0.859        0.887
 1461   1339     106    0.827 0.00877        0.810        0.844
 1826    927     106    0.767 0.01080        0.746        0.788


How can I get yearly hazard with std error for each year?

Sincerely Fredrik Lundgren


<Prev in Thread] Current Thread [Next in Thread>
  • yearly hazard with std error, Fredrik Lundgren <=