Checking assumptions of linearity and proortional hazards in the cox model

[kende jan]

Dear All,

How can I check with spss15, the assumptions of linearity and proportional hazards in the cox model ?
Thanks
Jan


      _____________________________________________________________________________
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail http://mail.yahoo.fr

====================To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Marta Garcia-Granero]

Hi Jan

For the linearity assumption (I expect you are talking about
quantitative predictors, someone asked me the same question once for a
polytomous variable... sigh), I suppose you have to work as in logistic
regression (see Hosmer&Lemeshow book "Applied Logistic Regression").
Let's call your your quantitative predictor "QP" (to keep the
explanation short):

- Compute the following variable:  LackLin=QP*Ln(QP)
- Add Lacklin to your model (together with QP, of course) and check its
significance. If it is significant, then there is a significant
deviation from linearity.

For the proportional hazards assumption, use a Cox model with
time-dependent covariates (I recommend you to use the GUI and paste the
syntax):

- Define T_COV_=QP*T_
- Add T_COV_ to your model (together with QP, of course). If T_COV_ is
significant, then the proportional hazards assumption fails.

Just in case, repeat the previous steps using T_COV_=QP*Ln(T_)
(I read somewhere it was a better way of checking prop hazards, sorry,
no ref. at hand).

I'll not be able to answer for quite a long time (I have to go to the
University to try to teach 30 graduate students statistics with SPSS,
and they are a talkative and restless bunch!)

I Might be able to answer after 20:00 CET, just in case you need
anything else.

Best regards,
Marta

kende jan escribió:

> Dear All,
>
> How can I check with spss15, the assumptions of linearity and proportional hazards in the cox model ?
> Thanks
> Jan
>
>
>       _____________________________________________________________________________
> Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail http://mail.yahoo.fr
>
> ===================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD