The curves generated after a Cox model are always going to be parallel because the Cox model assumes proportional hazards. If your Kaplan-Meier curves are crossing, this could indicate that the hazards are not proportional. It depends where the cross-over occurs. They will often cross-over towards the end of follow-up, but that's usually because the data is getting sparse and the survival estimates are becoming a bit erratic.
If the cross-over occurs at a time point where you still have a reasonable amount of data, then you need to check the proportionality assumption in the Cox model.
______________________________________________
Kieran McCaul MPH PhD
WA Centre for Health & Ageing (M573)
University of Western Australia
Level 6, Ainslie House
48 Murray St
Perth 6000
Phone: (08) 9224-2701
Fax: (08) 9224 8009
email: [email protected]
http://myprofile.cos.com/mccaul
http://www.researcherid.com/rid/B-8751-2008
______________________________________________
Man is a credulous animal, and must believe something; in the absence of good grounds for belief,
he will be satisfied with bad ones. Bertrand Russell
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Ricardo Ovaldia
Sent: Tuesday, 9 June 2009 7:33 AM
To: [email protected]
Subject: Re: st: Adjusted Kaplan-Meier curves
Thank you Maarten.
That is what exactly what I did, except that I used the -at()- option to plot one curve for each drug. I used the dummies to do that such that for example for drug=1 it would be -at(_Idrug_2=0 _Idrug_3=0)- and so on. However the curves are parallel, with jumps at every death and do not look like the unadjusted curves.
Ricardo.
Ricardo Ovaldia, MS
Statistician
Oklahoma City, OK
--- On Mon, 6/8/09, Maarten buis <[email protected]> wrote:
> From: Maarten buis <[email protected]>
> Subject: Re: st: Adjusted Kaplan-Meier curves
> To: [email protected]
> Date: Monday, June 8, 2009, 11:12 AM
>
> --- On Mon, 8/6/09, Ricardo Ovaldia wrote:
> > I have been asked to plot Kaplan-Meier curves adjusted
> for
> > covariates, such as age, gender, race.
> > My thought was to use -stcox- to adjust and then plot
> the
> > adjusted survival using -stcurve-.
> > But I am not sure I am doing this correctly. The KM
> curves
> > plotted with -sts graph,by()- crossover, but those
> plotted
> > with -stcurve- do not and also they have a lot more
> steps
> > than the original curves. Any ideas?
>
> Sounds like a reasonable strategy to me. I don't have any
> concrete ideas, except that I added the example below of
> how I would do this. Maybe you can see something in my
> example that is different from what you did (no guarantees
> that what did is better than what you did though).
>
> *--------- begin example --------
> sysuse cancer, clear
> stset studytime, failure(died)
> xi: stcox i.drug age, basesurv(S)
> stcurve, survival
> *--------- end example ----------
>
> Hope this helps,
> Maarten
>
> -----------------------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
> http://home.fsw.vu.nl/m.buis/
> -----------------------------------------
>
>
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
