[R] Fwd: Re: How to obtain nonparametric baseline hazard estimates in the gamma frailty model?

[Terry Therneau]

-------- Original Message --------
Subject: Re: How to obtain nonparametric baseline hazard estimates in the gamma frailty model?
Date: Mon, 04 Nov 2013 17:27:04 -0600
From: Terry Therneau <therneau.terry at mayo.edu>
To: Y <yuhanusa at gmail.com>

The cumulative hazard is just -log(sfit$surv).
The hazard is essentially a density estimate, and that is much harder.  You'll notice that
everyone does CDF curves for survival data ( Kaplan-Meier = estimate of 1-CDF), but no one
does histograms, which estimate a density.  That isn't because we wouldn't like density
estimates.

Terry T.


On 11/04/2013 04:47 PM, Y wrote:
> Hi Dr. Therneau,
>
> Thanks very much for your kind help! Does survfit() just give me the survival curve? What
> I wanted is the baseline hazard estimates, i.e., lambda_{0} (t). How can I obtain this
> estimate from coxph()? Or using basehaz()?
>
> Thanks,
> YH
>
>
>
>
>
>
>
>
>
>
> On Mon, Nov 4, 2013 at 5:29 PM, Terry Therneau <therneau at mayo.edu
> <mailto:therneau at mayo.edu>> wrote:
>
>
>
>     On 11/03/2013 05:00 AM, r-help-request at r-project.org
>     <mailto:r-help-request at r-project.org> wrote:
>
>         I can easily get the other parameter estimates by using coxph() but don't
>         know how to get the baseline hazard estimates from it. Does anyone know how
>         to obtain nonparametric baseline hazard estimates in the gamma frailty
>         model?
>
>         Thanks,
>         YH
>
>
>     I don't see what the problem is.
>
>     fit1 <- coxph(Surv(time, status) ~ age + ph.ecog + frailty(inst), data=lung)
>     sfit <- survfit(fit1)
>     plot(sfit)
>
>
>     Please give an actual example of the problem.
>
>     Terry Therneau
>
>