s-news
[Top] [All Lists]

[S] Weibull-model with gamma-frailty

To: S-News <s-news@wubios.wustl.edu>
Subject: [S] Weibull-model with gamma-frailty
From: Geir Egil Eide <Geir.Eide@NHH.No>
Date: Fri, 26 May 2000 14:44:23 +0200
Organization: Norwegian School of Economics and Business Administration
Sender: owner-s-news@wubios.wustl.edu
Dear S+ users,

I have without success tried to reproduce the results of Aalen et al
(1995) Statistics in Medicine 14, 1819-1829 using the frailty options of
SurvReg in S-Plus 2000 for Windows95. Can anyone explain to me why? Here
is my S+ commands and the results for the model without covariates which
I believed should correspond to the estimates in Aalen et al's Table II:

> survReg(formula = Surv(Obstime, Response, type = "right") ~ 
> 1+frailty(Patientn),
+        data = aalentn2, na.action = na.exclude, dist = "weibull",
scale = 0, 
+        control = list(maxiter = 30, rel.tolerance = 1e-005, failure =
1))
Call:
survReg(formula = Surv(Obstime, Response, type = "right") ~ 1 +
frailty(Patientn), data = 
        aalentn2, na.action = na.exclude, dist = "weibull", scale = 0, control
= list(maxiter = 
        30, rel.tolerance = 1e-005, failure = 1))

                  coef se(coef)   se2 Chisq   DF        p 
      (Intercept) 3.49 0.178    0.104 386    1.0 0.0e+000
frailty(Patientn)                     111   23.3 1.9e-013

Scale= 0.639 

Iterations: 10 outer, 24 Newton-Raphson
     Variance of random effect= 0.584   EM likelihood = -385 
Degrees of freedom for terms=  0.3 23.3  1.0 
Likelihood ratio test=114  on 22.6 df, p=4.47e-014  n= 566 


Thank you,


Geir Egil Eide,
Biostatistician,
Centre for Clinical Research,
Haukeland University Hospital,
Bergen, Norway
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu.  To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message:  unsubscribe s-news

<Prev in Thread] Current Thread [Next in Thread>