Re: Survival analysis

[Bassam A Dahman/HSC/VCU <bdahman@vcu.edu>]

For the survivors you have to put  (Survival
time = the  time till end of the study), it cannot be missing or a
fake value

Bassam

"Nader Deeb" <Nader.Deeb@sygeninternational.com>
Sent by: jmp-l-owner@lists.biostat.wustl.edu

02/01/2006 10:22 PM

Please respond to
jmp-l@lists.biostat.wustl.edu

To	<jmp-l@lists.biostat.wustl.edu>
cc
Subject	[jmp-l] Survival analysis

Subject: Survival data analysis using
Fixed Censoring (right censoring).
 
I have two treatments and group of
subjects within each treatment that some have died during the test period
while others survived to the end.
 
Those that died have a `death_age',
while all the survivals
(censored) have missing values.
 
I produced a new column called censored
where dead animals get a value of 0 and survivals get a value of 1.
 
The objective is to test if there is
a difference in survivability between the two treatments.
 
Is this the right way to arrange this
data?
Should I give the survival a `fake'
death_age (999)?
What option of the survival analysis
would work best (proportional hazard, survival, parametric etc.).
 
Thanks.