[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: ST: Hazard rates translating into probabilities

From   Maarten buis <[email protected]>
To   [email protected]
Subject   Re: st: ST: Hazard rates translating into probabilities
Date   Wed, 23 Jan 2008 10:05:58 +0000 (GMT)

--- Michael Kregar <[email protected]> wrote:
> I have run the following cox proportional regression
> How can I obtain an overall probability of the above
> model when using data from say one observation (for
> the four independent variables above). I am looking
> for overall probabilities of failure (not hazard
> rate). Can anyone give me a one by one calculation?

This can be interpreted in two ways: 
1) you want to probability of failure as an end result 
(at t=infinity). In most parametric models (-streg-) 
this is easy: that probability is 1 for everyone. So 
called cure models explicitly allow for a given 
proportion of observations that will never experience 
the event (are "cured"). I am not so sure if a Cox 
model makes any statement about what hapens at 
t=infinity. I could imagine that it either leaves that 
undefined or that it assumes that in the end everybody 
dies. I find it implausable that you could derive a 
curefraction from a Cox model. For more on cure models 
in Stata see (Lambert 2007).

2) A more interesting way to interpret your question
is to assume that you are asking for 1-Survivor 
function, i.e. the probability that someone has 
died at or before some specific timepoint. You can 
compute those as in the example below. More 
interestingly (and easy) you can plot them using the 
-stcurve- command (see: -help stcurve-).

*--------------- begin example ------------------
sysuse cancer, clear
stset studytime, failure(died)
xi: stcox i.drug age, basesurv(base)
predict double xb, xb
gen S = base^exp(xb)
*---------------- end example -------------------
(For more on how to use examples I sent to the Statalist, see )

Lambert, Paul (2007) "Modeling of the cure fraction
in survival studies". The Stata Journal, 7(3), pp. 

Hope this helps,

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Sent from Yahoo! Mail - a smarter inbox

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index