Survival Analysis Help

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Survival Analysis Help

Fahim Jafary
I have a dataset in which I'm looking at survival after a
procedure.  Patients are followed up by looking at their medical
record charts.  I have the following variables (amongst many):

date.procedure
date.death
date.last-follow-up
death (yes/no)

The follow-up period has a median of around 1.5 year (range days to 4
years) but for this particular analysis I want to plot a survival
curve and LIMIT my analysis to 30 days and one year.  In other words
I want to plot a survival curve that starts at time 0 and ends at 30
days (or 1 year).

I'm not sure how to go about the computing and coding in SPSS to
"draw a line" at 30 days (within which there will be those who have
died, those who are lost to follow up (censored) and those who don't
die (also censored).  To give an example

1.  say there's a case who has a procedure date and died at 6 months
but at 30 days was alive how do I get to declare this subject
censored and get the computer to put his follow-up time as 30 days.
I'm thinking the following:

IF DATEDIFF(date.death,date.procedure, "days") > 30 days death.30days = 0.
IF DATEDIFF(date.death,date.procedure, "days") <= 30 days death.30days = 1.

[this declares the subject a 30 day mortality].  But now how do I
compute a follow up time (which in this case is clearly 30 days as
the subject was alive till 6 months) ?


2.  say there's a subject that is alive at 20 days and then is
lost.  How do I get SPSS to declare this subject as censored yet with
a follow-up time of 20 days

Many thanks

Fahim H. Jafary, MD, FACC
Karachi, Pakistan