Date:         Sun, 17 Dec 2006 18:00:04 -0800
Reply-To:     Haris <Karovaldas@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Haris <Karovaldas@GMAIL.COM>
Organization: http://groups.google.com
Subject:      Re: How to allow for Left Censoring in PROC PHREG?
Comments: To: sas-l@uga.edu
In-Reply-To:  <1166405316.875594.168600@73g2000cwn.googlegroups.com>
Content-Type: text/plain; charset="us-ascii"

Look up "Left Truncation of Failure Times" under PHREG in SAS help.

proc phreg data=one;
      model (t1,t2)*dead(0)=x1-x10;
      baseline out=out2 survival=s;
      title 'Counting Process Style of Input';
run;

I forget now but there are rules about what missing t1 and t2 mean.  If
you have both times then you have left and right truncation interval.
If you have only one the meaning is different depending on which one it
is.

Allison's book on Survival Analysis from SAS press covers the topic
really well.  If you don't find the answer by tomorrow, let me know.  I
have a copy at work.


actstudent wrote:
> Hi,
>
> I've been looking at SAS procedure PROC PHREG to do some survival
> analysis on data with both left and right censoring. I haven't been
> able to find any example through google. Does anyone have any examples
> to share?
>
> The dataset looks like
> Code:
> Observed_    Observed_
>   Since         To        Days    Status
>
> 31MAY2005    31OCT2006    3620       0
> 31MAY2005    31OCT2006    3842       0
> 31MAY2005    31OCT2006    2325       0
> 31MAY2005    31OCT2006     732       0
> 31JUL2006    31OCT2006      67       0
> 31MAY2005    31OCT2006    3852       0
> 31MAY2005    31OCT2006    2752       0
>
>
> where days is age at exit (either through the end of investigation, or
> have "died") Status indicates censoring.
>
> There are examples of right censoring, but I couldn't find any examples
> about left censorying.
>
> Thanks.