Re: st: survival analysis

[Steve Samuels <sjsamuels@gmail.com>]

One other thought: if your assessment detects an event that occurred since the
previous assessment, but cannot be exactly dated, then you have interval, not
right, censoring. Choices in Stata include -intcens- and -stpm- (findit) and
grouped data approaches mentioned recently by Stephen Jenkins in
http://www.stata.com/statalist/archive/2012-07/msg01095.html.

Steve


The ending date for a censored observation is the last one for which you know
that the subject did not have the outcome of interest. So it will _not_ be the
date of the missed assessment, but an earlier or later one. An earlier date
might be the date of the previous assessment. A later date could be one on
which you have contacted the subject or proxy and can tell from the
communication that that the subject did _not_ have the outcome of interest,
because, for example, they had no symptoms.


Steve
sjsamuels@gmail.com


On Aug 20, 2012, at 12:58 PM, Nikolaos Pandis wrote:

Hi to all.

I would like to ask the following:

For survival data (before stset) how would I indicate censored observations?
For the observations with complete information (entry/exit dates and outcome) I have coded the outcome as 0 or 1 but I am not sure how to code patients lost before the end of the study. 
Censored observations also have just entry dates. Would exit date (for censored observatiosn) be the end of study date or date of last assessment in which patient(s) went missing? 

Any comments would be appreciated.

Thank you,

Nick  


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/