Re: st: Landmark analysis

[Steven Samuels <sjsamuels@gmail.com>]

You must -stset- your data. Then -sts graph , hazard- will estimate  
the hazard function over the entire range of the data; you can add a  
line for the time of your landmark event and annotate it. See  -help-  
for "added_line_options" and  "added_text_options".

-stcox- and the contributed program -stpm2- ("findit") handle allow  
time-dependent covariates. With them you can test equality of hazard  
functions before and after any event. If the times of your landmark  
event are not constant, you can  get graphs in -stcox- and -stpm2- by  
creating time-dependent strata: one pre-event, one post-event. To  
create these you could, for example, -stsplit- the data at the time of  
the event.


Note that "STATA" is not a correct spelling of "Stata" (Statalist FAQ  
8.4).  Names all in upper case are acronyms, names whose letters are  
letters (usually the first) of words in the original, sometimes  
abandoned, name.  So in statistics: "SAS" = "Statistical Analysis  
System"; "SPSS" = "Statistical Package for the Social Sciences";   
"SUDAAN"= "SUrvey DAta ANalysis".  "Stata" is not such a word; its  
letters never stood for anything else.

Steve
sjsamuels@gmail.com

On Feb 1, 2011, at 7:58 PM, Sripal Kumar wrote:

Dear all,

I am interested in doing a landmark surival analysis- an analysis
where you compare treatment effect before and after a certain
landmark.  For example, I am interested in constructing a hazard plot
for two treaments from start of treament to 30 days and then from 30
days to end of follow-up in the same plot.  Is there a way to do this
in STATA.
thanks,
Sripal.
*
*   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/