Re: st: Fwd: Time varying covariate data format for the flexible joint modelling of longitudinal

[Steve Samuels <sjsamuels@gmail.com>]

Sachin <sbatra13@gmail.com>:

I'm not very familiar with the joint-modeling literature you refer to.
But for a survival analysis, -stsplit- is not relevant to your problem.

1. Your data are in wide format; -reshape- to long format, with one line
per date.

2. The result will be what the Survival Manual calls "snapshot" data.
Use -snapspan- to convert to "timescan" data, which is needed for
-stset- and -stcox-, among other commands.

See the manual entries for -snapscan- and -stset-.


Steve
sjsamuels@gmail.com


On Jul 3, 2013, at 9:44 AM, sachin batra wrote:

Hi,
I wish to perform an analysis to find association between changes in
longitudinal blood gas values and mortality. I belive this would
require flexible joint modelling of longitudinal as described at
http://www.stata.com/meeting/uk11/abstracts/UK11_crowther.pdf or at
http://www.stata-journal.com/article.html?article=st0289.
Although the implementation of code is very well described, my
question is how can I adapt stsplit command in stata to set the data
in the time varying covariate format-splitting the data into discrete
periods of time spent at each po2.

My data is for example in the following format

id date_oxygenation1  o2_1  date_oxygenation2       o2_2
date_oxygenation3        o2_3            date_oxygenation4
o2_4           status             status_date
1      1/1/2000               60        2/8/2000                  55
             2/8/2002                     70
4/8/2003                45                 0
1/1/2010
2      4/6/2000               77        9/11//2000                60
            10/8/2004                    75
4/8/2008                55                 1
5/1/2010


Will appreciate the help in this matter.

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