Re: st: case-cohort data structures

[Enzo Coviello <coviello@mythnet.it>]

At 12.24 05/07/2002 -0400, you wrote:
> Hi, we've been using the stcascoh and stselpre commands to turn a cohort
> study dataset into a dataset designed for case-cohort analysis. Wonderful
> commands these.
> However, we don't want Stata to select the random sample from the cohort at
> baseline, we've already done that.  Rather, we wish to feed the stcascoh
> command the subcohort members and nonsubcohort cases, or an indicator
> variable stating which cohort members should be in the subcohort.
> Is there an extension to the stcascoh command that this that would force the
> macro to keep our selection,
Unfortunately not.


> or have we missed finding a different command
> to accomplish this?

You could try in this way:
1) Divide your dataset in two files: the first one for subcohort members,
   the second one for non subcohort cases.
2) appropriately stset new files.
3) feed the stcascoh command the subcohort file setting alpha just below 1 
(i.e. .999)
    and save it. Do some assert to verify that _subco _wSelpre _wBarlow 
are coded as you expect.
4) feed the stcascoh command  the non-subcohort case file setting alpha = 0
    and save it. Do some assert to verify that _subco _wSelpre _wBarlow 
are coded as you expect.
 5) append two files.

I am not sure that this works (be careful), but it is the best I can do now.
Ciao

Enzo







Enzo Coviello
Unit� di Epidemiologia ASL Ba/1
via Fornaci 201,  70031 ANDRIA (Ba) - Italy
tel +39 0883 299744-8 fax +39 0883 299417
tel home +39 0883 695055


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/