I'm tryng to analyze data in which the response is time to risk events (e.g., injection drug use, unprotected sex) following release from prison. During the follow-up period, some participants were in and out of the correctional facility multiple times. We are specifically interested in modeling time from release to risk event. I've either not structured the data correctly, or I'm not specifying the stset command quickly. Here's what the data would currently look like:
ID PERIOD TIME FAIL X1 X2 . . .
1 1 15 1
2 1 90 0
3 1 5 1
3 2 9 1
3 3 35 0
Subject subjects 1 and 2 each had only one post-incarceration period in which they were followed up. Subject 1 failed (used drugs) on day 15, subject 2 was observed for 90 days and did not fail.
Subject 3 was observed during three post incarceration follow-up periods. During the first period, this subject failed 5 days following release, this subject was re-incarcerated and failed on day 9 following the second incarceration. Subject 3 was incarcerated a third time, after which he was observed for 35 days without failure.
We don't want to treat this as one continuous time period. We want the "at risk" period to start immediatedly after each release.
Thanks,
Brad
*
* 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/
