Dear friends,I would greatly appreciate your help with this task. I understand that for time-varying covariates in the cox regression, SPSS offers the so-called "segmented time dependent covariates", where the covariates are not systematically related to time. The SPSS advanced models manual provides an example of that. I also saw some other examples elsewhere, but found that the number time periods in all of these examples are small, and it is not difficult to do it in such cases with the Boolean logic syntax as shown in the manual, which shows an example of a four-week study, where covariates are updated each week.However, what if I have a very large number of observation periods? In my current research, the covariates are updated in each month of the year for as long as 40 years, that means the number of time periods is as large as 12*40=480 months. How can I set up the syntax to include these covariates in the cox regression? An example from my data: The time to event/death for a particular respondent W is 400 months, and a covariate X (dichotomous 0 or 1) was recorded as 1 in the 200th month. Of course, for the respondents in my data, X could be updated as 1 at any time point (measured in month) up to the time of death or cencoring (otherwise X=0). I hope to include X as an explanatory variable in the Cox regression.How can I write the syntax for this?Thanks a bunch.Jun
_________________________________________________________________
Type your favorite song. Get a customized station. Try MSN Radio powered by Pandora.
http://radio.msn.com
Locked
