Re: st: survival analysis in the presence of competing risks and multi-level data

[Phil Clayton <philclayton@internode.on.net>]

A recent article in Statistics in Medicine dealt with this issue by extending shared frailty models to the competing risks situation:
Katsahian and Boudreau. Estimating and testing for center effects in competing risks. Stat Med (2011). doi: 10.1002/sim.4132

It's pretty hot off the press and would require a lot of work to implement. I am not sure if you can use a model based on Cox regression, eg Lunn & McNeil's method, in clustered data - but if you can it would almost certainly be a lot easier than implementing the above.
Lunn and McNeil. Applying Cox regression to competing risks. Biometrics (1995) vol. 51 (2) pp. 524-32

Phil

On 13/03/2011, at 4:52 AM, Salah Mahmud wrote:

> Hello,
> Just wondering if anyone is aware of reasonable approaches (preferably
> implementable in Stata) to fit competing risks time-to-event models to
> clustered data.
> 
> I have a dataset where participants are clustered by family and
> eventually by locality. The outcome is time to first hospitalization
> due to a certain condition, but observing this outcome could be
> precluded by death. So I would like to account for these competing
> risks  because my predictors will likely influence all outcomes.
> Normally I would use something like Fine & Gray models, but the data
> are clustered at more than one level and I would like to model the
> effects of covariates measured at different levels (and not just
> adjust my SE for correlation, eg, using robust SE estimates). Is there
> a way of fitting this model in Stata (eg, using gllamm)? in a
> different package?
> 
> Thanks,
> *
> *   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/