Hello,
I have a dataset with multiple spells (failures) per subject. There are left-censored, right-censored, left-truncated and right-truncated spells.
I would like to run a semi parametric Cox model with a Gamma shared frailty component and a time-varying covariate. I think that the proportional hazards assumption is violated for a model with frailty and/or time-varying covariates. What I think is true? Is it a problem when the proportional hazards assumption is violated for this model?
If the proportional hazards assumption is violated and the left-censored and left-truncated spells are discarded, will I have biased results about the shape of the hazard function?
Thanks a lot
Luciana
*
* 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/
