I want to consider death censoring in Multilevel survival models for length of hospital stay analysis

[Mark Hh]

Hi,

I have a question using the commands referred in the following web site.

https://www.stata.com/stata-news/new...rvival-models/

stset los
mestreg lap_surg age || hospital: || doctor:, distribution(weibull)

If we have data with patients who died before hospital discharge, is it possible to consider this death in the regression?

Or do we just need to drop patients who died before discharge?

Thank you in advance

[Mark Hh]

Or Perhaps, is there a solution using usual Cox regression (stcox) for the lenght of hospital stay analysis with considering death censoring?