Implementation of clustering in CoxPH

[jchts01]

I used the clustering function with CoxPH fitter while working on a set of medical data (on readmission) to account for the same patients having more than one episode, but I am uncertain as to how clustering was handled in lifelines. My question is: does applying 'cluster_col' invoke only the use of sandwich variance estimates? Or does it employ a shared frailty model as well? Thanks.

[CamDavidsonPilon]

Hi @jchts01 - it only invokes sandwich variance estimates

[jchts01]

Understood, thanks!