Hello,

I dont' think till now such packages are avialible in R; furthermore, the fitting model specification is a rare one. For a simple weighted Cox ph, you can deal with coxphw package. What about Random Forests for Survival data !!

Hadis Mozaffari - Thank you for your question. The enclosed video link may answers your question.

https://www.youtube.com/watch?v=PgW-IU4EX-M

The Cox proportional hazards model [36] has become by a wide margin the most used procedure for modeling the relationship of covariates to a survival or other censored outcome.

Hello,

I am not very clear about what you mean by multilevel modelling. If you are concerned on variables that may behave different depending on factors such as locations, sex or other conditions you could use strata() with coxphw, as suggested by Chellai Fatih , since it provides " estimation of average hazard ratios (Schemper et al., 2009) using Prentice weights with censoring correction and robust variance estimation; " with the default template "AHR".

http://finzi.psych.upenn.edu/library/coxphw/html/coxphw.html.

But If you are refering to a study where the selection of sub-cohorts is stratified according to some criteria I recomend you to have a look at cchs package:

https://journal.r-project.org/archive/2018/RJ-2018-012/RJ-2018-012.pdf

The package also comes with a dataset cchsData where the subcohort selection is stratified based on another variable, in this case, the result of an histological examination.

If neither of these approaches satisfies you, you could subset the data in the diferent levels and do a regression for each one sepparately and then a meta-analisys to get a overall HR but I would not recomend it.