Understanding frailty effect in frailty models using frailty HL R package

Frailty model, i.e. Cox proportional hazards model with frailty term, has been widely used for the analysis of univariate and clustered survival data. Here, the frailty is an unobserved random effect in the hazard function, which is useful for modelling the heterogeneity and dependence among the survival outcomes. In this paper, we review the meaning and role of frailty effect using h-likelihood. We demonstrate how to understand the frailty effect using a numerical study and two practical datasets via the frailtyHL R.package. We also show through the data analysis that the frailty effect can capture heterogeneity and/or classification inherent in survival outcomes. In particular, we provide the R codes which can be directly accessible to interested readers.