In this talk, we present recent works on extensions of the hierarchical (or h-) likelihood (Lee and Nelder, 1996) to time-to-event data (survival data), which overcomes various challenges due to incomplete observations caused by censoring, truncation and competing events, and also present further extensions of existing works, such as complicated structured frailties (or random effects) and joint models for repeated measures and time-to-events. In particular, we show that the penalized h-likelihood approach gives a useful methodology for interval estimation of the individual frailty and variable selection of covariates including high-dimensional cases in survival models (e.g. proportional hazards and AFT models) with random effects. We also demonstrate via the h-likelihood how to make inference various random-effect survival models using time-to-event data from multi-center clinical trials, and the inferential results are also compared with those of copula survival models which are useful for modelling clustered survival data. Furthermore, we discuss about a non-parametric mixture approach in the h-likelihood framework when baseline hazard and frailty distribution are unknown.