ABSTRACT
Sometimes survival data come in clusters, and multivariate, or frailty, models are appropriate to use. Thedistinction between individual and population mortality is described, and the individual and shared frailtymodels are introduced, the former being of less importance due to identification problems. It is shown howto use the available resources in R to analyze the shared frailty model in a proportional hazards setting. Analternative to the shared frailty model is stratification (in certain circles known as “fixed effects models”, amisnomer), and a comparison is made between the two.
