Correlated Frailty Models
In the last chapter we focused on the shared frailty model as a way of modeling associations between event times. The concept of shared frailty goes back to Clayton (1978), and was extensively studied in books by Hougaard (2000), and Duchateau and Janssen (2008). The shared frailty approach has proven to be a useful and popular extension of the Cox model when observations from subjects are not statistically independent of each other. As discussed in Section 4.9, the shared frailty model is especially useful for clustered event time data when the correlation structure in the clusters is symmetric and not of special interest as, for example, in multicenter clinical trials.