|
Abstract:
|
The process by which an individual experiences a series of recurrent events, such as tumor recurrence, is frequently subject to death. In a full cohort, a joint frailty model for the hazards of recurrent events and death may be used to investigate covariate effects on the two event types while accounting for their dependence. However, when certain covariates of interest are difficult to obtain, researchers may need to sub-sample patients on whom to collect complete data. One possibility is the nested case-control (NCC) design, in which risk set sampling is performed based on a single outcome of interest, possibly matching on certain factors. We propose a novel framework for estimation and inference for a joint frailty model for recurrence and death using data from an NCC study, under multiple schemes for risk set formation. We propose both a parametric maximum weighted likelihood approach and a maximum weighted penalized likelihood approach using flexible parametric models for the baseline hazards, assessing their operating characteristics via simulation. We illustrate our methods with data from a study on breast cancer investigating locoregional recurrence and death.
|
