Abstract:
|
Panel count data occurred when events are observed at finite fixed time points and the visit times vary from subject to subject, and the exact event times are unknown. In the past decades, there have been extensive researches to study the proportional mean model for panel count data, if we consider only time independent coefficients and covariates. When we account for time dependent coefficient, spline method is an important method, generally used to study for survival data. Nevertheless, limited work has been done in panel count data. Furthermore, in practice, the time-dependent covariate effects situation are common. Limited research has been found here. In this paper, we consider situations where coefficient and covariate effects are time-dependent simultaneously. Based on the conditional estimating equations method developed for time-dependent covariates, we approximate the coefficients by Bernstein splines, hence allow both coefficients and covariates to be time-dependent. We conduct a comprehensive simulation study to prove the operational characteristics of the proposed method. A real data example will be analyzed for illustration.
|