Analysis of interval‐grouped recurrent‐event data using piecewise constant rate functions

Abstract

AbstractWe consider situations where subjects in a longitudinal study experience recurrent events. However, the events are observed only in the form of counts for intervals which can vary across subjects. Methods for estimating the mean and rate functions of the recurrent‐event processes are presented, based on loglinear regression models which incorporate piecewise‐constant baseline rate functions. Robust methods and methods based on mixed Poisson processes are compared in a simulation study and in an example involving superficial bladder tumours in humans. Both approaches provide a simple and effective way to deal with interval‐grouped data.