Analyzing zero‐truncated recurrent event data by stratified regression with time‐varying coefficients

Abstract This article presents a strategy for conducting regression analysis of zero‐truncated recurrent event data. The research is partly motivated by a pediatric mental health care (PMHC) program based on administrative data. We are particularly interested in how the occurrence of an event depends on its past occurrences and the associated covariates over time. We propose a stratified Cox regression model with time‐varying coefficients. An easy‐to‐implement procedure is provided for estimating the model parameters using the zero‐truncated data integrated with readily available population census information. We examine the proposed estimator's finite‐sample performance through simulation and establish its asymptotic properties. The PMHC program data are used throughout the article to motivate and illustrate the proposed approach.