Regression Analysis of Multivariate Recurrent Event Data with a Time-Varying Dependence Structure

• Activity Number: 53 - New Developments in Survival Analysis
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 3:30 PM to 5:20 PM
• Sponsor: Biometrics Section
• Author(s): Wen Li* and Mohammad Hossein Rahbar and Sean I Savitz and Liang Zhu and Resmi Gupta and Jing Jing Zhang and Sori Kim and Amirali Tahanan and Jing Ning
• Companies: University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and University of Texas Health Science Center at Houston and The University of Texas MD Anderson Cancer Center
• Keywords: Multivariate analysis; Random effects; Recurrent events; Survival analysis; Time-varying dependence; Stroke

Abstract

Multivariate recurrent event data occur frequently in longitudinal studies, in which each patient may repeatedly experience more than one type of event. The statistical analysis of such data gets further complicated by the time-varying dependence structure among different types of recurrent events. The available literature regarding the joint modelling of multivariate recurrent event assumes a constant dependency over time, which is often violated in real life applications. To close the knowledge gap, we proposed a class of flexible shared random effect models that allow for time-varying dependence to adequately capture time-dependent correlation and model multivariate recurrent event data. We developed an expectation maximization algorithm for the model fitting and inference. Extensive simulation studies demonstrate that the estimators of the proposed approach have good finite sample performance in real life settings. We illustrate the approach using data from stroke patients identified from UT Houston Stroke Registry. The effects of risk factors on the incidences of different types of post-stroke readmission events and the associations between these events are examined.