Statistics and Its Interface

Volume 12 (2019)

Number 1

Semiparametric estimation of differences in treatment-specific recurrent event means with a terminal event

Pages: 1 – 9

DOI: https://dx.doi.org/10.4310/SII.2019.v12.n1.a1

Authors

Xiaowei Sun (Institute of Applied Mathematics, Academy of Mathematics and Systems Science, CAS, Beijing, China; and University of the Chinese Academy of Sciences, Beijing, China)

Liuquan Sun (Institute of Applied Mathematics, Academy of Mathematics and Systems Science, CAS, Beijing, China; and University of the Chinese Academy of Sciences, Beijing, China)

Abstract

Recurrent event data often arise from biomedical studies and a terminal event may preclude further occurrence of recurrent events. In comparing treatments, the marginal mean is frequently of interest, and treatment-specific differences in the mean number of events are often not constant over time. In this article, we propose a semiparametric method to compare treatment-specific recurrent event means by combining an additive hazards model for the terminal event and an additive rates model for the conditional recurrent event rate. The treatment effect is measured by the difference between treatment-specific recurrent event means. Estimation procedures are developed for the measure and the asymptotic properties of the proposed estimators are established. The finite sample performance of the proposed estimators is examined through simulation studies, and an application to a bladder cancer study demonstrates the usefulness of our method.

Keywords

Keywords and phrases: additive models, marginal mean, recurrent events, semiparametric method, terminal event, treatment effect

2010 Mathematics Subject Classification

Primary 62N01, 62N02. Secondary 62G05.

Received 22 November 2017

Published 26 October 2018