Title: Quantile Regression for Bivariate Sequential Time-to-Event Data Analysis 

Abstract: Bivariate sequential time-to-event data arise when individuals may experience two events in sequence, with the second gap time observed only after the occurrence of the first event. Quantile regression provides a useful framework for investigating heterogeneous covariate effects across the survival-time distribution. However, directly applying of existing methods developed for univariate censored survival data may be inappropriate when the sequential gap times are dependent. This thesis develops a two-stage estimation procedure for quantile regression of the marginal distribution of the second gap time while accounting for dependence between sequential event times. A grid-based optimization procedure is employed to solve martingale-based estimating equations and obtain parameter estimates for quantile regression model, while uncertainty is evaluated using a nonparametric bootstrap procedure. The proposed methodology is further illustrated using a colon cancer dataset to investigate covariate effects on the time from cancer recurrence to death.

---

Location: HH-3017

Date and Time: Friday, Sept. 4 at 01:00 PM - 02:00 PM (NDT)

---