Online Program Home
My Program

Abstract Details

Activity Number: 466 - First-Hitting-Time Based Threshold Regression and Applications
Type: Topic Contributed
Date/Time: Wednesday, August 1, 2018 : 8:30 AM to 10:20 AM
Sponsor: Section on Risk Analysis
Abstract #328874 Presentation
Title: Distribution-Free Inference Methods for Threshold Regression
Author(s): Mei-Ling Ting Lee* and George A Whitmore
Companies: University of Maryland and McGill University
Keywords: Disease progression; Failure time; First hitting time; Predictive inference; Time-to-event data; independent increments

In many medical and health-care contexts, a failure event (such as death, hospitalization or transplant) is triggered when a subject's deteriorating health first reaches a failure threshold. The failure process is well described as the sample path of a stochastic process hitting a boundary. The parameters and behaviors of such failure processes must often be inferred from data sets that include censored survival times and current health levels of survivors. A substantial input of expert experience with the health context is usually required to guide the data modeling. This talk describes a parsimonious model for the failure process that has only one distributional property, namely, stationary independent increments. As this property is frequently encountered in real applications, the stochastic model and its related statistical methodology have potential for general application in many fields. The mathematical underpinnings of the distribution-free methods for estimation and prediction will be described in the talk as well as techniques for incorporating covariates. Case examples will be presented.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program