Online Program Home
My Program

Abstract Details

Activity Number: 187 - Contributed Poster Presentations: Section on Nonparametric Statistics
Type: Contributed
Date/Time: Monday, July 29, 2019 : 10:30 AM to 12:20 PM
Sponsor: Section on Nonparametric Statistics
Abstract #306726
Title: Application of a Nonparametric Test for Comparing Transition Probabilities in Multi-State Models
Author(s): Ying Zhang* and Jun Park and Hong Wan and Valerie Teal and Robert Tipping and Giorgos Bakoyannis
Companies: Merck and Indiana University and Merck and Merck and Merck and Indiana University
Keywords: Transition Probabilities; Nonparametric Test; Multi-state Models

Multi-state models (MSMs) can be applied in helping to understand the drug effect on disease progressions in clinical trials. Many methods focus on modeling and comparing transition intensities between treatment arms in MSM settings, but they do not directly compare transition probabilities, which is often the parameter of interest in drug development programs. Since transition intensities and transition probabilities do not have 1-to-1 correspondence, the results of comparison transition intensities cannot be directly interpreted as comparisons between transition probabilities. Therefore, having a nonparametric test to compare transition probabilities on their own scale is critical. We considered the nonparametric test for comparing transition probabilities in non-homogeneous Markov processes proposed by Bakoyannis. The proposed test statistic is asymptotically normal. Bootstrapping method is used to estimate the variances of the test statistics. Monte Carlo simulations were performed to evaluate the performance of the method. Data from a Letermovir trial were analyzed to compare transition probabilities to cytomegalovirus infection and death after HSCT between treatment arms.

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

Back to the full JSM 2019 program