Control of Frequentist type I error rates in hierarchical linear models for multiregional clinical trials using a Bayesian method

Abstract

A multiregional clinical trial is a clinical trial conducted simultaneously in multiple regions under a single protocol. It is a new drug development strategy that has advantages in many ways. A significant part of these advantages comes from the hierarchical data structure of the multiregional clinical trials. Based on the recent work of Kim and Kang (2020) concerning hierarchical linear models to better reflect the structure, we address the problem of controlling the Frequentist type I error rate when the number of regions is small using a Bayesian method. More precisely, we present a strategy for controlling the type I error rate by setting a prespecified Bayesian credible level. Simulation results show that our proposed method controls the type I error rate reliably and well. Through sensitivity analysis, we also investigate the effect on the results when accurate information on between-region variability is given and when several different types of prior distributions are used.

Keywords:

• Bayesian credible levels
• Generalized linear mixed models
• Multiregional strategy

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Research Foundation(NRF), Korea, under project [BK21 FOUR (2021-11-0077)]. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (RS-2023-00239534).