A Bayesian approach based on discounting factor for consistency assessment in multi-regional clinical trial.

Multi-regional clinical trial (MRCT) has become an increasing trend for its supporting simultaneous global drug development. After MRCT, consistency assessment needs to be conducted to evaluate regional efficacy. The weighted Z-test approach is a common consistency assessment approach in which the weighting parameter W does not have a good practical significance; the discounting factor approach improved from the weighted Z-test approach by converting the estimation of W in original weighted Z-test approach to the estimation of discounting factor D. However, the discounting factor approach is an approach of frequency statistics, in which D was fixed as a certain value; the variation of D was not considered, which may lead to un-reasonable results. In this paper, we proposed a Bayesian approach based on D to evaluate the treatment effect for the target region in MRCT, in which the variation of D was considered. Specifically, we first took D random instead of fixed as a certain value and specified a beta distribution for it. According to the results of simulation, we further adjusted the Bayesian approach. The application of the proposed approach was illustrated by Markov Chain Monte Carlo simulation.

Regional efficacy evaluation in multi-regional clinical trials using a discounting factor weighted Z test.

Multi-regional clinical trial (MRCT) is an efficient design to accelerate drug approval globally. Once the global efficacy of test drug is demonstrated, each local regulatory agency is required to prove effectiveness of test drug in their own population. Meanwhile, the ICH E5/E17 guideline recommends using data from other regions to help evaluate regional drug efficacy. However, one of the most challenges is how to manage to bridge data among multiple regions in an MRCT since various intrinsic and extrinsic factors exist among the participating regions. Furthermore, it is critical for a local agency to determine the proportion of information borrowing from other regions given the ethnic differences between target region and non-target regions. To address these issues, we propose a discounting factor weighted Z statistic to adaptively borrow information from non-target regions. In this weighted Z statistic, the weight is derived from a discounting factor in which the discounting factor denotes the proportion of information borrowing from non-target regions. We consider three ways to construct discounting factors based on the degree of congruency between target and non-target regions either using control group data, or treatment group data, or all data. We use the calibrated power prior to construct discounting factor based on scaled Kolmogorov-Smirnov statistic. Comprehensive simulation studies show that our method has desirable operating characteristics. Two examples are used to illustrate the applications of our proposed approach.

Statistical Properties of the Design for Simultaneous Global Drug Development Program.

Recently, a design was proposed for the Simultaneous Global Drug Development Program (SGDDP) to assess the impact of ethnic factors on the effect of a new treatment for a targeted ethnic (TE) population. It used weighted Z tests to combine the information collected from the TE and non-TE (NTE) subgroups in the SGDDP based on the fundamental assumption on their shared biological commonality. In this article, we mathematically formulated this assumption as the quantitative interaction between treatment effect and subgroup. We used it to more rigorously describe the hypotheses, and showed the unbiasedness of the weighted Z test. Moreover, to study the loss of efficiency from down weighting the NTE information in this SGDDP design, we compared the power of their test with that of the uniformly most powerful (UMP) test, which we showed was also a weighted Z test. We discussed that the choice of weight should balance the maximization of power when the assumption holds and the minimization of bias otherwise.

A Quantitative Method for Weight Selection in SGDDP.

Ethnic factors pose major challenge to evaluating the treatment effect of a new drug in a targeted ethnic (TE) population in emerging regions based on the results from a multiregional clinical trial (MRCT). To address this issue with statistical rigor, Huang et al. (2012) proposed a new design of a simultaneous global drug development program (SGDDP) which used weighted Z tests to combine the information collected from the nontargeted ethnic (NTE) group in the MRCT with that from the TE group in both the MRCT and a simultaneously designed local clinical trial (LCT). An important and open question in the SGDDP design was how to downweight the information collected from the NTE population to reflect the potential impact of ethnic factors and ensure that the effect size for TE patients is clinically meaningful. In this paper, we will relate the weight selection for the SGDDP to Method 1 proposed in the Japanese regulatory guidance published by the Ministry of Health, Labour and Welfare (MHLW) in 2007. Method 1 is only applicable when true effect sizes are assumed to be equal for both TE and NTE groups. We modified the Method 1 formula for more general scenarios, and use it to develop a quantitative method of weight selection for the design of the SGDDP which, at the same time, also provides sufficient power to descriptively check the consistency of the effect size for TE patients to a clinically meaningful magnitude.

A simple and robust method for partially matched samples using the p-values pooling approach.

This paper focuses on statistical analyses in scenarios where some samples from the matched pairs design are missing, resulting in partially matched samples. Motivated by the idea of meta-analysis, we recast the partially matched samples as coming from two experimental designs and propose a simple yet robust approach based on the weighted Z-test to integrate the p-values computed from these two designs. We show that the proposed approach achieves better operating characteristics in simulations and a case study, compared with existing methods for partially matched samples.