Why is the linear mixed effect model used?
The linear mixed-effects model has been used for subgroup analysis to describe treatment differences among subgroups with great flexibility. The hierarchical Bayes approach has been applied to linear mixed-effects model to derive the posterior distributions of overall and subgroup treatment effects. In this article, we discuss the prior selection for variance components in hierarchical Bayes, estimation and decision making of the overall treatment effect, as well as consistency assessment of the treatment effects across the subgroups based on the posterior predictive p-value. Decision procedures are suggested using either the posterior probability or the Bayes factor. These decision procedures and their properties are illustrated using a simulated example with normally distributed response and repeated measurements.
What are the challenges of designing a phase 3 trial?
There are challenges in designing pediatric trials arising from special ethical issues and the relatively small accessible patient population. The application of conventional phase 3 trial designs to pediatrics is not realistic in some therapeutic areas. To address this issue we propose various approaches for designing pediatric trials that incorporate data available from adult studies using James-Stein shrinkage estimation, empirical shrinkage estimation and Bayesian methods. We also apply the concept of consistency used in multi-regional trials to pediatric trials. The performance of these methods is assessed through representative scenarios and an example using actual Type 2 diabetes mellitus (T2DM) trials.
What is MRCT in clinical trials?
The application of multi-regional clinical trials (MRCTs) is a preferred strategy for rapid global new drug development. In a MRCT, besides the other subgroup factors that are in general well defined, region is a special subgroup factor which can be a surrogate of many intrinsic and extrinsic factors. The definition of a region for a MRCT may be trial specific. It depends on where the MRCT will be conducted and how the sample sizes will be allocated across the regions. As a regional health authority will carefully review the regional treatment effect before the approval of the drug for the patients of the region, special attention should be paid to the regional subgroup analysis. In this chapter, we will discuss subgroup analysis in design and analysis of multi-regional clinical trials focusing on regional subgroup analysis. These include the considerations on region definition, analysis model, consistency assessment of regional treatment effects, regional sample size allocation and trial result interpretation. Numerical and real trial examples will be used to illustrate the applications of the methods.
What is the primary objective of a multiregional clinical trial?
The primary objective of a multiregional clinical trial (MRCT) is to assess the efficacy of all participating regions and evaluate the probability of applying the overall results to a specific region. The consistency assessment of the target region with the overall results is the most common way of evaluating the efficacy in a specific region. Recently, Huang et al (2012) proposed an additional trial in the target region to an MRCT to evaluate the efficacy in the target ethnic (TE) population under the framework of simultaneous global drug development program (SGDDP). However, the operating characteristics of this statistical framework were not well considered. Therefore, a nested group sequential program for regional efficacy evaluation is proposed in this paper. It is an extension of Huang’s SGDDP framework and allows interim analysis after MRCT and in the course of local clinical trial (LCT) phase. It is able to well control the familywise type I error in the program level and enhances the flexibility of the program. In LCT sample size estimation, we introduce virtual trial which is transformed from the original program by using discounting factor and an iteration method is employed to calculate the sample size and stopping boundaries of interim analyses. The proposed sample size estimation method is validated in the simulations and the effect of varied weight, effect size of TE population and design setting is explored. Examples with normal endpoint, binary endpoint and survival endpoint are shown to illustrate the application of the proposed nested group sequential program.
How are multi-regional clinical trials structured?
Data observed in multi-regional clinical trials are structurally hierarchical in the sense that the patient population consists of several regions and patients are nested within their own regions. In order to reflect such hierarchical structure, in this paper, we propose two-level hierarchical linear models in which the level-1 model is based on patient-level data such as treatment indicator and age, and the level-2 model is based on region-level data such as medical practices. The fixed effect model and the continuous random effect model are shown to be special cases of hierarchical linear models. We conducted simulation studies to investigate the empirical type I error rates of three methods for testing the overall treatment effect. The performance of the testing method with sample ratios as weights and the empirical Bayes estimator for between-region variability is better than that of the other two testing methods.
What is confirmatory evidence in ICH E9?
The one of the principles described in ICH E9 is that only results obtained from pre-specified statistical methods in a protocol are regarded as confirmatory evidence. However, in multi-regional clinical trials, even when results obtained from pre-specified statistical methods in protocol are significant, it does not guarantee that the test treatment is approved by regional regulatory agencies. In other words, there is no so-called global approval, and each regional regulatory agency makes its own decision in the face of the same set of data from a multi-regional clinical trial. Under this situation, there are two natural methods a regional regulatory agency can use to estimate the treatment effect in a particular region. The first method is to use the overall treatment estimate, which is to extrapolate the overall result to the region of interest. The second method is to use regional treatment estimate. If the treatment effect is completely identical across all regions, it is obvious that the overall treatment estimator is more efficient than the regional treatment estimator. However, it is not possible to confirm statistically that the treatment effect is completely identical in all regions. Furthermore, some magnitude of regional differences within the range of clinical relevance may naturally exist for various reasons due to, for instance, intrinsic and extrinsic factors. Nevertheless, if the magnitude of regional differences is relatively small, a conventional method to estimate the treatment effect in the region of interest is to extrapolate the overall result to that region. The purpose of this paper is to investigate the effects produced by this type of extrapolation via estimations, followed by hypothesis testing of the treatment effect in the region of interest. This paper is written from the viewpoint of regional regulatory agencies. © 2018 The Korean Statistical Society, and Korean International Statistical Society.
Summary
We consider two recent suggestions for how to perform an empirically motivated Monte Carlo study to help select a treatment effect estimator under unconfoundedness. We show theoretically that neither is likely to be informative except under restrictive conditions that are unlikely to be satisfied in many contexts.
1 INTRODUCTION
A large literature focuses on estimating average treatment effects under unconfoundedness (see, e.g., Blundell & Costa Dias, 2009; Imbens & Wooldridge, 2009; Abadie & Cattaneo, 2018 ).
2 EMCS DESIGNS
We first describe the two main approaches to conducting an EMCS, namely the placebo design of Huber et al. ( 2013) and the structured design of Busso et al. ( 2014 ). In either EMCS design, one simulates many “empirical Monte Carlo” replication samples from a known DGP.
3 THEORY
To understand the conditions under which an EMCS might be informative about the preferred estimator in some particular data set, we first construct a simple example. Here we have only two estimators, with a straightforward and restricted joint sampling distribution (bivariate Gaussian).
4 APPLICATION
To demonstrate the empirical relevance of the theoretical results discussed above, and to consider the extent to which they might be a problem in practice, we provide an application of EMCS procedures to a real‐world data set. In these data we have an experimental estimate of the treatment effect.
5 RESULTS
We now describe the results of our tests of the two EMCS procedures—placebo and structured—in the context of our real‐world data. As described in Section 4.2, we perform three sets of tests. First, we apply the two procedures to the NSW treatment sample, combined with the CPS‐1 comparison data set.
6 DISCUSSION
Advances in econometrics have left the empirical researcher blessed with a wealth of possible treatment effect estimators from which to choose. They have not yet provided clear guidance on which of these estimators should be preferred in which context.
Which is the most commonly used meta-analytic estimator?
The random effects estimator is arguably the most commonly used meta‐analytic estimator. It does not explicitly correct for publication selection other than giving greater weight to more precise estimates of βi. It estimates the population mean effect μ assuming the following specification:
Do simulation studies have bias?
Different simulation studies have implemented bias differently, have drawn sample sizes from different distributions, and have varied widely in the value and form of the simulated true underlying effects. This lack of overlap is not surprising given that there is an effectively infinite number of possible combinations of different conditions to explore and no way of determining which conditions actually underlie real‐world data. In other words, not only is there an inherent dimensionality problem in these simulation studies, but there is also no ground truth. These problems are often not discussed in reports of simulation studies, and indeed, many of the reports just cited—explicitly or implicitly—recommended the use of a single method, despite the fact that each study examined performance of only a handful of correction methods in only a limited subset of possible conditions.
Abstract
Randomized controlled trials (RCTs) are universally recognized as the preferred way to infer treatment effects because RCTs typically minimize validity challenges while maximizing estimation efficiency.
Background and Introduction
Randomized controlled trials (RCTs) are universally recognized as the preferred way to infer treatment effects because RCTs typically minimize validity challenges while maximizing estimation efficiency. In practice, simple RCTs—where individual study subjects are assigned to treatment or control status at random—are often infeasible.
An intervention to reduce falls in nursing homes
This paper uses an ongoing evaluation study as a running example; some nonpertinent details of the intervention being evaluated are omitted or simplified for the sake of brevity.
The CR and PCR designs
Cluster units (in our case, nursing home residents) are often intuitively apparent to the evaluator. Residents cluster within nursing homes, and the intervention is introduced at the nursing home level. Similar clustering examples are common in the evaluation literature—students cluster within schools, residents within a geographic region, etc.
The Monte Carlo simulation and design decisions
In this section, we illustrate use of preintervention data and Monte Carlo simulation to make design decisions for evaluating a program to reduce falls across nursing homes. The principal decisions regard choice of basic design (CR or PCR); choice of estimators (nonparametric or parametric); and whether to include covariates.
Conclusions
We have shown that, based on our Monte Carlo simulation, the best strategy for the nursing home evaluation is to use a PCR design and a nonparametric estimator. The resulting power is adequate for detecting a multiplicative treatment effect of about 0.25 (i.e., a 25 % reduction in falls) and perhaps smaller.
Acknowledgments
This study was funded by the Agency for Healthcare Research & Quality (AHRQ), Department of Health & Human Services (DHHS), under contract # HHSA290201000031I.
Authors
Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.
1 Introduction
Identifying subgroups of patients with an enhanced response to a new treatment has become an area of increased interest in the last few years. Hence subgroup analyses are commonly performed all throughout the clinical development process.
2 Multiple contrast tests for multiple populations
We consider a clinical trial, which has been performed in parallel groups of patients, which receive different doses d1,d2,...,dk of an experimental treatment. The number of patients in the dose groups are n1,...,nk and n:= k∑i=1ni, the total number of patients in the study.
3 Simulation study
In this section we discuss results of a simulation study to evaluate the properties of the multi-population testing approaches. The simulation setup is similar to the one used in [15] . The simulations are divided into two main parts.
4 Discussion
In this paper we discussed an extension of the MCP part of the MCP-Mod methodology to allow for testing in multiple populations. We focused on the situation of a dose-response trial, where a subgroup with a suspected enhanced treatment effect has been prespecified.
Acknowledgements
This work was supported by funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 633567 and by the Swiss State Secretariat for Education, Research and Innovation (SERI) under contract number 999754557 .
2 Additional simulation results for heteroscedastic scenarios
Figure 10: Probability to reject the global null hypothesis for single population (SP) and multi-population (MP) testing methods. Data are generated from a linear model under heteroscedasticity. MP-MultDF is used to approximate the joint distribution for MP testing methods.