Subgroup Analyses in Randomized Clinical Trials: Statistical and Regulatory Issues

Grouin, Jean‐Marie; Coste, Maylis; Lewis, John A.

doi:10.1081/bip-200067988

Cited by 47 publications

(32 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although there are arguments for why assessments of heterogeneity do not need adjustments for multiplicity (by reference to a publication by Grouin et al [13]) or could use criteria such as P < 0.1 or even P < 0.2 as signals for additional scrutiny in view of low power (Jackson [14]), we believe that the discussions of these authors have been misunderstood. In this regard, Grouin et al [13] wrote that a subgroup hypothesis is secondary to the overall one and that there is an implicit hierarchy for them. Jackson [14] wrote about the adjustments made for testing the feasibility of meta-analyses.…”

Section: Firstmentioning

confidence: 96%

Inconsistent trial assessments by the National Institute for Health and Clinical Excellence and IQWiG: standards for the performance and interpretation of subgroup analyses are needed

Hasford

Bramlage

Koch

et al. 2010

Journal of Clinical Epidemiology

View full text Add to dashboard Cite

Section: Firstmentioning

confidence: 96%

Inconsistent trial assessments by the National Institute for Health and Clinical Excellence and IQWiG: standards for the performance and interpretation of subgroup analyses are needed

Hasford

Bramlage

Koch

et al. 2010

Journal of Clinical Epidemiology

View full text Add to dashboard Cite

“…Subgroup analyses are often presented in study reports in the clinical literature. The most frequently reported reason for subgroup analysis is to show that treatment has a ‘statistically significant’ effect in one or more subgroups, although the trial shows no apparent overall effect 6–8. These subgroups are usually not predefined before the trial.…”

Section: Discussionmentioning

confidence: 99%

Targeting population entering phase III trials: A new stratified adaptive phase II design

Tournoux-Facon

Rycke

Tubert‐Bitter

2011

Statistics in Medicine

View full text Add to dashboard Cite

The primary goal of phase II studies is to assess the efficacy of the new treatment in order to decide whether it has sufficient activity to warrant further evaluation in a phase III comparative trial. However, many adequately conducted phase II trials are negative leading to termination of drug development. Heterogeneity of the population is often considered to be a cause of treatment effect dilution. One approach to determine the sensitive subpopulation is to conduct several phase II trials, one in each specific subset of patients. This option might unethically increase the number of non-sensitive patients under evaluation. Adaptive two-stage designs have been recently proposed. London and Chang proposed a global one-sample test for response rates for stratified phase II clinical trials, whereas Jones and Holmgren proposed an adaptive design that allows preliminary determination of efficacy that may be restricted to a specific subpopulation defined by biomarker status. These two methods do not allow early termination for efficacy in one or several subgroups as they are extensions of the Simon design. The authors propose an alternative method to deal with stratification in phase II clinical trials and identification of the best target population. This method is based on the multiple-stage Fleming design allowing for early stopping rules for either efficacy or inefficacy. It also integrates a procedure testing whether treatment effects are similar or heterogeneous between the two groups. The operating characteristics of this method were compared with those of a standard Fleming design using exact binomial probabilities.

show abstract

“…The degree of borrowing across strata was small, because the prior for , obtained after elicitation from experts, was concentrated in a range representing large to very large between-strata heterogeneity: 2 Gamma.2, 20/, with prior 95% interval (1.89, 9.09) for . Despite the fact that the prior excluded the possibility of substantial pooling, the posterior probabilities that CBR rates exceed the clinically relevant threshold 30% were small for all strata (between 0 and 0.134, table 4 in [16]), which was sufficient to conclude that 'imatinib is not an active agent in advanced sarcoma in these subtypes' [16].…”

Section: Hierarchical Modelsmentioning

confidence: 99%

Robust exchangeability designs for early phase clinical trials with multiple strata

Neuenschwander

Wandel

Roychoudhury

et al. 2015

Pharmaceutical Statistics

140

157

View full text Add to dashboard Cite

Clinical trials with multiple strata are increasingly used in drug development. They may sometimes be the only option to study a new treatment, for example in small populations and rare diseases. In early phase trials, where data are often sparse, good statistical inference and subsequent decision-making can be challenging. Inferences from simple pooling or stratification are known to be inferior to hierarchical modeling methods, which build on exchangeable strata parameters and allow borrowing information across strata. However, the standard exchangeability (EX) assumption bears the risk of too much shrinkage and excessive borrowing for extreme strata. We propose the exchangeability-nonexchangeability (EXNEX) approach as a robust mixture extension of the standard EX approach. It allows each stratum-specific parameter to be exchangeable with other similar strata parameters or nonexchangeable with any of them. While EXNEX computations can be performed easily with standard Bayesian software, model specifications and prior distributions are more demanding and require a good understanding of the context. Two case studies from phases I and II (with three and four strata) show promising results for EXNEX. Data scenarios reveal tempered degrees of borrowing for extreme strata, and frequentist operating characteristics perform well for estimation (bias, mean-squared error) and testing (less type-I error inflation).

show abstract

Subgroup Analyses in Randomized Clinical Trials: Statistical and Regulatory Issues

Cited by 47 publications

References 19 publications

Inconsistent trial assessments by the National Institute for Health and Clinical Excellence and IQWiG: standards for the performance and interpretation of subgroup analyses are needed

Inconsistent trial assessments by the National Institute for Health and Clinical Excellence and IQWiG: standards for the performance and interpretation of subgroup analyses are needed

Targeting population entering phase III trials: A new stratified adaptive phase II design

Robust exchangeability designs for early phase clinical trials with multiple strata

Contact Info

Product

Resources

About