Main page > Innovation > Flexible designs > Adaptive design working group > Adaptive design lecture series

This page is maintained by the Adaptive design working group (Alex Dmitrienko).

2009 adaptive design lecture series.

Planning and execution of learn-phase clinical trials presented by Dr. Parvin Fardipour (Wyeth) [Download slides].

A new group-sequential Phase II/III clinical trial design presented by Prof. Nigel Stallard and Tim Friede (University of Warwick) [Download slides].

Regulatory affairs and adaptive designs presented by Dr. Greg Enas (Eli Lilly) [Download slides].

An adaptive optimal design for the Emax model and its application in clinical trials presented by Dr. Sergei Leonov (GSK) [Download slides].

Strategies for including patients recruited during interim analysis of clinical trials presented by Prof. Andreas Faldum (University of Mainz, Germany) [Download slides].

Abstract
In clinical trials a periodical check of safety and efficacy data is often needed. For organizational reasons it is rarely desirable to stop a trial during such an interim analysis. Therefore, new study patients are included in the trial while the interim analysis is ongoing. Disregarding the additional information provided by these interim patients would be unsatisfactory, especially for an office of regulatory affairs. Consequently, the rules for group sequential or adaptive decisions must be adjusted to the recruitment of interim patients. In this presentation we consider strategies for modifying study designs to consider the analysis of interim patients.

A case study for an adaptive Phase II/III trial presented by Dr. Jeff Maca (Novartis) [Download slides].

An adaptive drug/biomarker design for the new millennium: I-SPY2 presented by Prof. Don Berry (MD Anderson Cancer Center) [Download slides].

Testing and estimation procedures in multiarmed designs with treatment selection presented by Prof. Gernot Wassmer (University of Cologne and ADDPLAN) [Download slides].

Abstract
In multiarmed clinical trials, treatment selection and sample size reassessment are straightforward and attractive applications when using an adaptive test design. Recently proposed procedures and designing options relevant for assessing the statistical performance of multiarmed adaptive designs are described. Particularly, calculation of repeated confidence intervals and overall p-values is considered.

Flexible group sequential designs: The conditional rejection probability approach presented by Prof. Hans-Helge Mueller (University of Marburg) [Download slides].

Abstract
During the course of a clinical trial it may be advantageous to adapt design elements although the statistical monitoring and analysis has been planned carefully. Redesigning the study by use of the Conditional Rejection Probability (CRP) principle (Müller and Schäfer, 2001, 2004) allows the researcher most flexibility while preserving the overall type I error rate.
In many studies there are two groups to be compared, an experimental group and a control. And for monitoring and analysis of the primary endpoint statistically modelling by means of a Brownian motion with drift and a fixed sample design or a group sequential design is appropriate, at least approximately by asymptotic theory. In such studies one can deal with the unexpected straightforward since the calculation of a conditional rejection probability is simple by use of standard statistical software packages and since the redesign is nothing else than carefully designing of a new study with the calculated conditional rejection probability taken as the type I error level. Methods for the construction of confidence sets reflecting early stopping for both, significance and futility, e.g. the confidence intervals by Tsiatis et al. (1984), can be developed further to cope with design modifications. However, these calculations will need some more efforts with standard software. And connected intervals are no longer guaranteed when using the most common method of Tsiatis et al.
After an introduction, the basic methods for flexible group sequential testing and confidence set estimation following design modifications based on the CRP-principle will be presented and illustrated. In an outlook, refinements of the CRP approach in the case of studies with small to moderate sample sizes where statistically modelling with nuisance parameters is more adequate (t-test as an example) and further developments of flexible multiple testing procedures incorporating selection strategies by use of the CRP approach will be adumbrated.

References
Tsiatis AA, Rosner GL, Metha CR. Exact confidence intervals following a group sequential test. Biometrics 1984;40:797-803.
Müller H-H, Schäfer H. Adaptive Group Sequential Designs for Clinical Trials: Combining the Advantages of Adaptive and of Classical Group Sequential Approaches. Biometrics 2001;57(3):886-891.
Müller HH, Schäfer H. A general statistical principle for changing a design any time during the course of a trial. Statistics in Medicine 2004;23:2497-2508.

Selection and bias -- Two hostile brothers presented by Prof. Peter Bauer (Medical University of Vienna) [Download slides].

Abstract
The situation is considered where at a first stage of a clinical trial several treatments are compared to a single control and the "best" treatment(s) are selected in an interim analysis. Bias and mean square error of conventional estimates after selection are quantified depending on the number of treatments and selection time during the trial. Results are shown for the simple strategies "select the best at interim", "select the two best at interim" and a more complex decision rule. The cases without and with reshuffling the planned sample size of the dropped treatments to the selected ones are treated. The mean bias shows very different patterns depending on the selection rule and the unknown parameters. Finally bias and mean square error are quantified for a simple but influential regulatory selection rule, to register a new medical therapy only when two pivotal trials both have proven an effect by a statistical test.

Adaptive dose-ranging studies: An update from the PhRMA working group presented by Dr. José Pinheiro (Novartis) [Download slides].

Abstract
The Adaptive Dose-Ranging Studies (ADRS) working group was formed by PhRMA with the goal of evaluating the potential benefit of adaptive and model-based approaches in improving dose selection and dose-response determination in clinical drug development. The ADRS WG undertook extensive simulation studies comparing a variety of dose-ranging and model-based methods, publishing a set of conclusions and recommendations on the use of these methods in drug development practice. This talk will present an overview of the second round of evaluations produced by the WG, focusing on additional adaptive dose-ranging methods, the use of exposure-response models in dose finding, and the impact of dose selection on the probability of success and expected net present value of development programs. Updated conclusions and recommendations will be presented and discussed.