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).

2010 adaptive design lecture series.

The DMC experience in response adaptive learn trials presented by Drs. Parvin Fardipour and Gary Littman (Pfizer) [Download slides].

Abstract
The planning and execution of clinical trials utilizing response-adaptive randomization is logistically complex. While a standardized approach to such designs is worthwhile, there are also unique challenges with each trial. Three detailed case studies are provided to illustrate the types of logistical issues that arise, as well as the decision problems coming before the data monitoring committees (DMCs). Several lessons learned from these case studies are presented. These range from consideration of more extreme scenarios for trial simulation during the planning stage to formal establishment of an executive steering committee to facilitate the decision process (even for learn-phase trials).

Type I error rate control in adaptive designs for confirmatory clinical trials with treatment selection at interim presented by Dr. Frank Bretz (Novartis AG and University of Hannover) [Download slides].

Abstract
Clinical trial designs allowing for design adaptations based on interim data have raised increasing interest in the past few years. An important application are adaptive designs for multi-armed studies. These studies start comparing several treatments (say) with a control. One (or more) treatment(s) are then selected after the first stage based on the available information at an interim analysis, including interim data from the ongoing trial, external information and expert knowledge. Recruitment continues, but now only for the selected treatment(s) and the control, possibly in combination with a sample size reassessment. The final analysis of the selected treatment(s) includes the patients from both stages and is performed such that the overall Type I error rate is strictly controlled, thus providing confirmatory evidence of efficacy. We review several approaches to control the Type I error rate in adaptive designs, describe their underlying assumptions and discuss their limitations if some of the assumptions are violated.

Manufacture and drug supply recommendations for adaptive trials presented by Nancy Burnham (GSK), Micheline Marshall (Pfizer) and Nitin Patel (Cytel) [Download slides].

Abstract
Appropriate management of the drug supply is a critical factor in the success of most adaptive trials. Therefore it is important to understand how adaptive designs affect traditional clinical supply chain strategies. The Adaptive Design Working Group's work stream on Manufacture and Drug Supply has consolidated experiences across industry and produced a paper intended to provide guidance and recommendations for drug supply management in support of adaptive trial designs. A summary of this work and their recommendations will be presented and discussed.

Adaptive sample size re-estimation in randomized clinical trials presented by Dr. Cyrus Mehta (Cytel) [Download slides].

Abstract
The sample size of a clinical trial is usually determined by statistical power calculations. Often, however, there is uncertainty and debate as to what magnitude of alternative hypothesis treatment difference and between-patient variability it is appropriate to pre-specify. Hence there exist methods to facilitate re-estimation of the sample size in the light of interim results from the ongoing trial. Statistical methods based only on interim estimates of within-patient variability are well-understood and not the focus of this presentation. We are concerned here with adaptive sample size re-estimation in major Phase III trials based on unblinded interim estimates of the primary effect size. This controversial topic generates two areas of concern:
1) the need for a robust statistical methodology for sample size re-estimation and its consequences for making inferences from the final trial data,
2) the practical organization of such an adaptive approach paying due regard to the confidentiality of interim data and the need to preserve the integrity of the trial's conduct throughout. We will discuss these issues, and illustrate their importance through examples of actual clinical trials.

How liberal can learning-free Type I error rate be in confirmatory adaptive design trials? presented by Dr. Sue-Jane Wang (Office of Biostatistics, CDER, U.S. FDA) [Download slides].

Abstract
A two-stage adaptive design trial is a single trial that combines the learning data from Stage 1 (often considered as Phase II or Phase IIB) and the confirming data in Stage 2 (often considered as phase III) for formal statistical testing. We call it a "Learn and Confirm" trial. The studywise Type I error rate remains to be at issue in a "Learn and Confirm" trial. For studying multiple doses or multiple endpoints, a "Learn and Confirm" adaptive design can be more attractive than a fixed design approach. This is because intuitively the learning data in Stage 1 should not be subject to type I error scrutiny if there is no formal interim analysis performed and only an adaptive selection of design parameters is made at Stage 1. We will call it learning-free Type I error rate. In this work we show that regardless of whether or not there is a formal interim analysis for making an adaptive selection, the learning-free Type I error rates are always at risk of inflation and should not be overlooked.
The views presented are the author's professional opinions and do not represent those of the U.S. FDA.

Information-based sample size adaptation for binomial trials presented by Dr. Keaven Anderson (Merck Research Laboratories) [Download slides].

Abstract
The FDA draft guidance recommends blinded sample size re-estimation as a well-understood method of adaptation. We demonstrate methods for information-based sample size re-estimation in a group sequential setting for binomial trials depending on whether the alternate hypothesis of most interest is stated as a risk difference, odds-ratio or relative risk. This is useful when a minimal clinical difference of interest is known, and the primary adaptation is for the overall event rate in the trial. Definitive interim analysis results could then be used to adapt for other treatment effects. Information is computed as a function of the variance of the appropriately scaled treatment difference. Practical considerations are dealt with and routines based on the gsDesign R package will be available to implement the methods. We provide an example for risk-difference and compare an unblinded sample size reassessment scheme. We provide a second example based on relative risk. Each example demonstrates good power over a range of control group response rates. We provide a preliminary comparison with the recently presented conditional power method by Gao et al.

Adaptive designs for dose-ranging trials: ADRS Working group simulation study presented by Dr. Vlad Dragalin (Quintiles, Innovations) [Download slides].

Abstract
Adaptive dose-ranging study (ADRS) designs may result in power gains to detect dose response and higher precision in estimating the target dose and the dose response curve. ADRS Working group has recently complemented the library of available methods with five new adaptive dose-ranging designs. Due to their inherent complexity, the operating characteristics of these designs can be assessed only through intensive simulations. We present here results of a comprehensive simulation study that compares and contrasts these designs for a variety of different scenarios.

Good practices for adaptive clinical trials in pharmaceutical product development presented by Dr. Brenda Gaydos (Eli Lilly) and Keaven Anderson (Merck Research Laboratories) [Download slides].

Abstract
This presentation will be based on the paper, "Good Practices for Adaptive Clinical Trials in Pharmaceutical Product Development". A summary will be provided of best practices for the planning and implementation of adaptive designs compiled from experiences gained in the pharmaceutical industry. The target audience is anyone involved in the planning and execution of clinical trials. Strategic points for consideration at the design and implementation stage will be discussed from an operational, regulatory, clinical and statistical perspective.