Implementation of a Phase 1 Adaptive Clinical Trial in a Treatment of Type 2 Diabetes
- PDF / 101,227 Bytes
- 11 Pages / 612 x 792 pts (letter) Page_size
- 44 Downloads / 186 Views
Fabián S. Tibaldi, PhD Senior Research Scientist, Eli Lilly and Co., MontSaint-Guibert, Belgium Benoit H. L. Beck, Ir, PhD Principal Research Scientist, Eli Lilly and Co., MontSaint-Guibert, Belgium Alun Bedding Director, Biostatistics and Programming Development Partners, Drug Development Sciences, GlaxoSmithKline, Harlow, Essex, United Kingdom
Key Words Adaptive designs; Bayesian models; R software; WinBUGS; Markov chain; Monte Carlo methods Correspondence Address Fabián S. Tibaldi, PhD, Eli Lilly and Co., Rue Granbonpré 11, 1348 MontSaint-Guibert, Belgium (email: [email protected]).
455
Implementation of a Phase 1 Adaptive Clinical Trial in a Treatment of Type 2 Diabetes Phase 1 clinical trials are in general conducted to determine the dose response of a new drug with respect to safety and tolerability (1). There is an increasing interest in reducing the trial time in order to have new and better drugs available to patients in a minimum amount of time. Adaptive designs for dose response have been intensively studied in recent years and have been considered as a novel approach to minimize the number of subjects or patients used in
INTRODUCTION Pharmaceutical companies are becoming more interested in conducting clinical trials in such a way that the time to market is significantly reduced such that the best treatments are available to patients. Therefore, it is important to identify the good candidates as early as possible in the development process and to select the optimal dose levels for any new treatment that will be taken forward to later phases of clinical development. Phase 1 clinical trials traditionally use standard parallel or crossover designs. Without a loss of generality in this work, we focus only on parallel designs. Various initiatives have been initiated to tackle the issues with adaptive designs (7). For example, if an adaptive randomization is employed to examine dose response, treatments are allocated to patients based on previous responses in the trial, such that as much knowledge as possible is gained about the dose-response curve. This learning process is efficient, leading to improved precision for the estimated quantities of interest for any given number of subjects. While formal adaptive designs are often used in phase 2 trials (8–10), in phase 1 these approaches are not so common outside the oncology area (10), although interest is growing. From a practical point of view, the implementation can be done through high-level programming
clinical trials and to reduce the risk of serious adverse events. Advantages and disadvantages of this kind of approach have been discussed in the literature and are not an object of discussion for this article (2). However, the implementation of this type of study is still challenging (3–7). This article focuses on the practical advantages of a Bayesian adaptive design and shows how it can be implemented using the R2WinBUGS freely downloadable software.
languages (eg, Pascal, FORTRAN, C++, etc); however, this implementation can be cumbersome and
Data Loading...
