Optimal Designs for Multi-Arm Phase II/III Drug Development Programs

Abstract

In drug development programs, phase II studies may be carried out as multi-arm trials aiming to identify treatment(s) with the optimal benefit-to-risk profile(s). Succeeding confirmatory phase III trials then attempt to demonstrate efficacy and safety of the designated treatment(s). While upmost important to consider multi-arm phase II/III study(ies) holistically to optimize drug development, sample size allocation with sound go/no-go decision rules has become more complex to handle but critical to guide decision making. For example, one has to decide whether to conduct the phase III trial with a single (the most promising) treatment only, or with multiple treatments (if sufficiently promising). We propose a framework for program-wise phase II/III drug development planning aiming to optimize the sample size allocation and go/no-go decision rule with respect to the maximal expected utility for time-to-event endpoints. The approach is based on a utility function taking into account, for example, fixed and variable costs of the program, expected benefits after successful market launch, assumed true treatment effects, and event rates. We illustrate our method by application to practical examples. Our results show that program-wise drug development planning is indispensable. This can be done by using a user-friendly R Shiny application. Supplementary materials for this article are available online.

KEYWORDS:

• Assurance
• Decision rule
• Optimization
• Probability of success
• Sample size allocation

Supplementary Materials

The supplementary materials include detailed formulas for calculating the expected cost and gain for Strategy 1 and 2 (A1), an example on how to introduce a time factor into the utility function to model the impact of program duration (A2), and the optimization results for scenarios where patients are limited (i.e., N = 1000; A3).

Conflict of Interests

Heiko Götte has no conflict of interest with the subject matter of this article while being an employee of Merck Healthcare KGaA.

Additional information

Funding

We would like to thank the Deutsche Forschungsgemeinschaft (DFG) for supporting this research by grant KI 708/2-1.