MedianaDesigner

Package

Mediana has released a beta version of an R package (MedianaDesigner) that supports efficient simulation-based power and sample size calculations for a broad class of late-stage clinical trials, including adaptive designs.

The following seven modules are currently included in the package:

  • Adaptive designs with data-driven sample size or event count re-estimation.
  • Adaptive designs with data-driven treatment selection.
  • Adaptive designs with data-driven population selection.
  • Optimal selection of a futility stopping rule.
  • Event prediction in event-driven trials.
  • Adaptive trials with response-adaptive randomization.
  • Traditional trials with multiple objectives.

These modules are described below.

The latest version of this package is available on Github. In addition, free web applications that facilitate simulation-based trial design are available on our Cloud platform. The applications can be accessed using the following login information:

  • Login: MedianaDesigner
  • Password: Rt5UKD8pzm

Additional  information and multiple case studies can be found in the online manual.

For more information on open-source and free software developed by Mediana, visit the free software page.

Adaptive designs with sample-size or event count re-estimation

The most common type of adaptive designs in Phase III trials or seamless Phase II/Phase III trials supports an option to examine the unblinded efficacy data at an interim look to re-estimate the total sample size or total number of events in the trial. This data-driven adjustment helps boost the probability of success at the final analysis.

This module supports two-arm adaptive designs with three decision points (two interim analyses followed by a final analysis). The first interim analysis is introduced to enable a futility assessment, i.e., the trial will be stopped at this interim look if the treatment is unlikely to be effective. The second interim analysis includes an option to improve the trial’s probability of success by increasing the number of patients or events up to a pre-defined cap.

For more information about this module, download the technical manual:

Adaptive designs with data-driven treatment selection

Adaptive designs are often employed in multi-arm Phase III trials to identify the most promising doses or regimens of an experimental treatment, e.g., doses or regimens with the best risk-benefit ratio. These trial designs are known as adaptive treatment selection designs.

This module assumes that a Phase III trial or seamless Phase II/Phase III trial will be conducted to evaluate the efficacy and safety of several active arms versus control. Two interim analyses will be employed in the trial. The first interim look is aimed at a futility assessment during the Phase II portion of the trial and the second interim look represents the end of the Phase II portion. A pre-defined treatment selection rule will be applied at this interim look to select the most promising active arms. The selected active arms and the control arm will continue into the Phase II portion of the trial.

For more information about this module, download the technical manual:

Adaptive designs with data-driven population selection

Adaptive population selection designs are conceptually similar to adaptive treatment selection designs defined above and are aimed at an efficient characterization of the efficacy and safety profiles of experimental treatments across several populations of patients. Most commonly, two patient populations are considered, i.e., the overall population with the condition of interest and a subset of patients with a pre-defined characteristic such as a demographic variable or a biomarker. Adaptive population selection designs are often used in seamless Phase II/Phase III trial.

This module supports adaptive population selection designs in two-arm Phase III trials or seamless Phase II/Phase III trials with two pre-specified patient populations. Two interim looks will be employed in the trial, the first one being a futility analysis and the second one focusing on population selection. The efficacy signals in the two populations will be examined at the second interim analysis will identify the best strategy for the final analysis. The available options include the evaluation of the treatment effect in only one of the two patient populations or in both populations. If the subset is chosen as the most promising populations, patients from the complementary subset will not be enrolled after the second interim analysis.

For more information about this module, download the technical manual:

Optimal selection of a futility stopping rule

Consider Phase II or Phase III trials with futility assessments at an interim analysis. A futility stopping rule will be applied at this interim look and the trial will be stopped early due to futility if the treatment is unlikely to be effective (predicted probability of success at the final analysis is low).

This module derives an optimal futility stopping rule, which maximizes the sensitivity and specificity rates.

For more information about this module, download the technical manual:

Event prediction in event-driven trials

This module supports event predictions in Phase II or Phase III trials with an event-driven design, e.g., trials with endpoints based on overall survival or progression-free survival. In the context of adaptive designs with several decision points, it is important to reliably predict the timing of interim analyses that are conducted after a pre-defined number of events has been accrued. This module examines the available blinded data to project the number of events at any point in the future.

Download an example data set with patient enrollment, event and patient dropout information (CSV file).

For more information about this module, download the technical manual: