Flexible Clinical Trial Design, Simulation, and Analysis with rpact

Group sequential and adaptive designs

Statistical Methods

Trial Design

Intermediate

Authors

Daniel Sabanes Bove (RCONIS)

Friedrich Pahlke (Tech Entrepreneur)

Overview

Intermediate Trial Design Statistics

Explore the capabilities of {rpact} - a comprehensive, validated, open-source R package for clinical trial planning, design simulation, and data analysis. Under continuous development since 2017 with extensive documentation.

What You’ll Learn

📊 Trial design basics - Group sequential and adaptive designs
🎲 Simulation - Design characteristics assessment
🔬 P-value combination tests
📈 Adaptive designs - Multi-armed and enrichment
💻 RPACT Cloud - User-friendly platform

Prerequisites

Required Knowledge:

Intermediate statistics
Clinical trial basics
R programming fundamentals

Recommended:

Experience with trial design
Knowledge of group sequential methods

Key Tools

{rpact}

RPACT Cloud

Workshop Materials

Resources

Presentation: rpharma.presentation.2025.rpact.com

Documentation: www.rpact.org

RPACT Cloud (Free): cloud.rpact.com

Package Features

Design Types

Group Sequential Designs:

Multiple interim analyses
Early stopping rules
Alpha spending functions
Futility boundaries

Adaptive Designs:

Sample size reassessment
Treatment selection
Population enrichment
Combination tests

Data Types Supported

✅ Continuous endpoints
✅ Binary outcomes
✅ Survival data
✅ Count data

RPACT Cloud

User-friendly web platform for trial design:

Features:

No coding required
Interactive design exploration
Automatic documentation
Export to R code
Free version available

Use Cases:

Quick design evaluation
Stakeholder presentations
Teaching and training
Proposal preparation

Practical Applications

Sample Size Calculation

library(rpact)

# Group sequential design
design <- getDesignGroupSequential(
  kMax = 3,  # 3 interim analyses
  alpha = 0.025,
  beta = 0.2,
  typeOfDesign = "asOF"  # O'Brien-Fleming
)

# Calculate sample size
sampleSize <- getSampleSizeMeans(
  design = design,
  meanRatio = 1,
  normalApproximation = FALSE,
  alternative = 0.5,  # Treatment effect
  stDev = 1
)

Simulation

# Simulate adaptive design
simulation <- getSimulationMultiArmMeans(
  design = design,
  activeArms = 3,
  plannedSubjects = c(50, 100, 150),
  meanRatio = 1,
  stDev = 1,
  maxNumberOfIterations = 10000
)

Learning Outcomes

✅ Design group sequential trials
✅ Implement adaptive designs
✅ Simulate trial characteristics
✅ Use RPACT Cloud platform
✅ Apply {rpact} to real studies

Similar Workshops

Bayesian Survival Models - Alternative statistical approach
Debugging Stan - For Bayesian designs

Next Steps

For Bayesian designs: Try Bayesian Survival workshop

Last updated: November 2025 | R/Pharma 2025 Conference

--- title: "Flexible Clinical Trial Design, Simulation, and Analysis with rpact" subtitle: "Group sequential and adaptive designs" author: - "Daniel Sabanes Bove (RCONIS)" - "Friedrich Pahlke (Tech Entrepreneur)" categories: [Statistical Methods, Trial Design, Intermediate] --- ## Overview [Intermediate]{.badge .badge-intermediate} [Trial Design]{.badge .badge-category} [Statistics]{.badge .badge-category} Explore the capabilities of **{rpact}** - a comprehensive, validated, open-source R package for clinical trial planning, design simulation, and data analysis. Under continuous development since 2017 with extensive documentation. ### What You'll Learn - 📊 **Trial design basics** - Group sequential and adaptive designs - 🎲 **Simulation** - Design characteristics assessment - 🔬 **P-value combination** tests - 📈 **Adaptive designs** - Multi-armed and enrichment - 💻 **RPACT Cloud** - User-friendly platform ## Prerequisites ::: requirements **Required Knowledge:** - Intermediate statistics - Clinical trial basics - R programming fundamentals **Recommended:** - Experience with trial design - Knowledge of group sequential methods ::: ## Key Tools ::: tool-tag {rpact} ::: ::: tool-tag RPACT Cloud ::: ## Workshop Materials ::: callout-note ## Resources **Presentation:** [rpharma.presentation.2025.rpact.com](https://rpharma.presentation.2025.rpact.com/) **Documentation:** [www.rpact.org](https://www.rpact.org/) **RPACT Cloud (Free):** [cloud.rpact.com](https://cloud.rpact.com/) ::: ## Package Features ### Design Types **Group Sequential Designs:** - Multiple interim analyses - Early stopping rules - Alpha spending functions - Futility boundaries **Adaptive Designs:** - Sample size reassessment - Treatment selection - Population enrichment - Combination tests ### Data Types Supported - ✅ Continuous endpoints - ✅ Binary outcomes - ✅ Survival data - ✅ Count data ## RPACT Cloud User-friendly web platform for trial design: **Features:** - No coding required - Interactive design exploration - Automatic documentation - Export to R code - Free version available **Use Cases:** - Quick design evaluation - Stakeholder presentations - Teaching and training - Proposal preparation ## Practical Applications ### Sample Size Calculation ``` r library(rpact) # Group sequential design design <- getDesignGroupSequential( kMax = 3, # 3 interim analyses alpha = 0.025, beta = 0.2, typeOfDesign = "asOF" # O'Brien-Fleming ) # Calculate sample size sampleSize <- getSampleSizeMeans( design = design, meanRatio = 1, normalApproximation = FALSE, alternative = 0.5, # Treatment effect stDev = 1 ) ``` ### Simulation ``` r # Simulate adaptive design simulation <- getSimulationMultiArmMeans( design = design, activeArms = 3, plannedSubjects = c(50, 100, 150), meanRatio = 1, stDev = 1, maxNumberOfIterations = 10000 ) ``` ## Learning Outcomes ✅ Design group sequential trials\ ✅ Implement adaptive designs\ ✅ Simulate trial characteristics\ ✅ Use RPACT Cloud platform\ ✅ Apply {rpact} to real studies ------------------------------------------------------------------------ ### Similar Workshops - [Bayesian Survival Models](bayesian-survival.qmd) - Alternative statistical approach - [Debugging Stan](debugging-stan.qmd) - For Bayesian designs ### Next Steps - **For Bayesian designs:** Try [Bayesian Survival workshop](bayesian-survival.qmd) ------------------------------------------------------------------------ *Last updated: November 2025 \| R/Pharma 2025 Conference*