1st Edition
Innovative Strategies, Statistical Solutions and Simulations for Modern Clinical Trials
"This is truly an outstanding book. [It] brings together all of the latest research in clinical trials methodology and how it can be applied to drug development…. Chang et al provide applications to industry-supported trials. This will allow statisticians in the industry community to take these methods seriously." Jay Herson, Johns Hopkins University
The pharmaceutical industry's approach to drug discovery and development has rapidly transformed in the last decade from the more traditional Research and Development (R & D) approach to a more innovative approach in which strategies are employed to compress and optimize the clinical development plan and associated timelines. However, these strategies are generally being considered on an individual trial basis and not as part of a fully integrated overall development program. Such optimization at the trial level is somewhat near-sighted and does not ensure cost, time, or development efficiency of the overall program. This book seeks to address this imbalance by establishing a statistical framework for overall/global clinical development optimization and providing tactics and techniques to support such optimization, including clinical trial simulations.
- Provides a statistical framework for achieve global optimization in each phase of the drug development process.
- Describes specific techniques to support optimization including adaptive designs, precision medicine, survival-endpoints, dose finding and multiple testing.
- Gives practical approaches to handling missing data in clinical trials using SAS.
- Looks at key controversial issues from both a clinical and statistical perspective.
- Presents a generous number of case studies from multiple therapeutic areas that help motivate and illustrate the statistical methods introduced in the book.
- Puts great emphasis on software implementation of the statistical methods with multiple examples of software code (both SAS and R).
It is important for statisticians to possess a deep knowledge of the drug development process beyond statistical considerations. For these reasons, this book incorporates both statistical and "clinical/medical" perspectives.
- Overview of Drug Development
- Clinical Development Plan and Clinical Trial Design
- Clinical Development Optimization
- Globally Optimal Adaptive Trial Designs
- Trial Design for Precision Medicine
- Clinical Trial with Survival Endpoint
- Practical Multiple Testing Methods in Clinical Trials
- Missing Data Handling in Clinical Trials
- Special Issues and Resolutions
- Issues and Concepts of Data Monitoring Committees
- Controversies in Statistical Science
Introduction
Drug Discovery
Target Identi_cation and Validation
Irrational Approach
Rational Approach
Biologics
NanoMedicine
Preclinical Development
Objectives of Preclinical Development
Pharmacokinetics
Pharmacodynamics
Toxicology
Intraspecies and Interspecies Scaling
Clinical Development
Overview of Clinical Development
Classical Clinical Trial Paradigm
Adaptive Trial Design Paradigm
New Drug Application
Summary
Clinical Development Program
Unmet Medical Needs & Competitive Landscape
Therapeutic Areas
Value proposition
Prescription Drug Global Pricing
Clinical Development Plan
Clinical Trials
Placebo, Blinding and Randomization
Trial Design Type
Confounding Factors
Variability and Bias
Randomization Procedure
Clinical Trial Protocol
Target Population
Endpoint Selection
Proof of Concept Trial
Sample Size and Power
Bayesian Power for Classical Design
Summary
Benchmarks in Clinical Development
Net Present Value and Risk-Adjusted NPV Method
Clinical Program Success Rates
Failure Rates by Reason
Costs of Clinical Trials
Time-to-Next Phase, Clinical Trial Length and
Regulatory Review Time
Rates of Competitor Emerging
Optimization of Clinical Development Program
Local Versus Global Optimizations
Stochastic Decision Process for Drug Development
Time Dependent Gain g,
Determination of Transition Probabilities
Example of CDP Optimization
Updating Model Parameters
Clinical Development Program with Adaptive Design
Summary
Common Adaptive Designs
Group Sequential Design
Test Statistics
Commonly Used Stopping Boundaries
Sample Size Reestimation Design
Test Statistic
Rules of Stopping and Sample-Size Adjustment
Simulation Examples
Pick-Winner-Design
Shun-Lan-Soo Method for Three-Arm Design
K-Arm Pick-Winner Design
Global Optimization of Adaptive Design - Case Study
Medical Needs for COPD
COPD Market
Indacaterol Trials
US COPD Phase II Trial Results
Optimal Design
Summary & Discussions
Introduction
Overview of Classical Designs with Biomarkers
Biomarker-enrichment Design
Biomarker-Stratified Design
Sequential Testing Strategy Design
Marker-based Strategy Design
Hybrid Design
Overview of Biomarker-Adaptive Designs
Adaptive Accrual Design
Biomarker-Informed Group Sequential Design
Biomarker-Adaptive Threshold Design
Adaptive Signature Design
Cross-Validated Adaptive Signature Design
Trial Design Method with Biomarkers
Impact of Assay Sensitivity and Specificity
Biomarker-Stratified Design
Biomarker-Adaptive Winner Design
Biomarker-Informed Group Sequential Design
Basket and Population-Adaptive Designs
Basket Design Method with Familywise Error Control
Basket Design for Cancer Trial with Imatinib
Methods based on Similarity Principle
Summary
Overview of Survival Analysis
Basic Taxonomy
Nonparametric Approach
Proportional Hazard Model
Accelerated Failure Time Model
Frailty Model
Maximum Likelihood Method
Landmark Approach and Time-Dependent Covariate
Multistage Models for Progressive Disease
Introduction
Progressive Disease Model
Piecewise Model for Delayed Drug Effect
Introduction
Piecewise Exponential Distribution
Mean and Median Survival Times
Weighted LogRank Test for Delayed Treatment Effect
Oncology Trial with Treatment Switching
Descriptions of the Switching Problem
Treatment Switching
Inverse Probability of Censoring Weighted LogRank Test
Removing Treatment Switch Issue by Design
Competing Risks
Competing Risks as Bivariate Random Variable
Solution to Competing Risks Model
Competing Progressive Disease Model
Hypothesis Test Method
Threshold Regression with First-Hitting-Time Model
Multivariate Model with Biomarkers
Summary
Multiple-Testing Problems
Sources of Multiplicity
Multiple-Testing Taxonomy
Union-Intersection Testing
Single-Step Procedure
Stepwise Procedures
Single-Step Progressive Parametric Procedure
Power Comparison of Multiple Testing Methods
Application to Armodafinil Trial
Intersection-Union Testing
The Need for Coprimary Endpoints
Conventional Approach
Average Error Method
Li-Huque's Method
Application to a Glaucoma Trial
Priority Winner Test for Multiple Endpoints
Finkelstein-Schoenfeld's Method
Win-Ratio Test
Application to Charm Trial
Summary
Missing Data Problems
Missing Data Issue and Its Impact
Missing Mechanism
Implementation of Analysis Methods
Trial Data Simulation
Single Imputation Methods
Methods without Specified Mechanics of Missing
Inverse-Probability Weighting Method
Multiple Imputation Method
Tipping Point Analysis for MNAR
Mixture of Paired and Unpaired Data
Comparisons of Different Methods
Regulatory and Operational Perspective
Overview
Drop-Loser Design Based on Efficacy and Safety
Multi-stage Design with Treatment Selection
Dunnett Test with Drop-losers
Drop-Loser Design with Gatekeeping Procedure
Drop-loser Design with Adjustable Sample Size
Drop-Loser Rules in Term of Efficacy and Safety
Simulation Study
Clinical Trial Interim Analysis with Survival Endpoint
Hazard Ratio versus Number of Deaths
Conditional Power
Prediction of Timing for Target Number of Events
Power and Sample Size for One-Arm Survival Trial Design
Estimation of Treatment Effect with Interim Blinded Data
Likelihood
MLE Method
Bayesian Posterior
Analysis of Toxicology Study with Unexpected Deaths
Fisher versus Barnard's Exact Test Methods
Wald statistic
Fisher's Conditional Exact Test p-value
Barnard's Unconditional Exact Test p-value
Power Comparisons of Fisher's versus Barnard's Tests
Adaptive Design with Mixed Endpoints
Summary
Overview of the DMC
Operation of the DMC
Role of the DMC Biostatistician
Requirement for a DMC
Use of a DMC in Rare Disease Studies
Statistical methods for Safety Monitoring
Statistical methods for interim efficacy analysis
Summary and Discussion
What is a Science?
Similarity Principle
Simpson's Paradox
Causality
Type-I Error Rate and False Discovery Rate
Multiplicity Challenges
Regression with Time-Dependent Variables
Hidden Confounders
Controversies in Dynamic Treatment Regime
Paradox of Understanding
Summary and Recommendations
Biography
Mark Chang, John Balser, Jim Roach, Robin Bliss
"This is the first edition of a comprehensive book covering the most recent methodology on innovative clinical trial designs for drugs and biological products. It is a great reference book for statisticians, clinicians, and other stakeholders involved in drug discovery and development. ... Chang et al aimed to provide the statistical framework to reach the overall development program optimizations in this book. In addition, innovative methodology to mitigate the risks of failed efficacy, safety, strategy, commercial and operation failures have been described by Chang et al. Special techniques such as clinical trial simulations are highly recommended by the authors. ....
In summary, this is an excellent reference book for statisticians, clinicians, and all stakeholders involved in clinical development program with a common goal to reach clinical development optimization.”
—Holly Huang in the Journal of Biopharmaceutical Statistics, October 2019"The book has a number of detailed examples, and SAS and R code to implement some of the methods described in the text...In summary, this book covers a wide range of interesting topics in clinical trials, and provides an appealing and useful reference to researchers."
"This first edition of this book provides the most recent methodology and statistical considerations in the design and management of clinical trials. It is focused on professionals in drug development, specifically statisticians and clinical researchers...It is, however, a useful insight into strategies for specific situations including personalized medicine, missing data, adaptive design, and multiple testing. The authors include a large number of practical clinical trials from various therapeutic areas, and they put emphasis on the use of clinical trial simulations...The authors present a remarkable amount of SAS code examples that could be directly used in daily practice. An extensive overview of modern innovative strategies for clinical trials helps to broaden the horizons of scientists interested in drug or treatment development."
- Ionut Bebu, JASA 2020
- Iveta Selingerová, ISCB News, July 2020