適應醫療實驗：醫療實驗設計、管理、分析的技術創新

The pharma industry is gradually coming to realize that the classically structured clinical trial does not offer enough flexibility to make use of continuously emerging knowledge that is generated as the trial progresses. This report is a comprehensive assessment of the benefits, challenges, and accumulated industry experience with regard to adaptive clinical trials.

• A critique of the structural, conceptual, and ethical issues inherent in the traditional clinical trial.
• An in-depth review, based on actual case studies (e.g., Napo, Genaera, Pfizer, Lilly, Millennium, and various academic institutions) of the use of adaptive and Bayesian approaches in Phases I, II, and III
• An assessment of various hybrid and seamless designs in which the line between trial stages is blurred
• Evolving regulatory positions of FDA, EMEA, and ICH on adaptive designs; industry response and initiatives
• A review of specialized software vendors (e.g., Cytel, Tourtellotte, Pharsight, CTriSoft) and their applications that have emerged to support adaptive designs.
• A CHI Insight Pharma survey of the views and experiences of individuals involved with adaptive designs
• Three future scenarios for the integration of adaptive designs in clinical trials by 2015: (1) United States Leads the Way, (2) Globally Integrated Midphase Revamping, and (3) the Late Phoenix Scenario
• Roundtable interviews with senior executives in industry and consulting who bring decades of combined experience in adaptive and Bayesian clinical trial designs

Unacceptable levels of attrition in the clinical stage of development are driving profound changes in the architecture, design, and analysis of clinical trials. The majority of respondents to our survey said that reduction in patient numbers, less exposure to study drug, and drops in overall trial duration were key points in favor of adaptive designs; however, a majority also had specific concerns with adaptive trials�\concerns that involved methodological, logistical, and regulatory uncertainties:

Methodological: Will adaptive and/or "seamless" designs lead the sponsor to erroneous conclusions if used in Phase II and (in particular) in pivotal trials?

Logistical: Can such trials be fully kept under control without major organizational change/expansion and/or increased dependency on outside statistical and monitoring advice?

Regulatory: Will regulatory authorities (FDA, EMEA) accept adaptive designs and- more importantly- will they accept the trial sponsor' s interpretation of the results of such trials?

Adaptive Clinical Trials: Innovations in Trial Design, Management, and Analysis examines the key challenges involved in adaptive trials, such as:

• Staff training requirements
• EDC to enable near-real-time capture, validation, and analysis of trial-emergent data
• Working with Data Monitoring Committees (DMC)
• Ways in which adaptive modifications �\ e.g., dropping and replacing a dosage arm �\ can have ripple effects on a project' s critical path
• The challenges of prognosis, analysis and interpretation

Adaptive trials are emerging as a solution to lengthening development timelines, runaway costs, and overall lack of predictivity in the drug development process. Drug companies, regulators, and academic investigators have shown a willingness to explore adaptive and Bayesian options as part of a general revamping of the drug development process. We predict that increasing guidance and endorsement from regulatory bodies will lead the industry to fully embrace adaptive trials by 2015.

Table of Contents

INTRODUCTION

• 1.1. General Objectives of a Clinical Trial Program
  • Phase I
  • Phase II
  • Phase III
• 1.2. Science, Regulatory Affairs, and Business: Interdependence and Inertia

THE TRADITIONAL CLINICAL TRIAL: AN EXPERIMENT IN MEDICAL STATISTICS

• 2.1. The Classical Phase I Trial
• 2.2. Implicit and Explicit Assumptions for Phase II and III Efficacy Trials
  • Representativeness of the Trial Population
  • Matching Comparison Groups
  • Strength of Treatment Effect and Compliance
  • Choice of Treatment Duration and Outcome Measures
  • The Neyman-Pearson Approach: Fixed Parameters and the Null Hypothesis
• 2.3. Phase III Today: Unmanageably Large, Prohibitively Expensive
• 2.4. Rescue Attempts for Efficacy Trials with Unexpected Negative Outcome
• 2.5. The Ethical Side
  • The Call to Minimize Exposure in Phase I
  • Forced Unblinding, Active Comparators, and Add-on Trials

ADAPTIVE TRIAL DESIGNS

• 3.1. The Need for Alternate Approaches to Clinical Trials
• 3.2. Adaptive Integration of Pre-existing and Trial-Emergent Knowledge
  • Building an Adaptive Design: "Flexible" Does Not Mean "Less Controlled"
  • Adaptive Statistics: Bayes' Theorem
  • Monte Carlo Methods: Not Gambling, but Probability Design
• 3.3. Application to Phase I: Continual Reassessment Methods and Accelerated Titration Schemes
  • The Expanding Spectrum of CRMs
  • Other Methods
• 3.4. Applications to Phase II: Achieving Dose Finding with Adaptive Randomization
  • The PhRMA Working Group on Adaptive Dose-Ranging Studies
  • Case Studies of Phase IIa Dose-Finding Studies Using Adaptive Designs
• 3.5. Adaptive Strategies for Phase III: The Greatest Challenge
  • Challenges of Trial Design
    • Response-Adaptive Randomization
    • Modifying Trial Endpoints and Outcome Measures
  • Challenges of Conduct: Monitoring Adaptive Trials
    • Staff Training Requirements
    • The Pervasive Near-Real-Time Paradigm
    • Problems in Trial Organization and Logistics
  • Challenges of Prognosis, Analysis, and Interpretation
    • The Simulated Trial: Probabilistic Predictions of the Trial Course
    • Synthesizing and Reporting Outcomes of Adaptive Trials
      • Incorporating Direct and Indirect Evidence
      • A New Way of Handling Missing Data
      • Reporting Issues
  • Case Studies in Adaptive Trials and Their Simulation
    • "Post-hoc Prognosis": Modeling Bayesian Stop Decision Based on Real Trial Data
      • Chemoprophylaxis for Bacterial Coinfection
      • Cure Rate Models for Malignant Melanoma
    • Selected Examples of Actual Adaptive Phase III Trials
      • PURSUIT: Eptifibatide in Unstable Angina
      • ASTIN: A Bayesian Adaptive Dose-Response Trial in Acute Stroke
      • ADVENT: Anti-Diarrhea Therapy in HIV Disease
      • COPERNICUS: A Beta Blocker in Advanced Heart Failure

HYBRID AND SEAMLESS DESIGNS

• 4.1. Blurring the Line Between Trial Stages
• 4.2. Phase I/II Hybrid Designs
• 4.3. The Integrated Phase II/III Adaptive Trial
  • PhRMA Adaptive Design Workshop

THE ADAPTIVE APPROACH, INDUSTRY, AND REGULATORY AUTHORITIES

• 5.1. The Industry Position
• 5.2. FDA, EMEA, and ICH Positions
  • US Food and Drug Administration
  • EMEA Guidelines in Preparation
  • ICH: Toward Future Global Harmonization of Global Trial Principles?
• 5.3. Summary

CHI INSIGHT PHARMA REPORTS- ADAPTIVE TRIALS IN CURRENT PRACTICE SURVEY- MAY 2007

• 6.1. Participant Demographics
• 6.2. Analysis of the Number and Type of Trials Conducted
• 6.3. Motivations, Plans for the Future, and Perceived Impediments to Adaptive Trials
• 6.4. Summary Interpretation ansd Comments

ADAPTIVE TRIALS TO 2015: SCENARIOS FOR ACCEPTANCE

• 7.1. A Multilayered Web of Concern
• 7.2. Three Scenarios to 2015
  • Scenario A
  • Scenario B
  • Scenario C
  • Comparison of Scenarios
• 7.3. Other Factors Beyond Anticipation
• 7.4. Summary

INTERVIEWS

References

Glossary of Selected Terms

Company Index with Web Addresses