• A well designed adaptive trial can mitigate the risk of having a failed clinical trial
• Systematically and rigorously addressing key uncertainties that exist during the design of a confirmatory trial
• Avoiding “anticipated regret” and avoiding undue focus on minor threats to trial success
• When a trial is adaptive “by design,” all relevant statistical threats to trial validity can be rigorously addressed via simulation
• The productive use of adaptive designs is limited mostly by the conservative nature of those designing trials and by a paucity of statisticians familiar with adaptive design and simulations methods, rather than by substantive theoretical or regulatory concerns.

Traditional oncology trials are inefficient, expensive, and suffer from high failure rates
Adaptive design approaches can be incorporated in all phases of oncology research, including:

More accurate determination of maximum tolerated dose in Phase I
Better understanding of dose response and target population in Phase II
Overcoming high failure rates of phase III trials
A major opportunity for mitigating risk in oncology drug development is improvement in Phase II strategies to reduce risk of late Phase III failure and to improve the chance of overall success

Traditional oncology trials are inefficient, expensive, and suffer from high failure rates
Adaptive design approaches can be incorporated in all phases of oncology research, including:

More accurate determination of maximum tolerated dose in Phase I
Better understanding of dose response and target population in Phase II
Overcoming high failure rates of phase III trials
A major opportunity for mitigating risk in oncology drug development is improvement in Phase II strategies to reduce risk of late Phase III failure and to improve the chance of overall success

• The research question – Alzheimer’s Disease
• Applying adaptive design methodology to the research question
• Biomarkers as enablers for the adaptive design
• Case Study

• Oncology Ph I: Challenges & Design Requirements
• Design Overview
• Implementing a Bayesian Design
• Case Study

• Two principle approaches for designing multi-arm multi-stage clinical trials
• Discussion: Options for selecting treatments
• Software to design such studies is presented

• Main uncertainties and interactions between adaptive trial design, patient recruitment,  randomization
• Adaptive patient recruitment prediction and trial cost modelling
• Data-driven predicting trial performance and site productivity
• Predictive event modelling and adaptive recruitment adjustment
• Optimization of different stages of drug development

• Preclinical data is very often underused, whereas they can provide very valuble prior information for Bayesian approach
• Adaptive design can also be implemented in non-clinical studies to maximise the information extracted from the study
• This extenses approachs opens the way to a global approach of drug development; fully data driven and allowing early selection of a  winning area for a project.

• Examples of adaptive designs in early stage pediatric trials are documented and hopefully growing
• However for late stage confirmatory trials (to gain regulatory approval and license) examples seem fewer and further between efficient trial designs should give the same benefits to patients and clinical development programs, regardless of patient age.
• By discussion of the potential barriers and challenges within pediatric studies, with the help of recent case studies, perhaps we can improve pediatric trial designs.

• Hypothesis Testing in Clinical Trials
• Should Type I Error be Fixed in Drug Development?
• An Alternative to Maximising Power for Fixed type I Error
• Traditional values for Type I and Type II Error Rates and Implications for the Relative Costs of These Errors
• An Alternative Form of the Neyman-Pearson Fundamental Lemma
• The Likelihood Principle and Adaptive Trials

• Motivation for enrichment trials in oncology
• Adaptive enrichment design for PFS endpoints
• Simulation tools to evaluate power under different scenarios
• Future directions

• Understanding the adaptive p-value
• Comparing fixed vs adaptive p-values-when is an adaptive p-value "better"
• A case study for the promising zone in asthma

• Use a predictive model to provide supporting information for dose-escalation decisions for a data monitoring committee
• Design the model to provide high-quality predictions; the model must be fully trusted by the clinical team and hence based on prior information dwarn form expert opinion and on prior research
• Case study illustrating design of such a model and use in practice

• Designing a clinical development programme includes a string of decisions. The design problem is multi-staged, iterative and collaborative, with extensive communication among teams of experts in various disciplines, and with a problem statement that often is ill-defined to start with and gradually becomes clearer.
• A framework and tool were developed and implemented for transparently translating target product claims into study specifications so that scientific questions as well as underlying assumptions and risks are clearly defined.
• Case study.

• Methods for adaptive sample size re-estimation at interim analyses
• Understanding Mehta and Poock's "indifference zone" approach
• Improving Mehta and Pocock's approach to make better use of additional sample size
• Use of group sequential stopping rules with a delayed response

• Protocol development
• Quality of data and monitoring
• Governing the logistics
• Chronology of the steps and critical points

• Introduction of BVS
• Application to dose-response modelling
• Comparison with other methods
• Case study
• Review and discussion