Robert T. O’Neill Ph.D.
Director , Office of Biostatistics
Center for Drug Evaluation and Research
Presented at the 2005 FDA/Industry Statistics Workshop: September 14-16, 2005
Marriott Wardman Park Hotel, Washington, DC
Refers to the product development path from candidate selection to product launch
Covers drugs, biologics, and medical devices – but today’s talk is mostly about drugs / biologics
Initiative was announced publicly by Dr.
McClellan Tuesday, March 16, 2004
A serious attempt to bring attention & focus to the need for more scientific effort and publicly-available information on evaluative tools
Evaluative tools: The techniques & methodologies needed to evaluate the safety, efficacy & quality of pharmaceuticals as they move down the path
Despite Advances in Science, Success
Rate of Product Development has
NOT Improved
New compounds entering Phase I development today have 8% chance of reaching market, vs. 14% chance 15 years ago.
Phase III failure rate now reported to be 50%, vs. 20% in Phase III, 10 years ago.
Current applied science and infrastructure date from last century
Funding and progress in Development science has not kept pace with basic biomedical science.
Science to evaluate safety and efficacy of potential new medical products, and enable manufacture, is different from basic discovery science.
Need to fill gap in applied science-- to increase productivity and efficiency --to reduce cost of development process.
Overwhelming concurrence with:
recognition of science infrastructure problem
CP Initiative focus on research and collaboration,
We heard this from : drug industry, patient groups, device companies and groups, biotech companies, others
Docket Demand for FDA Action Exceeds
FDA Capacity: Far more proposed than FDA can undertake.
Principles for setting priorities for FDA actions are on Science Board agenda.
Clinical Trials
Biomarkers and Endpoints
Phase III trials are failing at a rate that is higher than expected - root causes ?
What is the typical planning process for drug development / phase 3 trials
What can we change; what new tools can we use, and what can we do better in the future to improve Phase III success and efficiency of drug development
Possible solutions / strategies
Can statisticians help ?
Are new study designs needed
Impetus for Adaptive designs, two stage designs, enriched target population designs
Are we planning correctly - Rethink how the study planning process occurs
It’s the dose
It’s the scenario needing better planning - or analysis methods
Bring consensus / closure to most pressing statistical issues at the core of decision making
Get involved in new emerging subject matter areas and impact them -genomics, proteonomics, nanotechnology
Broaden the multi-disciplinary roles, in industry, academia and regulatory bodies - internationally
Conduct Research , Gain Consensus, and
Develop Guidance to Remove Obstacles to
Efficient Drug Development and Enhance
Success Rates of Clinical Trials
Improve the Processes and Approaches to
Quantitative Analysis of Clinical Safety Data from Clinical Trials to Enhance Risk
Assessment and Management Initiatives
Improve the Statistical Understanding and
Application of Modern Statistical
Approaches to Product Testing and Process
Control
Clinical Trial Proposals for the Critical Path
Missing data due to patient withdrawals and dropouts in clinical trials
Flexible / adaptive clinical trial designs to improve the information and success rate of trials
Non-inferiority active control studies when placebos can't be used - getting to consensus on appropriate methods for margin setting, data analysis and interpretation for various data rich and data poor scenarios
Development of consensus on the statistical handling of multiple endpoints in clinical trials .
Clinical trial modeling and simulation as a tool for better design and interpretation of clinical trials
Application of Bayesian Methods to Enhance the Success
Rate of Clinical Trials
Guidance Development
Multiple endpoints
Non-inferiority
Topics of high interest
Adaptive / Flexible designs
Modeling / simulation / planning/Phase 2a
Other Critical Path needs: safety , product quality
Methods of application
Planning, data collection, statistical analysis plan
Process
Newly formed statistical safety team for more concentrated and focused advice
Earlier planning, modeling and simulation
Premarketing Risk Assessment
(Premarketing Guidance)
Development and Use of Risk
Minimization Action Plans (RiskMAP
Guidance)
Good Pharmacovigilance Practices and
Pharmacoepidemiology Assessment
(Pharmacovigilance Guidance)
Modern in process testing raises the possibility that alternatives to product quality should be considered
There have also been advances in Process
Analytical Technology (PAT) which depends on in process assessment of product quality all along the drug manufacturing process
The Non -Inferiority Problem
Current guidance is inadequate and the issues are poorly understood - must be fixed
Term introduced in ICH E9 ‘Statistical
Principles for Clinical Trials’
Some issues described in ICH E10 ‘Choice of Control Groups’
A study design that provides an indirect measure of evidence of efficacy / safety
To prove efficacy of test treatment by indirect inference from the active control treatment
To establish a similarity of effect to a known very effective therapy - e.g. anti-infectives
To infer that the test treatment would have been superior to an ‘imputed placebo’ ; ie. had a placebo group been included for comparison in the current trial. - a new and controversial area choice of margin is the key
To preserve a specified % effect of the AC
How is the margin “ “ chosen based upon prior study data
For a large treatment effect, it is easier - a clinical decision of how similar a response rate is needed to justify efficacy of a test treatment - e.g. anti-infectives is an example.
For modest and variable effects, it is more difficult ; and some approaches suggest margin selection based upon several objectives.
Complexities in choosing the margin
(how much of the control treatment effect to give up)
Margins can be chosen depending upon which of these questions is addressed:
how much of the treatment effect of the comparator can be preserved in order to indirectly conclude the test treatment is effective - a clinical decision for very large effects; a statistical problem for small and modest effects
how much of a treatment effect would one require for the test treatment to be superior to placebo , had a placebo been used in the current active control study - a lesser standard than the above
Do clinical trials of the comparator treatment consistently and reliably demonstrate a treatment effect - when they do not, what is the reason ?
Study is too small to detect the effect - under powered for a modest effect size
The treatment effect is variable, and the estimate of the magnitude will vary from study to study, sometimes with NO effect in a given study - a BIG problem for active controlled studies (Sensitivity to drug effect)
Importance of the assumption of constancy of the active control treatment effect derived from historical studies
It is relevant to the design and sample size of the current study, to the choice of the margin, to the amount of bias built into the comparisons, to the amount of effect size one can preserve (both of these are likely confounded), and to the statistical uncertainty of the conclusion.
Before one can decide on how much of the effect to preserve, one should estimate an effect size for which there is evidence of a consistent demonstration that effect size exists.
Four approaches to the problem
The simple case: specify a delta - not estimated
Indirect confidence interval comparisons (ICIC)
(CBER/FDA type method, etc.)
- thrombolytic agents in the treatment of acute MI
Virtual method (Hasselblad & Kong, Fisher, etc.)
- Clopidogrel, aspirin, placebo
Bayesian approach (Gould, Simon, etc.)
- treatment of unstable angina and non-Q wave MI
Current Guidance on Multiple
Endpoints is inadequate
Multiple primary endpoints
Multiple secondary endpoints
Composite endpoints
Multiple composites
Hierarchies
Patient reported outcomes
Decision Criteria for success
A collaborative effort: PhRMA 2004 meeting on coprimary endpoints, manuscript
Adaptive designs
Enrichment / pharmacogenomics
Sample size re-estimation
Design modification
Enriching trials with patients having genomic profiles likely to respond or less likely to experience toxicity
Goal of an adaptive / flexible design
Mid study changes that prospectively plan for modifications that preserve
Type 1 errors and maximize chances for success
Controversial
Statististical Methodology is Available
Why and where to use them?
Why the need for adaptation?
Design specifications often entail at least partial knowledge of the values of many planning (primary or nuisance) parameters that are unknown or at best might be guessed crudely
Sample size planning entails “educated” guess of effect size.
Selection of a composite endpoint requires
“educated” guess of where the potential effects lie and what noises may be.
Others…..
Hung
Addressing a process issue:
Scenario Planning:
A Tool to Increase the Success Rate of Phase III trials and to Enhance Drug Development Planning
Incorporates:
Several linked linked study phases - continuum
Multiple endpoints
Missing data
Use of all information in the process
Safety Planning
Modeling and simulation
Flexible designs / development sequence / international
Modern approach to protocol planning and choice of clinical study designs
Utilizing models for disease progression and endpoint selection
Utilizing simulation strategies for what if scenarios
Assumes input from other studies and planning efforts - planned sequences of studies may matter
An aid for prospectively planning integrated analyses
Disease Progression Modeling
Endpoint selection and evaluation
Trial Duration determination
Frequency and number of subject measurements
Tradeoffs between clinical endpoints and patient reported outcomes
Evaluate impact of missing data, informative treatment related censoring
Evaluate multiplicity implications
What would be observed if subjects had stayed in trial ?
Impute values from subects staying in longer
Test
Control
Which path do you choose ?
Higher is bad
1 2
Visit
3 4 5
What model captures the functional relationship of the disease progression and the clinical outcome(s) to be used to measure treatment effect
Can one function capture each of the clinical outcomes adequately
If not are several disease progression models used to express ‘response’
Different statistical tools and strategies
Challenge and explore assumptions
More multidisciplinary involvement
It is more than sample size planning
Structured planning meetings that are different that current – formal Q & A’s not broad enough
Links between phase planning and modeling efforts – currently too limited and stove piped
Concluding remarks
Multidisciplinary / collaborative planning and evaluation is needed now more than ever because issues becoming more complex - guidances can’t solve this - resources, exposure, experience, training will
Efforts to move available appropriate statistical methods and concepts , possibly more complex, into the main stream by emphasis on understanding by the audience appropriate to the application
Guidances don’t help here - need resources that can understand and communicate
Efforts to maximize contributions of industry, academic and regulatory statisticians
Choosing the most pressing needs and the chances for success - currently being updated
This is a national effort - not just FDA’s initiative - it will take a major coordinated effort to make progress