Critical Path Initiative: What it means for pharmaceutical industry statisticians

1

• Critical Path Initiative What it means for
  pharmaceutical industry statisticians

Walter Offen, Lilly Brenda Gaydos, Lilly José
Pinheiro, Novartis
2
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

3
Introduction

• March, 2004 FDA published Innovation-Stagnation
  Challenge and Opportunity on the Critical Path
  to New Medical Products
• http//www.fda.gov/oc/initiatives/criticalpath/whi
  tepaper.html (or ___ . pdf)
• Industry has been working on a number of the
  issues raised
• Strong synergies between FDA, academia/NIH, and
  industry are possible

4
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

5
Highlights of Critical Path Paper my
underlines

• the current medical product1 development path
  is becoming increasingly challenging,
  inefficient, and costly.
• costs of product development have soared over
  the last decade.

6
Highlights of Critical Path Paper my
underlines

• Not enough applied scientific work has been done
  to create new tools to get fundamentally better
  answers about how the safety and effectiveness of
  new products can be demonstrated, in faster time
  frames, with more certainty, and at lower costs.
• Finally, the path to market even for successful
  candidates is long, costly, and inefficient, due
  in large part to the current reliance on
  cumbersome assessment methods.

7
Highlights of Critical Path Paper

• The goal of critical path research is to develop
  new, publicly available scientific and technical
  tools -- including assays, standards, computer
  modeling techniques, biomarkers, and clinical
  trial endpoints -- that make the development
  process itself more efficient and effective and
  more likely to result in safe products that
  benefit patients.

8
Highlights of Critical Path Paper

• correlate early markers of safety and benefit
  with actual outcomes in patients.
• these new technologies could provide tools to
  detect safety problems early, identify patients
  likely to respond to therapy, and lead to new
  clinical endpoints.

9
Highlights of Critical Path Paper

• much more attention and creativity need to be
  applied to disease-specific trial design and
  endpoints intended to evaluate the effects of
  medical products.
• problems are often uncovered only during
  clinical trials or, occasionally, after
  marketing.

10
Highlights of Critical Path Paper

• Clinical testing, even if extensive, often fails
  to detect important safety problems, either
  because they are uncommon or because the tested
  population was not representative of eventual
  recipients. Conversely, some models create
  worrisome signals that may, in fact, not be
  predictive of a human safety problem.

11
Highlights of Critical Path Paper

• Adopting a new biomarker or surrogate endpoint
  for effectiveness standards can drive rapid
  clinical development. For example, FDA adoption
  of CD4 cell counts and, subsequently, measures of
  viral load as surrogate markers for anti-HIV drug
  approvals allowed the rapid clinical workup and
  approval of life-saving antiviral drugs

12
Highlights of Critical Path Paper

• FDA adoption of the eradication of H. pylori as
  a surrogate for duodenal ulcer healing greatly
  simplified the path of those therapies to the
  market.

13
Highlights of Critical Path Paper

• There are many important additional
  opportunities in the area of clinical trial
  design and analysis. More clinically relevant
  endpoints need to be developed for many diseases.
  Enrichment designs have the potential for
  providing much earlier assurance of drug
  activity. Bayesian approaches to analysis need to
  be further explored.

14
Highlights of Critical Path Paper

• This must be a joint effort involving the
  academic research community, industry, and
  scientists at the FDA

15
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

16
Overview of 8 PISC Initiatives

• PISC Pharmaceutical Innovation Steering
  Committee
• Improving Efficiency of Late-Stage Clinical
  Research (ECR) Walt Offen
• CTs have become increasingly large and expensive
  in recent years team hopes to identify means to
  improve efficiencies, including improved
  post-approval safety data collection and
  evaluation, study design improvements, and use of
  technology

17
Overview of 8 PISC Initiatives

• Novel Adaptive Clinical Trial Design Brenda
  Gaydos
• Collaborate with FDA, academia, and across the
  industry to develop accepted methodologies
  required to achieve development efficiency
  advantages
• Rolling Dose Studies José Pinheiro
• Develop and investigate dynamic CT designs with
  changing number of doses to efficiently and
  reliably characterize benefit/risk ratio of dose
  response.

18
Overview of 8 PISC Initiatives

1. Biomarker Working Group
2. Enriched patient population trial designs
3. Data Mining Tool Validation
4. Accelerated Proof of Concept
5. Predictive Models for Safety and Efficacy Working
   Group

19
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

20
ECR

• 3 key topics
• Obtaining sufficient safety data
• Data reduction and operational efficiency
• Efficient study design

21
ECR Safety Information

• Goal Increasing knowledge of safety while
  improving efficiency of late-stage clinical
  studies
• Rare serious AEs cannot be adequately assessed
  pre-approval
• Phase 3 duration and sample size cannot be
  sufficiently extended

22
ECR Safety Information

• Consider post-marketing LSSS (Large Simple Safety
  Study)
• Internet based study
• Relatively inexpensive, yet includes 10,000
  100,000 exposures or more
• Study of large prescribing database
• FDA Drug Safety and Risk Management Advisory
  Committee Meeting, May 18, 2005
• Issue Lack of control group, randomization

23
ECR Data Reduction and Operational Efficiency

• Lessen frequency of expensive procedures
• e.g. lab data, lab reference ranges
• Reduce study monitoring

24
ECR Data Reduction and Operational Efficiency

• Efficiency in Clinical Operations
• Electronic Data Capture (EDC)
• Standard database designs
• Standard statistical analysis programs/tables
• Internet-based trials
• Handheld devices

25
ECR Efficient Study Design Multiple
co-primary endpoints

• PhRMA Multiple Endpoints Expert Team (MEET) has
  researched this problem
• Position paper shared with FDA and submitted to
  DIJ
• Optimal solution is medical one reduce
  dimensionality to a single primary endpoint
• Choose one
• Create composite

26
ECR Efficient Study Design Multiple
co-primary endpoints

• Under complete null space, no upwards adjustment
  to nominal alpha levels is permissible
• Statistical adjustment under reasonable
  restricted null space is very modest
• Recent actions
• IMMPACT (Initiative on Methods, Measurement, and
  Pain Assessment in CTs) single primary endpoint
  for pain
• Migraine AC (Aug 4) 2-hour pain response is
  single primary endpoint

27
ECR Efficient Study Design Multiple
co-primary endpoints

• What about key secondary endpoints?
• Suggest moving away from gatekeeping strategy
• If academic, FDA, and industry scientists/experts
  can agree on a set of key secondary endpoints
  that help define and describe the disease,
  then..
• All of these should be summarized in Clinical
  Studies Section of product labeling ( or -)
• Helps address personalized medicine

28
ECR Efficient Study Design Non-inferiority
designs

• Margin selection
• The following two paradigms lead to vastly
  differing size of study
• Indirect demonstration of superiority to placebo
  (had a placebo group been in the trial)
• Preservation of a certain fraction of the active
  control's effect

29
ECR Efficient Study Design Flexible dosing

• In such a design, patients and/or physicians are
  allowed to alter dose based on response
• Alternative to searching for the single dose
  which is right for all patients
• Diseases where response can be assessed in short
  period of time are candidate
• e.g., migraine, acute and chronic pain

30
ECR Efficient Study Design Flexible dosing

• Comparisons between dose groups is problematic
• Can summarize patients receiving available
  doses
• Separation of drug and placebo groups is
  maximized
• Mimics clinical practice

31
ECR Efficient Study Design Enrichment Designs

• Biomarkers, ____-omics, or other attribute can
  lead to restriction of study population to those
  most likely to respond to study drug
• Two drugs may be identical on an average basis,
  but one might be best for one subpopulation, the
  other for another subpopulation
• If cant predict going into the study, a
  crossover design might be a candidate to evaluate
  this aspect

32
ECR Efficient Study Design Additional Topics

• Surrogate endpoints
• Categorization of continuous data
• Appropriate methods for handling missing data
  Mixed Models Repeated Measures (MMRM) vs. LOCF
• Lieberman et al (Neuropsychopharmacology 2005 30,
  pp 445-460 Contemporary approaches to handling
  missing data (Mallinckrodt et al, 2003, 2001)
  Entsuah, 1996) are highly preferable.
• Pure ITT

33
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

34
Adaptive Designs Opportunity

• Improve quality, speed and efficiency of
  decision making within clinical development
• Bring more winners onto market quickly
• Discard losers early
• Shift towards a more seamless integrated approach
  to clinical drug development
• Optimize patient treatment within a trial
• Maximize patient exposure to doses/drugs that
  work
• Minimize patient exposure to doses/drugs that
  dont work

35
Adaptive Designs What

• Any design which uses accumulating data to modify
  aspects of the trial
• Adaptations can include
• Sample size (stopping early, increasing sample
  size)
• Treatment allocation ratios
• Dose / Treatment arms (dropping, adding arms)
• Adapting hypothesis (primary objective, primary
  endpoint)
• Patient population (entry criteria)
• Observational scheme
• Test statistics
• Stages of the experiment (e.g. seamless phase
  II/III)
• Dynamic randomization based on baseline covariates

36
Adaptive Designs How

• Facilitate understanding and implementation of
  adaptive designs through the deliverables of the
  working group
• Adaptive designs not yet routinely used
• Perception that there might be regulatory
  concerns
• Additional time/upfront investment required to
  design/implement non-standard designs
• Lack of internal/external buy-in to concept
• Lack of infrastructure for timely data collection
  and data analysis
• Lack of training and experience in best practices
  for adaptive design methods

37
Adaptive Design Focused Topics

• Rationale when to adapt
• Definition and classification of adaptive designs
• Dose-response finding
• Seamless phase II/III
• Implementation issues
• Sample size re-estimation
• Case studies

38
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

39
Rolling Dose Studies Why?

• Poor understanding of dose response (efficacy and
  safety) of drugs plagues clinical development
• Indicated by both FDA and Industry as one of
  leading causes of late phase attrition and
  post-marketing problems with approved drugs
• Current designs and methods for dose finding
  focus on selection of MED out of fixed, generally
  small number of doses, via hypothesis testing ?
  inefficient

40
Rolling Dose Studies What?

• Flexible designs for investigating dose response,
  allowing dynamic allocation of patients to a
  larger, possibly variable number of doses
• Main goal efficiently learn about dose response
  profiles for efficacy and safety to characterize
  benefit/risk over dose range
• Better, faster decision making on dose selection
  and improved labeling
• Emphasis on modeling and estimation, as opposed
  to hypothesis testing

41
Rolling Dose Studies How?

• Identify and investigate existing designs and
  methods for flexible dose finding
• Adapt current methods and develop new ones to
  create suite of designs and statistical methods
  for efficient dose response learning under
  various CT scenarios (e.g., availability of
  biomarker, single drug or combination)
• Evaluate potential benefits over traditional
  designs to make recommendations on practical
  usefulness of rolling dose studies

42
Outline

• Introduction
• Highlights of FDAs Critical Path Paper
• PhRMA initiatives to address Critical Path
• Overview of 8 PISC Initiatives
• Improving Efficiency of Late-Stage Clinical
  Research (ECR)
• Adaptive Designs
• Rolling Dose Studies
• Biomarkers

43
Biomarkers

• Some are useful for predicting in early phase
  clinical research which drugs will be
  successfully approved for marketing
• Some help identify the right patient (targeted
  therapeutics)
• May or may not be ____-omics based
• Some may be elevated to become a surrogate marker

44
(No Transcript)
45
Summary Comments

• FDAs Critical Path white paper has opened the
  door to exciting opportunities for improving the
  current drug development paradigm
• All opportunities are in need of statistical
  input and direction