Regulatory Affairs and Adaptive Designs
Download
Report
Transcript Regulatory Affairs and Adaptive Designs
Regulatory Affairs and
Adaptive Designs
Greg Enas, PhD, RAC
Director, Endocrinology/Metabolism
US Regulatory Affairs
Eli Lilly and Company
Key Messages
• Adaptive designs open doors to realize the
promise of translational medicine advocated by
regulators
• At each stage of drug development, regulators
use a “lens” for their review of adaptive design
proposals, protocols, and study results
• Sponsors must not be shy about approaching
regulatory agencies with adaptive design
proposals, especially Phase 1 and 2.
• Sponsors should thoughtfully consider adaptive
design proposals as part of Phase 3, especially
in the context of larger confirmatory program.
Outline
• What’s important to regulators
• When and with what should sponsors
communicate with regulators
• New horizons for adaptive designs in the
regulatory context
Note: US FDA administrative focus but regulatory
science principles generalizeable
What’s Important to Regulators?
• Probability that the wrong dose or the wrong
drug is administered to the wrong patient at the
wrong time must be known to be acceptably low
• At each stage of drug development, regulators
use a “lens” for their review
–
–
–
–
Meaning of any claims or conclusions made
Relationship between claims and evidence
Quality of the evidence
Totality of the evidence (e.g. “substantial”)
Meaning of Claims
• Claims must be stated in terms of clearly defined
hypotheses that convincingly dominate other
plausible hypotheses
• Regulatory scientific paradigm for confirmation
of claims is constructed via hypothesis testing
framework
– H0: This agent should not be approved
H1: This agent should be approved
• Regulators adopted frequentist approach for
drugs and biopharmaceuticals; Bayesian
approach also acceptable for device approval
Relationship between Claims and
Evidence
• Clinical development program design
– Individual study designs
• Hypotheses to test and relationships to model
• Sample size, Type 1 and 2 error limits, effect size,
confidence interval width
• Primary and secondary endpoints
• Patient populations (inclusion, exclusion criteria)
• Post accrual adaptations within and between
patients (concomitant medications, rescue Rx, etc)
Relationship between Claims and
Evidence (continued)
• Designed to progressively discharge
uncertainty related to desired claims
– Start smaller / end bigger
• Enriched proof of concept, then generalize
– Start bigger / end smaller
• Explore heterogeneity, then target
– Biomarker and diagnostic test qualification
– Overall and subgroup exploration / testing
– Program-wise error rates (across studies)
Quality of the Evidence
• Randomization and Blinding
• Minimize unplanned changes post
randomization
• Certify that planned post randomization
changes do not bias inferences
• Transparency – tell regulators what you
plan to do, then do what you say you will
do and inform them that you did (or did
not) do it the way you planned
Totality of the Evidence
• Patient safety first ! Benefits (B) must outweigh
risks (R)
– If B<R then minimize exposure
– If B>R then maximize exposure (or must B>>R ?)
– If B = R then assume B<R
• Total patient exposures needed to estimate R
dictate how much needed when to estimate B
• Adaptive designs may be driven by formal utility
functions u(R, B) or other explicit considerations
of total evidence
• Substantial evidence of B>R for approval
Total Evidence can be generated
via Adaptive Designs
• Generally acceptable for early “learning”
phase development programs in Phase 1
and 2
• Discouraged in confirmatory Phase 3
setting unless
– exceptional circumstances (e.g. rare disease,
special populations)
– Additional (redundant?) fixed design studies
dominate the confirmatory development plan
Clinical Development Planning with
Regulators
• Meaning of any claims or conclusions made
• Target Product Profile (TPP)
• Relationship between claims and evidence
• Clinical development plan from “first in man” to approval
• Study protocols and simulations supporting error rate control
• Data Monitoring Committee charters, information control, and
sponsor decision-making plan, statistical analysis plans
• Quality of the evidence
• Locked database documentation
• DMC reports, information control results
• Totality of the evidence (e.g. “substantial”)
• Study reports/synopses, integrated summaries, etc.
Timing for Regulatory
Communication
• Communicate when you have the
evidence needed to support your plans
– As early as possible (e.g. Pre IND meetings or
written feedback)
– How much / how detailed depends on stage
of development
– If you want definitive regulatory action then
provide fairly definitive proposals
New Regulatory Horizons
• Exploratory INDs
• End of Phase 2A meetings
• Large, “simple”, “outcome” studies
– Pre-approval Risk Management
• Cardiovascular assessment of anti-diabetes drugs
– FDA Amendments Act of 2007
• Risk Evaluation and Mitigation Strategy (REMS)
– Observational
– Comparative effectiveness
Regulator’s Perspective
• Pro’s
– Risk and benefit information can drive safe dose
selection, patient populations to study, etc.
– Stratified medicine via appropriately targeted labeling
• Con’s
– Summary test statistics and other inferential tools
depend on a less certain and more complicated
sample space
– Chance of wrong decisions goes up
– Longer, more intensive reviews needed
Time to Act is Now
• Sponsors can develop adaptive design
proposals that align with the regulator’s
review lens at appropriate stages of drug
development
• Window of opportunity – don’t be shy
about proposing adaptive designs.
• Success in early phase development will
support enhanced use in confirmatory
phase development.