Quality by Design (QbD) –

Integration of Prior Knowledge

and Pharmaceutical Development
into CMC Submission and Review

Ajaz S. Hussain, Ph.D.

OPS/CDER/FDA
Outline
z What is QbD?
z Integration of prior
knowledge and
pharmaceutical
development
z Creating regulatory
flexibility via design space
What is “Quality by Design”?
z Quality
• “Good pharmaceutical quality represents an
acceptably low risk of failing to achieve the desired
clinical attributes.”
z Quality by Design (QbD)
• “Means that product and process performance
characteristics are scientifically designed to meet
specific objectives, not merely empirically derived from
performance of test batches.”

Janet Woodcock (2004)

To achieve QbD objectives..
z Product and process characteristics important to desired
performance must be derived from a combination of prior
knowledge and experimental assessment during product
development.

z From this knowledge and data, a multivariate model

linking product and process measurements and desired
attributes may be constructed.

z Clinical study would then be viewed as confirmatory

performance testing of the model.

Janet Woodcock (2004)

To achieve QbD objectives..
z The final link between the product and the
customer-driven quality attributes is the quality
system for manufacturing.
z Ideally, the quality system reflects and
addresses
• customer requirements
• ensures integration of product and process knowledge
gained during development
• ensures ongoing control of manufacturing processes,
and
• enables continuous improvement
Janet Woodcock (2004)
Quality by Design: Summary
z Design features and decisions that deliver a high
degree of confidence that manufactured lots will have
the intended product quality and performance
• Design to specifications (not setting specifications after the
fact)
• Redundant (i.e., an additional layer of security – risk based)
end product testing (not testing to document quality)
• An ability to describe and justify why proposed design
features deliver with confidence the intended quality (not
awaiting test results to be submitted in a post approval
supplement)
“….a combination of prior knowledge
and experimental assessment…”
z We know….
z We want to know…
z We learned …..
z We control …….
z We understand, agree & approve
z We expect a state of control
Benchmarking “Prior Knowledge” in
Current Regulatory Decisions
z Example: Post approval changes - testing and reporting
recommendations (SUPAC-IR, 1995)
• Scale-up factor 10X
• Up to 10X and greater than 10X ……
• “Significant body of data”… number of years of manufacturing
experience…
• 1 or up to 3 batches for stability
• In vitro to in vivo correlation established
• Waiver of in vivo BE
• Bioequivalence (BE) goal post of 80-125%
z (Popper) - Validating a piece of knowledge requires another
piece of knowledge
• regressio ad infinitum
• Stopped only with assumptions about the domain (conjectural)
We need to do better …
z Domain-specific assumptions
• Generally valid in precise contexts
• Assessment of validity requires domain specific
knowledge (e.g., for dosage forms - industrial
pharmacy/pharmaceutics)
• Poses significant challenge (to FDA) – recruiting, training, etc.
• Often becomes a check-box
• Domain specific assessment may not be ideal from a
public health perspective
• With increasing complexity a growing need for rigorous scientific
assessment of domain specific assumptions and practices
• Need’s a common scientific assessment framework – with
peer review to assure consistency
We can do better …
z Domain-specific assumptions
• Can be used to generate scientific hypothesis
• (Regulatory) Assessment can then be based on
established scientific principles
• Can serve as an engine for structured knowledge
generation, assimilation, verification and
generalization
• Continuous learning leading to continuous
improvement and regulatory flexibility
Clinical study would then be viewed as
confirmatory performance testing of the
model (?)

Clinical Safety &

Efficacy

Phase I “Bridging Studies” “To-be-Marketed”

Product/Process

Post-Approval
Changes
Data
Information
Continuous
Knowledge
Improvement
Pharmaceutical
Development
Prior Knowledge
 In Vivo BE* Bridging Studies: Justifying
Changes During Development (example
inspired by an actual case)

Capsule Tablet Film Site

(WG) Coat Change

Change BE Study
Solvent Site
in Drug
-Coat Change
Manuf. Failed BE

To-Be-Marketed
Tablet Multi- Multi- Approval
Scale-up Strength Strength
(DC)

*Generally 3-6 clinical bioequivalence tests are conducted in a NDA

From this knowledge and data, a multivariate model
linking product and process measurements and desired
attributes may be constructed (?)
Pharmaceutical Equivalent
Products
Reference Test
Possible Differences
Drug particle size, ..
Excipients
Normal healthy subjects
Manufacturing process
Crossover design
Equipment Overnight fast
Glass of water
Site of manufacture
90% CI within 80-125%
Batch size …. of Ref. (Cmax & AUC)

Bioequivalence =
Therapeutic Equivalence
Product and process characteristics important to desired
performance must be derived from a combination of prior
knowledge and experimental assessment during product
development (?)

z Prior knowledge guides product development

• Pre-formulation characterization For conventional

• Product development dosage forms

increasing opportunity

• Process development to leverage

prior knowledge

z Deliver a high degree of confidence that

manufactured lots will have the intended
product quality and performance
• Manufacturability, quality and performance
over the intended shelf-life
 Constructing a design space
“A battery of orthogonal characterization methods”

Phase I “Bridging Studies” “To-be-Marketed”

Product/Process

Post-Approval
Changes
Process understanding Data
Information
Continuous
Knowledge
Improvement
Pharmaceutical
Development
Raw Prior Knowledge
materials Facilities
Formulations Packaging
Unit operations
Design Space
z Multi-dimensional space that
encompasses combinations of product
design, manufacturing process design,
manufacturing process parameters and
raw material quality that provide
assurance of suitable quality and
performance
A Process is well understood when…

z All critical sources of variability are

identified and explained;

z Variability is managed by the process;

and,

z Product quality attributes can be

accurately and reliably predicted over
the design space …
The PAT Guidance
Depending on the degree of
complexity and uncertainty..
z The boundaries of a design space may be
defined by acceptable end product release test
• Solution dosage forms …
• Large process design space – as long as it conforms
with cGMP
z Moving towards control of in-process material
attributes provides a means to justify a large
process parameter design space
z For highly complex systems extensive
experimentation may be necessary to reduce
uncertainty and to justify a design space
Specifications - Standards and
Continuous Improvement
Process capability based
acceptance criteria

Incentive for continuous improvement?

Risk-based specifications

State of control

Action/alert limits

Real time release

Constructing and Justifying “Design
Space”

z Minimizing the potential for confusion

and incorrect interpretation due to
“minimal” & “optional” expectations
• Build on “minimal” expectations such as
stability, bioavailability, and other
performance assessment to “test of
hypothesis”
• Scientific risk assessment
• Opportunity to demonstrate the level of process
understanding and reliability of proposed “design
space”
FDA Proposal for Q8 Addendum. Brussels 2005
 Desired State

ACPS May 2005

2004, PAT & draft ICH Q8

1995 1997 2000

Current State
SUPAC Dissolution BCS
IVIVC ICH Q6A

ACPS May 2005

Regulatory Flexibility: Why?
z Facilitate continuous improvement and
innovation to improve quality, efficiency,
knowledge, and availability
z High level of process understanding and
control capability can further improve our
ability to ensure quality on every batch
• compared to a validated process with insufficient
understanding and for which “state of control” is
based primarily on compendial testing

CPG 7132c.08 The PAT Guidance

The ICH Q8 “Desired State”: Connecting to
Quality System (Q10)
z Product quality and z Develop effective CAPA – eliminate
performance achieved “special cause” variability
Utilize Process capability analysis –
and assured by design of z
reduce/control “common cause”
effective and efficient variability
manufacturing processes
z Identify, understand and acquire ability
to predict critical to quality attributes
z Product specifications (CQA)
based on mechanistic (product/process/measurement)
understanding of how z Focus on the “critical few”
formulation and process
factors impact product z Establish CQA target values and
acceptable variability around the target
performance value
z Utilize a monitoring system that
demonstrates “state of control”
z An ability to effect preferably based on critical material
continuous improvement attributes (not just end product testing)
and continuous "real
time" assurance of quality
Regulatory Flexibility: How?
z Quality by design
• Structured product and process development
• Process understanding and control capability
• Design space
z Integration of prior knowledge and pharmaceutical
development into C, M, C submission and review
• Present the knowledge gained to provide a more
comprehensive understanding of the product and
manufacturing process for reviewers and inspectors
• Risk based assessment and investigations and
knowledge sharing (over a product’s life cycle)
CPG 7132c.08 The PAT Guidance ICH Q8 (9 &10)
Thank you!
Hesitating to act because the whole
vision might not be achieved, or
because other do not yet share it, is
an attitude that only hinders progress.
- M. K. Gandhi