Clinical Outcome Assessment to Demonstrate Treatment Benefit: An

Clinical Outcome Assessment to
Demonstrate Treatment Benefit:
An FDA Perspective
Elektra Papadopoulos, MD, MPH
Acting Associate Director
Clinical Outcome Assessment Staff
Center for Drug Evaluation and Research (CDER), FDA
March 31, 2015
The views expressed in this presentation are
those of the speaker, and do not necessarily
represent an official FDA position.
Clinical Benefit
• Clinical benefit is demonstrated by evidence that
the treatment has a positive impact on how a
person with the condition or disease:
– Survives
– Feels or Functions in daily life
Types of Outcome Assessments
• Survival
• Clinical outcome assessments (COAs)
Performance outcomes (PerfOs)
Clinician-reported outcomes (ClinROs)
Observer-reported outcomes (ObsROs)
Patient-reported outcomes (PROs)
• Surrogates
– Often a biomarker* that is intended as a substitute for how a
patient feels, functions, or survives
– Two types for use in clinical trials to support product approval:
• Established Surrogates (for regular approval)
• Reasonably likely to predict clinical benefit (for accelerated
approval; require post-marketing studies to confirm clinical
*biomarker: a physiologic, pathologic, or anatomic characteristic that is objectively measured and
evaluated as an indicator of some normal or abnormal biologic function, process or response to a
therapeutic intervention
Selecting the COA type
Concept(s) of interest
Context of use
Consider appropriateness of COA type
Observable concepts
(e.g., signs, events, behaviors)
Unobservable concepts
(e.g., feelings)
No clinical
FDA’s PRO Guidance for Industry (2009)
• PRO is a measurement based on a report
that comes from the patient (i.e., study
subject) about the status of a patient’s health
condition without amendment or
interpretation of the patient’s report by a
clinician or anyone else.
•Describes good measurement principles
many of which are also applicable to other
types of clinical outcome assessment tools
•Provides an optimal approach to PRO
development; flexibility and judgment
needed to meet practical demands
Establishing Content Validity
• Begins after confirmation that the concept and the context of
use are appropriate and the type of outcome assessment has
been selected
• Evidence that the instrument measures the targeted concept in
the context of use
– And that the score represents the concept
– Supported by literature review, expert input and patient input
• Testing other measurement properties (e.g., test-retest
reliability, construct validity and ability to detect change) will
not replace or rectify problems with content validity.
V. Modify Instrument
• Identify a new COU
• Change wording of items, response options,
recall period, or mode/method of
administration/data collection
• Translate and culturally adapt
• Evaluate modifications using spokes I – IV
• Document all changes
Consider submitting to FDA for qualification
of new COA, as appropriate.
• Assess ability to detect change and construct validity
• Identify responder definition(s)
• Provide guidelines for interpretation of treatment benefit
and relationship to claim
• Document all results
• Update user manual
Submit to FDA for COA qualification as effectiveness
endpoint to support claims.
III. Cross-sectional Evaluation of Other Measurement Properties
• Outline hypothesized concepts and
potential claims
• Determine intended population
• Determine intended
application/characteristics (type of scores,
mode and frequency of administration)
• Perform literature/expert review
• Develop hypothesized conceptual
• Position COA within a preliminary
endpoint model
• Document COU and COI
IV. Longitudinal Evaluation of
Measurement Properties/
Interpretation Methods
I. Identify Context of Use (COU)
and Concept of Interest (COI)
Assess score reliability (test-retest or inter-rater) and construct validity
Establish administration procedures & training materials
Document measure development
Prepare user manual
Consider submitting to FDA for COA qualification for use in exploratory
studies prior to longitudinal evaluation.
II. Draft Instrument and Evaluate
Content Validity
Obtain patient or other reporter input
Generate new items
Select recall period, response options and format
Select mode/method of administration/data collection
Conduct cognitive interviewing
Pilot test draft instrument
Finalize instrument content, format and scoring rule
Document content validity
U.S. Food and Drug Administration
Center for Drug Evaluation and Research
Office of New Drugs
Measurement of Symptoms and
Patient-reported signs
• Content validity considerations
What are the core signs/symptoms of a condition?
Which are most bothersome, important to patients?
Which signs/symptoms are expected to improve with the therapy?
How severe are the signs/symptoms at study entry?
Variability: Are signs/symptoms heterogeneous across patients/
within patients over time?
– Do patients understand the questionnaire and respond as intended?
• Patient input in the form of qualitative research from the targeted
patient population important
Importance of defining the
context of use
• Patient characteristics
– Eligibility criteria: e.g., disease definition, other clinical
characteristics, baseline severity and age
– Language, cultural considerations
• Endpoint definition
– E.g., Symptom-free days; mean symptom severity over a period of a
week; time-to-worsening
• Analysis:
– E.g., Responder vs. comparison of mean change from baseline
Clinician-reported outcomes
• Content validity supported by clinician input, literature review
• Similar considerations to evaluation of content validity of PROs,
such as:
– Do rating clinicians interpret the instructions and items in the
instrument consistently and in the way intended?
– Do clinicians agree that the points on the rating scale correspond to
clinically meaningful gradations of severity in the concept of
– Are the appropriate aspects of the condition being captured as part
of the measure?
– What are the conditions, in which the instrument is appropriate for
• Note: clinician-reported signs and patient-reported symptoms
may not correlate highly
Pediatric Measurement
• Self-report of symptoms and impacts provides direct evidence
of treatment benefit and should be used when possible and
• Use verifiable report of observable concepts (e.g., signs,
behaviors) when self-report not possible or appropriate (e.g.,
young children)
• Example:
– A parent/caregiver should not be asked to rate unobservable
concepts such as pain
– A parent/caregiver can validly report on observable signs (e.g.,
crying, holding a body part and so forth)
Drug Development Tool Qualification Guidance
(Final January 2014)
• Qualification process described for
Biomarkers, Animal Models, and
Clinical Outcome Assessments (COA)
• COA qualification:
– a conclusion that within the
stated context of use, the results
of measurement can be relied
upon to represent a specific
concept (i.e., outcome) with a
specific interpretation when used
in drug development and
regulatory decision-making
Helpful Links
• FDA’s Patient-Reported Outcome (PRO) Guidance for Industry:
• DDT Clinical Outcome Assessment Qualification Program webpage:
• Includes Roadmap and Wheel and Spokes diagrams
• FDA’s DDT Qualification Program Guidance for Industry: