DEFINING QUALITY OF LIFE FOR ONCOLOGIC DRUG APPROVAL

1
DEFINING QUALITY OF LIFE FOR ONCOLOGIC DRUG
APPROVAL

• Carol M. Moinpour, PhD
• Southwest Oncology Group Statistical Center
• Fred Hutchinson Cancer Research Center
• Seattle, Washington/USA
• ODAC QUALITY OF LFE SUBCOMMITTEE
• February 10, 2000

2
What is Quality of Life?

• Subjective
• Patient perception is critical
• Health-related
• Yes, we are primarily measuring health status
• Multidimensional
• Which dimensions are relevant?
• Usually not driven by psychosocial theories

3
Proposal Data Source

• Expert re patient benefit patient
• Best equipped to evaluate claims about
  impact of treatment

4
Quality of Life Vs Health-Related QOL (HRQL)

• Not feasible to measure the myriad of non-medical
  influences
• Restrict QOL domains likely to be affected by
  medical intervention (HRQL)
• Attribution problem
• Randomization addresses unmeasured factors
• Combine quality with quantity of life?

5
Proposal QOL vs HRQL

• Restrict quality of life assessments to HRQL
• Patients asked to report current status
• Address known covariates
• HRQL can include duration of life
• A complicated question deserving our attention

6
HRQL Domains
7
Examples of Quality of Life Domains

• Physical Functioning Spirituality
• Occupational/Role Functioning Future Orientation
• Social Functioning Sexuality/Intimacy
• Emotional Functioning Health Concerns
• Symptom Status Family Well-Being
• Financial Concerns Satisfaction w/Care
• Global/Overall Perception of Quality of Life

8
HRQL Domains

• Inclusion of symptoms important because
• Corroborate physician-rated toxicities
• Document palliation in advanced stage disease
• Need to examine the reach of better/worse
  symptoms to general functioning

9
Assessment of Multiple HRQL Dimensions

• Specifies how treatments affect patients
• Complements physician-related toxicity
  information
• Informs patients and physicians about
  risks/benefits tradeoffs
• Identifies ways to improve cancer treatments
• Identifies survivors rehabilitation needs

10
Proposal HRQL Domains

• Three basic domains
• Psychological
• Physical
• Social
• Global HRQL vs summary score?
• Measure symptoms but also document their effect
  on basic domains

11
Proposal HRQL Domains

• Symptom status is not a manifestation of patient
  HRQL
• Sx outcomes alone should not be called HRQL
• Sx outcomes alone may be appropriate in Phase II,
  single institution, or supplemental submissions
• Sx outcomes can be designated primary
• Clinical issues should drive content of sx
  measures

12
Proposal Role of Theory

• Psychological or social science theories usually
  not driving HRQL assessment design
• Psychometric theory has done so
• First obligation identify HRQL issues critical
  to evaluating treatments
• Can suggest broad impacts on the patient
• Tx rationale usually doesnt address HRQL

13
Additional Slides
14
HRQL Research Issues

• Independence of HRQL dimensions
• Incorporation of importance ratings?
• Impact of substantial individual variation
  conclusions re patient benefit
• Integration of symptom and basic domain data
• To document scope of patient benefit/harm
• Validity of total scores with symptom component
• Role of methodological and theory-driven research
  in defining HRQL for drug claims?

15
Problem with Single Item Measures

• Only 1 chance to capture concept or dimension
• OK for detecting moderate to large differences
• Attractive for clinical trials (burden issue)
• Likely to miss differences at individual patient
  level
• Compromise - brief, multi-item scales

16
World Health Organization Definition

• Health is not only the absence of infirmity and
  disease but also a state of physical, mental and
• social well-being.