Clinical Trials for Quality of Life Endpoints in Oncology

1
Clinical Trials for Quality of Life Endpoints
in Oncology
Jeff A. Sloan, Ph.D. Mayo Clinic, Rochester,
MN, USA

• Oncology Education Session
• Rochester, November 1, 2005

2
QOL challenges

• Reliability if I were to use this tool under the
  same conditions would I get the same results?
• Validity am I measuring what I want to measure?
• Missing data imputation, design considerations
• Response shift hospice patients at 75

3
Take home messagethere is good news

• There are problems with using QOL assessments as
  indicators of efficacy in clinical trials.
• There are scientifically sound solutions to these
  problems. The problems have been disseminated
  widely and consistently. The solutions have not.

4
Checklist for designing, conducting and reporting
HRQL - PRO in clinical trialsPatient Reported
Outcomes (PRO) and Regulatory Issues A European
Guidance Document for the improved integration of
health-related quality of life assessment in the
drug regulatory process. Chassany O et ERIQA
Working Group. Drug Information Journal 2002.

• HRQL / PRO objectives
• Added value of HRQL / PRO
• Choice of the questionnaires
• Hypotheses of HRQL / PRO changes
• Study design
• Basic principles of RCT fulfilled ?
• Timing and frequency of assessment
• Mode and site of administration...
• HRQL / PRO measure
• Description of the measure (items, domains)
• Evidence of validity
• Evidence of cultural adaptation

• Statistical analysis plan
• Primary or secondary endpoint
• Superiority or equivalence trial
• Sample size
• ITT, type I error, missing data
• Reporting of results
• Participation rate, data completeness
• Distribution of HRQL / PRO scores
• Interpreting the results
• Effect size,
• Minimal Clinically Important Difference
• Comparison with other criteria / scores
• Number needed to treat

5
EMEA RECOMMENDATIONS Points to consider
(CPMP/EWP/562/98) on clinical investigation of
medicinal products in the chronic treatment of
patients with COPD, 1999

• In the major efficacy studies of symptomatic
  benefit the primary endpoint should reflect the
  clinical benefit the applicant wishes to claim in
  the future SPC (Summary of Product
  Characteristics)
• It should include the FEV1 as a measure of lung
  function and include a measure of symptomatic
  benefit
• A significant benefit for both endpoints, should
  be demonstrated so that no multiplicity
  adjustment to significance levels is indicated
• The primary symptomatic benefit endpoint should
  be justified by referencing published data which
  supports its validity one example is the St
  Georges Respiratory Questionnaire
• There are number of secondary endpoints which may
  provide useful information. e.g. symptom
  scales, and quality of life assessment

6
QOL The big picture
7
Symptoms and QOLIs there a difference?

• If you count the number of emetic episodes, you
  are assessing a symptom
• If you ask the patient how bad their nausea is,
  you are assessing QOL
• The measurement issues and analytical procedures
  are the same
• Literature is converging to the term
  patient-reported outcomes (PRO)

8
Developmental Timeline of Commonly Used QOL
Measurement Tools
9
What is an Appropriate QOL Instrument?

• Research objective (HYPOTHESIS DRIVEN)
• Specific rationale for the QOL part of the study
• Relevant domains of QOL (LIST MATCH)
• Disease and patient population characteristics
• Psychometric characteristics (reliability
  validity) of QOL instrument
• Practical considerations (e.g. respondent burden,
  language translations)

10
Timing of QOL Assessment

• Study objective
• Characteristics and natural course of disease
• Baseline and one follow-up QOL assessment are
  necessary
• Treatment regimen
• Similar timing of QOL assessment across treatment
  arms
• Expected effects of the treatment

11
QOL Research Themes

• 1. Assess QOL within clinical trials with
  efficiency, consistency, specificity
• 2. Improve QOL methodology
• 3. Develop intervention studies targeted at QOL
  endpoints

12
QOL in NCCTG Clinical Trials

• Since 1995, 84 trials with QOL component
• gt50 different QOL questionnaires used
• gt20 papers per year published with QOL
• Average baseline compliance rate 94

13
What underlies these QOL metrics?

• NCCTG does not experience the problems that
  other groups report with respect to QOL.
• Efforts to make the inclusion of QOL components
  in treatment trials easy and efficient have been
  well received by investigators. (Integrating
  cancer control research into the CCOP network a
  case study of the NCCTG, NCI, 2004)

14
QOL Team Resources

• MD tumor group liaisons
• Operations manual
• Forms bank
• Literature bank
• Background templates
• Web-based utilities

15
Cancer Patient Assessment

• Cancer patient assessment involves tumor growth
  and survival data.
• We measure these scientifically and the effect of
  interventions on these endpoints.
• Cancer also involves other things besides tumors
  and reduced lifespan that can be measured..

16
by answering scientific questions

• What is the value added of loooooong QOL
  assessments to treatment trials?
• What is the evidence for the use of single-item
  QOL assessments?
• How do you deal with multiple endpoints?
• How do you handle missing data?
• What is the clinical significance of QOL
  assessments?

17
What is the value added of additional questions?
18
Less is More

• Numerous studies indicate shorter assessments are
  just as good as longer assessments
• Bernhard. single item quality of life indicators
  in cancer clinical trials. Brit J Cancer
  84(9)1156-1165, 2002
• Vickers. Contolled Clinical Trials, 24 731
  735, 2003
• Abdel-Khalek. Measuring anxiety. Death Studies
  22(8)763-772, 1998
• Gardner. Ed Psych Measurement 58(6)898-915,
  1998
• Sloan. Overall QOL. JCO 163662-3673, 1998
• Sloan. Clinical significance of single items
  relative to summated scores. Mayo Clinic Proc
  77 479-487, 2002

19
Sloan et al, Biopharm Stat 14(1) 73-96, 2004.
20
Single-Item or Multiple-Item PRO?
Sloan et al, Mayo Clinic Proc 77 479-487, 2002.
21
A Comparison of Simple Single-Item Measures and
the Common Toxicity Criteria in Detecting the
Onset of Oxaliplatin-Induced Peripheral
Neuropathy in Patients with Colorectal Cancer
R. F. Morton, J. A. Sloan, A. Grothey, D. J.
Sargent, H. McLeod, E. M. Green, C. Fuchs, R. K.
Ramanathan, S. K. Williamson, R. M. Goldberg
ASCO 2005
22
Background

• Peripheral neuropathy (PN) is common during
  treatment with Oxaliplatin
• Assessment of PN is historically done via the
  Common Toxicity Criteria (CTC)
• We developed a single-item numerical analogue
  scale assessment to help measure PN
• We compared the two measures to look at the
  sensitivity of the CTC in detecting the onset of
  PN

23
Methods

• 696 patients randomized to FOLFOX4
• PN assessed bi-weekly during treatment
• NAS filled out at baseline and every 12 weeks
  during treatment

24
NCCTG/Intergroup Trial N9741
IFL Irinotecan 5-FU/LV
RANDOMI ZAT ION
FOLFOX4 Oxaliplatin 5-FU/LV
IROX Irinotecan Oxaliplatin
Goldberg et al, JCO 2004
25
NAS Tools
26
An Empirical Anomaly

• According to CTC only 20 of patients experienced
  serious PN
• Clinical knowledge suggested the incidence rate
  should be much higher (about 80)

27
Agreement
2 Point Change in QOL 2 Point Change in QOL
No (N420) Yes (N276) Agreement Kappa Statistic
Grade 2 PN No (N440) 308 132 65 0.25
Grade 2 PN Yes (N256) 112 144 65 0.25
Grade 3 PN No (N597) 380 217 63 0.13
Grade 3 PN Yes (N99) 40 59 63 0.13
The agreement of lt 65 indicates CTC and NAS
measure different aspects of PN.
28
Dose to PN CTC versus NAS Which Comes First?
Median dose to NAS CSD of 424 mg/m2 versus 765 (
961) mg/m2 for CTC grade 2 (3) event
29
Time to PN CTC versus NAS Which Comes First?
Patients notice an increase in PN two or three
months earlier via the NAS
30
Conclusions

• Grade 2 PN is found to be a significant problem
  according to the NAS
• Using CTC, PN is under-reported
• NAS may allow for earlier detection
• NAS should be used in conjunction with CTC

31
Are the occurrence of adverse events and
clinically significant changes in symptom
specific and global quality of life measures
predictable?

• Sumithra J. Mandrekar, Ph.D.
• Mashele M. Huschka, B.S.
• James R. Jett, M.D.
• Jeff A. Sloan, Ph.D.
• Mayo Clinic
• Rochester, MN

32
NCCTG Lung Cancer Trials
Study Number Description Sample Size Assessments Assessment Schedule
95-20-53 A Pilot Study of High-Dose Thoracic Radiation Therapy w/ Concomitant Cisplatin/Etoposide in Limited-Stage SCLC 76 Uniscale LCSS Baseline, prior to irradiation, prior to last cycle and at 3 months, 1 year 2 year follow-up visits
95-24-52 A Phase II Trial of Edatrexate in Combo w/ Vinblastine, Adriamycin, Cisplatin Filgrastim in Pts w/ Advanced NSCLC 34 Uniscale FACT-L v3 Baseline and prior to each treatment cycle
97-24-51 Phase III Randomized, Double-Blind Study of CAI Placebo w/ Advanced NSCLC 177 Uniscale FACT-L v4 Baseline and monthly during course of treatment
98-24-52 Randomized Phase II Study of Docetaxel Gemcitabine for Stage IIIB/IV NSCLC 99 Uniscale LCCS Baseline and prior to each treatment cycle
N0021 Phase II Study of Gemcitabine and Epirubicin for the Treatment of Mesothelioma 68 Uniscale SDS Baseline, at each evaluation and 3 months 1 year follow-up visits
N0022 Oral Vinorelbine For the Treatment of Metastatic Non-Small Cell Lung Cancer in Patients gt 65 Years of Age A Phase II Trial of Efficacy, Toxicity, and Patients' Perceived Preference for Oral Therapy 58 Uniscale LCSS Baseline and immediately after completion of second cycle of chemotherapy
33
QOL Assessments

• Spitzers Uniscale
• 1 question for the global assessment of QOL
• Functional Assessment of Cancer Therapy Lung
  (FACT-L)
• 27 questions divided into 4 well-being
  constructs physical, social/family, emotional,
  and functional
• 10 questions specific to lung cancer
• Lung Cancer Symptom Scale (LCSS)
• 9 questions pertaining to lung cancer symptoms
• Symptom Distress Scale (SDS)
• 12 questions related to symptoms commonly
  experienced by cancer patients

34
Determine the relationship of a single-item
assessment with the multiple-item summated scales
35
Post-Baseline QOL
FACT-L Uniscale (N148) LCSS Uniscale (N164) SDS Uniscale (N46) Total (N358)
Uniscale        
Mean (SD) 71.1 (19.13) 68.6 (25.44) 65.5 (23.05) 69.6 (22.31)
    Median 75.0 76.5 68.5 75.0
    Range (0.0-97.0) (1.0-100.0) (4.0-97.0) (0.0-100.0)

Multiple-items
    Mean (SD) 74.9 (12.26) 72.0 (16.26) 73.9 (14.53) 73.6 (14.22)
    Median 75.7 74.0 77.9 75.2
    Range (30.7-99.3) (0.0-99.3) (38.5-96.2) (0.0-99.3)
Spearman Rank Correlations between the Uniscale
and the FACT-L, LCSS, and SDS were 0.66, 0.57,
and 0.49 respectively
36
When QOL is high Uniscale gt LCSS When QOL is
low Uniscale lt LCSS ? Greater variability in
Uniscale Scores
Correlation0.43
37
Determine if clinically significant declines are
more readily detected by a single-item or
multiple-item assessment
38
Individual Patient Data over time Greater
variability in Uniscale Scores
39
Clinically Significant Decline (CSD) 10-point
decline on a 0-100 scale
FACT-L Uniscale (N120) LCSS Uniscale (N152) SDS Uniscale (N45) Total (N317)
Uniscale      
n () 73 (60.8) 91 (59.9) 20 (44.4) 184 (58.0)

Multiple-items
n () 46 (38.3) 66 (43.4) 13 (28.9) 125 (39.4)

Percent Agreement 56 59 71 59
Represents the number of patients that completed the Uniscale at baseline and at least once post-baseline and completed a multiple-item assessment at baseline and at least once post-baseline Represents the number of patients that completed the Uniscale at baseline and at least once post-baseline and completed a multiple-item assessment at baseline and at least once post-baseline Represents the number of patients that completed the Uniscale at baseline and at least once post-baseline and completed a multiple-item assessment at baseline and at least once post-baseline Represents the number of patients that completed the Uniscale at baseline and at least once post-baseline and completed a multiple-item assessment at baseline and at least once post-baseline Represents the number of patients that completed the Uniscale at baseline and at least once post-baseline and completed a multiple-item assessment at baseline and at least once post-baseline

• Uniscale more likely to detect a CSD in QOL than
  the multiple-item assessments (58 vs. 39)
• The overall percent agreement in detecting a CSD
  in QOL between Uniscale and multiple-item
  assessments was 59

40
Determine how single-item assessment and
multiple-item summated scales relate to adverse
events data
41
Adverse Events (AE)

• Severe adverse event is defined as a grade 3, 4,
  or 5, regardless of attribution
• 33 experienced a severe AE post baseline
• Nine AEs experienced by at least 2 of the
  population that can also be collected via a QOL
  instrument
• Alopecia, Anorexia, Constipation, Diarrhea,
  Dyspnea, Fatigue, Nausea, Neurosensory, Vomiting
• 95 experienced at least one of the nine AEs
• 20 had at least one of the nine graded as severe
• CSD in AE is defined as a baseline AE of grade 0,
  1, or 2 that changes to a grade 3, 4, or 5 post
  baseline

42
Severe AE and CSD in QOL
FACT-L Uniscale LCSS Uniscale SDS Uniscale Total
Uniscale        
Number evaluable 122 155 46 323
Severe AE 26 (21.3) 74 (47.7) 17 (37.0) 117 (36.2)
    CSD in QOL 74 (60.7) 92 (59.3) 20 (43.5) 186 (57.6)
Percent agreement 46 51 46 48

Multiple-items
Number evaluable 140 156 45 341
    Severe AE 30 (21.4) 76 (48.7) 17 (37.8) 123 (36.1)
    CSD in QOL 52 (37.1) 67 (43.0) 13 (28.9) 132 (38.7)
Percent agreement 64 53 60 59
Represents the number of patients that had an adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had an adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had an adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had an adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had an adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline
43
CSD in AE and CSD in QOL
Alopecia Anorexia Constipation Diarrhea
Uniscale        
Number evaluable 139 94 69 73
CSD in AE 2 (1.4) 8 (8.5) 6 (8.7) 13 (17.8)
    CSD in QOL 75 (54.0) 61 (64.9) 37 (53.6) 44 (60.2)
Percent agreement 46 37 52 41

Multiple-items
Number evaluable 145 99 72 77
    CSD in AE 2 (1.4) 9 (9.1) 6 (8.3) 15 (19.5)
    CSD in QOL 59 (40.7) 50 (50.5) 17 (23.6) 31 (40.3)
Percent agreement 59 44 74 53
Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline
44
CSD in AE and CSD in QOL
Dyspnea Fatigue Nausea Neuro-sensory Vomiting
Uniscale        
Number evaluable 155 226 208 189 142
CSD in AE 43 (27.7) 42 (18.6) 34 (16.4) 9 (4.8) 23 (16.2)
    CSD in QOL 90 (58.1) 139 (61.5) 118 (56.7) 116 (61.4) 72 (50.7)
Percent agreement 46 42 42 39 50

Multiple-items
Number evaluable 159 236 216 202 150
    CSD in AE 43 (27.0) 45 (19.1) 33 (15.3) 11 (5.5) 23 (15.3)
    CSD in QOL 67 (42.1) 96 (40.1) 84 (38.9) 73 (36.1) 60 (40.0)
Percent agreement 57 59 58 64 63
Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline Represents the number of patients that had a baseline and post-baseline adverse event (any grade) and completed a QOL assessment at baseline and at least once post-baseline
45
(No Transcript)
46
70
12
47
83.7

• 6 events reported via CTC
• 25 CSD reported via SDS

25.3
48
Summary

• Uniscale demonstrates greater variability than
  the multiple-item indices
• The Uniscale is better able to detect a CSD in
  QOL than the multiple item assessments, and
  captures a CSD earlier than the multiple item
  assessments
• Correlations and percent agreement between
  Uniscale and multiple-item assessments were modest

49
Summary

• There is indication that a CSD in QOL occurs
  earlier than CTC AE reporting
• Consistent with a recent finding that single-item
  QOL assessments detect a patient-perceived
  problem in peripheral neuropathy more than six
  weeks earlier than CTC (Morton et al, ASCO 2005)
• The multiple-item assessments are in better
  agreement with occurrence or CSD in AE compared
  to the Uniscale

50
What is the evidence for the use of simple
(single-item) LASAs?
51
The literature for simple assessments is
considerable

• Grunberg S.M. (1996). Comparison of conditional
  quality of life terminology and visual analogue
  scale measurements. Quality of Life Research 5
  65-72.
• Gudex C. (1996). Health state valuations from
  the general public using the Visual Analogue
  Scale. Quality of Life Research, 5 521-531.
• Hyland ME. Development of a new type of global
  quality of life scale and comparison and
  preference for 12 global scales. Quality of Life
  Research. 5(5) 469-480. 1996.
• Sriwatanakul, K. (1983). Studies with different
  types of visual analog scales for measurement of
  pain Clinical Pharmacology and Therapeutics
  34(2) 234-239.
• Wewers ME. (1990). A Critical Review of Visual
  Analogue Scales in the Measurement of Clinical
  Phenomena. Research in Nursing Health, 13
  227-236.
• Bretscher M. (1999). Quality of Life in Hospice
  Patients A Pilot Study, Psychosomatics, 40,
  309-313.

52
The Visual Analogue Uniscale

• Please mark with an X the appropriate place
  within the bar to indicate your rating of this
  persons quality of life during the past week.
• Lowest quality applies to someone completely
  dependent physically on others, seriously
  impaired mentally, unaware of surroundings, and
  in a hopeless position.
• Highest quality applies to someone physically and
  mentally independent, communicating well with
  others, able to do most of the things enjoyed,
  pulling own weight, with a hopeful yet realistic
  attitude.

Highest Quality
Lowest Quality
(Please mark one X within the bar)
53
Uniscale-NAS(Numeric Analog Scale)
54
Linear Analogue Self Assessment (LASA)

• General measure of global QOL dimensional
  constructs
• Overall QOL Uniscale question plus domain
  specific questions
• LASA 6 questions
• covering domains QOL, Mental, Social, Spiritual,
  Emotional, Physical
• e.g. How would you describe your overall
  physical well-being during the past week,
  including today? (0 as bad as it can
  be 10 as good as it can be)
• LASA additional items (any understandable
  construct)
• e.g. How would you describe your anxiety during
  the past week, including today?
• (0 anxiety as bad as it can be 10 no
  anxiety)

55
LASA Validity Data

• Median split adds 3 months to median survival in
  advanced cancer patients (Sloan, JCO, 1998)
• Qualitative study score of 5 or less indicates
  need for intervention (Frost, unpublished)
• Stable populations average roughly 7, with SD
  roughly 2 on 10-point scale (20 on 100 pointt
  scale) (Locke, in preparation)

56
LASA Norms (Various)

• Hospice patients 7.6
• Advanced cancer patients 7.2
• Recovering surgical patients 6.6
• Healthy volunteers 8.2
• Medical students 4.4

57
A Structured Multidisciplinary Psychosocial
Intervention Improves the Quality of Life of
Patients with Advanced Stage Cancer
T Rummans, M Clark, J Sloan, M Frost, P Atherton,
M Bostwick, G Gamble, M Johnson, J
Richardson Mayo Clinic, Rochester, MN

• In press, JCO

58
Background

• Some studies have suggested a psychosocial
  intervention has a positive effect on survival,
  while others have not demonstrated such an effect
  or suggested a negative effect on survival.
  (Spiegel, 1990 Goodwin NEJM 2001 Spiegel,
  Cancer, 2002)
• Most interventions are single - focus and have
  targeted mood (Fawzy, AGP,1993 Jacobsen JCO
  2002 Kolden, Psycho-Onc. 2002)

59
Motivation for the present study

• A multidisciplinary intervention had not been
  tried nor tested for feasibility
• Overall QOL is the composite, multidimensional
  psychosocial target

60
Study Schema
Arm A Structured multi-disciplinary psychosocial
intervention. 8 - 90 minute sessions over 4 weeks
Patients with Advanced Stage Disease scheduled to
undergo radiation therapy
R
Arm B Standard Care
QOL assessed at baseline and week 4
(EOT) Stratification tumor type, ECOG PS, age
61
Secondary endpoint assessment tools
Linear Analogue Self Assessment (LASA) items
Profile of Mood States Short Form
(POMS) Symptom Distress Scale (SDS) FACIT
Spiritual Well-Being
62
Which is the real symptom endpoint?
63
Primary Result Overall QOL at 4 weeks
At week 4, overall QOL was 10 points higher in
the intervention arm than in the standard care
arm (80 versus 70 on the 100-point scale
respectively, p0.047). The treatment group
improved 3.3 points from baseline, while the
control group decreased 8.9 points on average,
p0.009. More than three times as many patients
in the treatment group reported a 10-point
improvement in QOL from baseline compared to the
control group (30 versus 9, p0.004).
64

65
best
worst
66
(No Transcript)