EURURO Vol. 71 No. 4

536 692

point for assessing treatment differences was at 61 wk. For both models,
data up to 61 wk were used to remove bias introduced by high patient
dropout beyond this point (<15% of bicalutamide patients with available
data beyond 61 wk).
Time to first HRQoL deterioration or pain progression was calculated
from the date of randomisation to first HRQoL deterioration or pain
progression. The treatment effect of enzalutamide versus bicalutamide,
based on time to HRQoL deterioration or pain progression, was tested
with an unstratified log-rank test, and Kaplan-Meier curves were used to
estimate distribution. The hazard ratio (HR; enzalutamide/bicalutamide)
and 95% confidence intervals (CI) were determined using Cox
proportional-hazards model with treatment as the only covariate.
HRQoL in patients who discontinued treatment (and HRQoL
assessments) due to progressive disease may not be representative of
patients who continued HRQoL assessments without progressive
disease. Patients with progressive disease may be more likely to
experience HRQoL deterioration and may represent informative
censoring. Therefore, sensitivity analyses of HRQoL deterioration using
a composite endpoint were performed, where progressive disease and
HRQoL deterioration were considered an event.
Minimal clinically important differences (MCIDs)—patient-derived
scores reflecting intervention-induced changes that are meaningful for
the patient were defined for each instrument . The
lower limit of the MCID range was used to interpret change from
baseline analysis between and within treatments, whereas the upper
limit was used for deterioration analysis. For threshold values for
deterioration or pain progression used in sensitivity analyses, see
Supplementary Table 1.
Table 1 – Patient-reported outcome instruments used: description, number of items, theoretical range, minimal clinical important
difference (MCID) range, and threshold values required to meet the definition of deterioration or progression in the respective analyses of
this study
Outcome No. of
items
Theoretical
range
MCID range Change
threshold
Description
FACT-P
Physical well-being 7 0–28 2–3
Deterioration of
3 points from baseline
FACT-P is a 39-item questionnaire with five
subscales: emotional well-being; functional
well-being; physical well-being; social
well-being; and the PCS three
additional indices (Trial Outcome Index
, FAPSI-8 , and PCS pain-related
score ) were also evaluated. Items were
rated on a 5-point Likert scale (0–4). Higher
scores indicate better HRQoL.
Functional well-being 7 0–28 2–3
Emotional well-being 6 0–24 2–3
Social well-being 7 0–28 2–3
PCS 12 0–48 2–3
FAPSI-8 8 0–32 2–3
PCS pain-related score 4 0–16 1–2
Deterioration of
2 points from baseline
Trial Outcome Index 26 0–104 5–9 Deterioration of
9 points from baseline
FACT-G total score 27 0–108 3–7 Deterioration of
7 points from baseline
FACT-P total score 39 0–156 6–10 Deterioration of
10 points from
baseline
EQ-5D
EQ-5D utility index 5 –0.594 to 1 0.04–0.14 Deterioration of
0.12 points from
baseline
EQ-5D is a generic instrument comprising a
VAS (which rates health states from
0 [worst imaginable] to 100 [best
imaginable]) and a self-reported
questionnaire (with five dimensions
[mobility, self-care, usual activities, pain/
discomfort, and anxiety/depression]
combined to give a unique health state for
each patient) . Health states are
converted to an index-based summary
score; we used the preference-based
algorithm from the UK general population
to derive the EQ-5D index Higher
scores indicate better HRQoL.
EQ-5D VAS 1 0–100 7–11 Deterioration of
11 points from
baseline
BPI-SF
Item 3: pain at its worst 1 0–10
Increase 30% or
2 points from
baseline
Increase 30% from
baseline
The BPI-SF assesses pain through four
items rated from 0 (no pain) to 10 (as bad as
imaginable). It also measures interference
of pain with seven daily activities on 0–10
scales (0 = no interference; 10 = interferes
completely). Higher scores indicate higher
degree of pain.
Item 4: pain at its least 1 0–10
Item 5: pain on average 1 0–10
Item 6: pain now 1 0–10
Pain severity score 4 0–10
Pain interference score 7 0–10 Increase 50%
of baseline SD
Increase 50% of
baseline SD
EQ-5D = European Quality of Life 5-Domain Scale; FACT-G = Functional Assessment of Cancer Therapy–General; FACT-P = Functional Assessment of Cancer
Therapy–Prostate; FAPSI-8 = FACT Advanced Prostate Symptom Index-8; HRQoL = health-related quality of life; PCS = prostate cancer subscale; PRO = patient-
reported outcomes; SD = standard deviation; VAS = visual analogue scale.
}
}
E U R O P E A N U R O L O G Y 7 1 ( 2 0 1 7 ) 5 3 4 – 5 4 2 536