Abbott Proclaim 3660 Clinician Manual page 54

Three (3) subjects underwent an invasive procedure for a new clinically significant or

disabling chronic pain problem, and their overall VAS score was imputed using the hot deck
method or the last observation carried forward (LOCF) method, which replaces missing
values by using the last value observed for the score.
The following table shows the results of testing the primary endpoint. In both randomization
groups, burst stimulation scored lower overall than tonic stimulation on the VAS by 3.6 in the
tonic/burst group and 6.5 in the burst/tonic group. The estimated difference in the overall VAS
score between burst and tonic stimulation was -5.1, and the 95% upper confidence bound (UCB)
for the mean difference between burst and tonic stimulation was -1.14, which is less than the
non-inferiority margin of 7.5. Additionally, the p-value for the test of the non-inferiority hypothesis
was <0.001. Therefore, the primary endpoint was met, and it is concluded that burst stimulation
is non-inferior to tonic stimulation.
Table 28. Primary endpoint (overall VAS score)
Randomization
Burst
Mean ± SD (N)
Minimum, Median, Maximum
Tonic
Mean ± SD (N)
Minimum, Median, Maximum
Burst-Tonic
Mean ± SD (N)
Minimum, Median, Maximum
Burst-Tonic Average ± Pooled SD
Across Arms
95% UCB on Difference (Burst-Tonic)
Non-Inferiority Margin
p
-value for Non-Inferiority Test
Endpoint Met?
t
95% UCB and p-value for non-inferiority are based on a t-distribution with n
freedom, where n and n
1
42.7±26.1 (45)
0.9, 41.1, 100.0
46.3±22.8 (45)
1.0, 49.3, 88.1
-3.6±26.3 (45)
-64.6, -3.7, 79.0 -70.4, -2.6, 48.4
are the number of subjects in each arm.
2
Arm 1: Arm 2:
tonic/burst burst/tonic
44.2±25.3 (55)
0.0, 45.1, 98.1
50.0±24.8 (55)
0.0, 49.3, 98.7
-6.5±21.0 (55)
-5.1±11.8
-1.14
<0.001
50
Pooled
43.5±25.6
(100)
—
48.7±23.9
(100)
—
—
—
t
7.5
t
Yes
+n -2 degrees of
1 2