When a trial is designed as an equivalence or non inferiority study and the two treatments are equivalent, i.e. differences in treatment effects do not equal or exceed preset margins, results should be classified as positive.
Postpathology patients exceeded the Trial's preset noninferiority margin of 2.5%% (5.4 vs 1.7 %, p = 0.069); prepathology patients did not (2.1 vs 1.0 %, p = 0.31).
With a 90% confidence interval, the lower limit was 9.7%, approaching the preset 10% margin, which led to some criticism [ 28].
In the intent-to-treat analysis with 57 in each group, the upper 97.5% confidence boundary for the difference between the 80.7% (46 of 57) non-failures in the 5 days on/2 days off, and the 70.2% (40 of 57) non-failures in the continuously treated arm was 5.4% which is again well within the preset non-inferiority margin of 15%.
The upper 97.5% confidence boundary for the difference between the percent of non-failures in the 5 days on/2 days off (88.5% non-failure) arm compared to continuous treatment (78.4% non failure) was 4.8% which is well within the preset non-inferiority margin of 15%.
The upper 97.5% confidence boundary for the difference between the percent of non-failures in the 5 days on/2 days off arm (88.5% non-failure) compared to continuous treatment (78.4% non failure) was 4.8% which is well within the preset non-inferiority margin of 15%.
In the modified intent-to-treat analysis which designated the 5 in each group who discontinued treatment as failures, the upper 97.5% confidence boundary for the difference between the percent of non-failures in the 5 days on/2 days off, 80.7% (46 of 57), and that in the continuously treated, 71.4% (40 of 56) was 6.6% which is within the preset non-inferiority margin of 15%.
The acceptable margin for non-inferiority was preset at −9 mm, ie, less than the minimum perceptible clinical improvement of −10 mm.
In Nagano, everything was preset".
("Danger, Will Robinson!") ISpeak has 11 preset voices.
