Skip to main content
. 2017 Jun 27;14(5):507–517. doi: 10.1177/1740774517716937

Table 1.

The performance of several sequential SW-CRT designs (Designs 1–6), along with that of the corresponding classical SW-CRT design (Design 7), is summarised, for Scenarios 1 and 2.

Scenario 1
Design T Stopping γe γf m E(M|0) P(RejectH0|0) E(M|δ) P(RejectH0|δ) minM maxM
Design 1 {2,3,4,5} E&F 0.5 0.5 104 1043.49 0.05 1113.17 0.90 832 2080
Design 2 {3,5} F NA 1 75 1031.73 0.05 1464.44 0.90 900 1500
Design 3 {3,4,5} E 1 NA 97 1912.03 0.05 1288.63 0.93 1164 1940
Design 4 {2,3,4,5} E&F 1.5 1 84 946.52 0.05 1040.49 0.90 672 1680
Design 5 {3,5} F NA 1.5 73 1032.61 0.05 1433.30 0.90 876 1460
Design 6 {3,4,5} E 0.5 NA 104 2044.88 0.05 1353.52 0.95 1248 2080
Design 7 {5} NA NA NA 70 1400.00 0.05 1400.00 0.90 1400 1400
Scenario 2
Design 1 {2,4,7,9} E&F 0.5 0.5 11 878.21 0.05 1063.59 0.81 440 1980
Design 2 {5,9} F NA 1 9 856.89 0.05 1037.91 0.82 900 1620
Design 3 {3,6,9} E 1 NA 8 1416.43 0.05 1031.39 0.81 480 1440
Design 4 {2,4,7,9} E&F 1.5 1 9 1583.44 0.05 1084.97 0.82 360 1620
Design 5 {5,9} F NA 1.5 8 924.10 0.05 1365.12 0.82 800 1440
Design 6 {3,6,9} E 0.5 NA 8 952.90 0.05 1382.72 0.83 480 1440
Design 7 {9} NA NA NA 7 1260.00 0.05 1260.00 0.81 1260 1260

E&F: efficacy and futility; E: efficacy; F: futility; NA: not applicable.

All rounding is to two decimal places.