Skip to main content

Table 1 Scenarios 20 and 28 – performance measures for estimation of control arm RMST

From: Two-stage estimation to adjust for treatment switching in randomised trials: a simulation study investigating the use of inverse probability weighting instead of re-censoring

Scenario details Method Percent bias Empirical SE of % bias RMSE of % bias Coverage (%) Convergence (%)
Scenario number: 20
True RMST:
Control: 357
Experimental: 430
Mean switch: 57%
True ave. HR: 0.57
True ave. AF: 1.53
Mean censored: 50%
No switching 0.0 3.7 3.7 94.4 100.0
ITT 7.5 3.4 8.3 46.7 100.0
TSE −2.3 6.9 7.3 97.8 100.0
TSEnr 3.5 3.9 5.3 86.1 100.0
TSEipcw −5.1 16.3 17.1 96.0 95.8
min/max MC error 0.1/0.5 0.1/0.4 0.1/0.5 0.5/1.6
Scenario number: 28
True RMST:
Control: 228
Experimental: 322
Mean switch: 57%
True ave. HR: 0.56
True ave. AF: 1.85
Mean censored: 26%
No switching −0.1 5.7 5.7 94.7 100.0
ITT 15.1 5.5 16.0 29.1 100.0
TSE −3.5 9.1 9.8 93.0 100.0
TSEnr 4.0 6.5 7.6 91.6 100.0
TSEipcw 1.1 11.3 11.3 97.1 99.9
min/max MC error 0.2/0.4 0.1/0.3 0.1/0.8 0.5/1.4
  1. Note: RMST restricted mean survival time, HR hazard ratio, AF acceleration factor, SE standard error, RMSE root mean squared error, MC Monte-Carlo, ITT intention to treat, TSE two-stage estimation, TSEnr two-stage estimation without re-censoring, TSEipcw two-stage estimation with inverse probability of censoring weights