From: Should RECOVERY have used response adaptive randomisation? Evidence from a simulation study
Operating characteristic where sample size matches RECOVERY | FuR | FeR | Tf | RMCf | Ts | RMCs | ||||
---|---|---|---|---|---|---|---|---|---|---|
i | ii | iii | i | ii | iii | |||||
61.6 | 38.6 | 32.3 | 60.7 | 38.3 | 29.6 | |||||
Proportion allocated to dexamethasone (%) | 33.3 | 50 | 62.9 | 62.4 | 57.0 (average | 57.8 (average) | ||||
38.4 | 61.4 | 67.7 | 39.3 | 61.7 | 70.4 | |||||
Mortality rate (deaths/6400 and percentage) | 1583 (24.7%) | 1556 (24.3%) | 1528 (23.9%) | 1527 (23.9%) | 1516 (23.7%) | 1514 (23.7%) | ||||
Statistical power (%) | 65.2 | 67.7 | 66.7 | 66.1 | N/A | 52.3 | 97.4 | N/A | 53.9 | 97.6 |
Type 1 error rate (%) | 4.7 | 5.3 | 5.0 | 5.0 | 5.8 | 5.2 | ||||
Bias (×10−2) | 0 | 0 | −2.8 | −3.3 | −0.7 | 0.8 | 0.7 | −0.4 | 2.6 | 1.5 |
Mean Squared error (× 10−4) | 1.27 | 1.18 | 1.21 | 1.28 | 6.57 | 3.57 | 16.0 | 6.67 | 3.55 | 16.2 |