Table 3 Virtual executions of the OSCAR trial using the Bayesian sequential designs
From: Using Bayesian adaptive designs to improve phase III trials: a respiratory care example
| Â | Design 2: Interim analysis at 250, 500 and 750 patients | Design 3: Interim analysis at 335 and 670 patients | Design 4: Interim analysis at 335, 500 and 670 patients | Design 5: Interim analysis at 503 and 755 patients | |
|---|---|---|---|---|---|
Interim 1 | Decision | Stopping criteria not met | Stopping criteria not met | Stopping criteria not met | Stopping criteria not met |
Randomisation allocation (control: HFOV) | 129: 121 | 174: 161 | 174:161 | 251:252 | |
Primary outcome (control; HFOV) | 49/118 (41.5%); 44/113 (38.9%) | 70/165 (42.4%); 57/153 (37.3%) | 70/165 (42.4%); 57/153 (37.3%) | 98/233 (42.1%); 93/240 (38.8%) | |
Posterior probability HFOV superior | 0.6406 | 0.8222 | 0.8222 | 0.7556 | |
Pmaxa | 0.2410 | 0.3958 | 0.3958 | 0.1747 | |
Interim 2 | Decision | Stopping criteria not met | Stop for futility | Stopping criteria not met | Stop for futility |
Randomisation allocation (control: HFOV) | 249:251 | 339:331 | 249: 251 | 380:375 | |
Primary outcome (control; HFOV) | 96/230 (41.7%); 93/239 (38.9%) | 136/330 (41.2%); 129/322 (40.1%) | 96/230 (41.7%); 93/239 (38.9%) | 154/375 (41.1%); 152/364 (41.8%) | |
Posterior probability HFOV superior | 0.7350 | 0.6490 | 0.7350 | 0.4146 | |
Pmaxa | 0.1315 | 0.0128 | 0.1315 | 0.0000 | |
Interim 3 | Decision | Stop for futility | NA | Stop for futility | NA |
Randomisation allocation (control: HFOV) | 377:373 | NA | 339:331 | NA | |
Primary outcome (control; HFOV) | 154/372 (41.4%); 152/363 (41.9%) | NA | 136/330 (41.2%); 129/322 (40.1%) | NA | |
Posterior probability HFOV superior | 0.4372 | NA | 0.6490 | NA | |
Pmaxa | 0.0000 | NA | 0.0128 | NA | |