Table 3.
Adaptive study design power analysis showing two possible effect estimates and various scenarios
| Scenario | Success rate (%) | Powera | FWERb | |||
|---|---|---|---|---|---|---|
| Standard | mHealth | Socio-behavioral | mHealth + socio-behavioral | |||
| Null | 70% | 70% | 70% | 70% | --- | 0.019 |
| Single | 70% | 70% | 70% | 90% | 0.95 | 0.010 |
| Two | 70% | 70% | 90% | 90% | 0.97 | 0.006 |
| All | 70% | 90% | 90% | 90% | 0.98 | --- |
| Additive | 70% | 80% | 80% | 90% | 0.92 | --- |
| Null | 75% | 75% | 75% | 75% | --- | 0.020 |
| Single | 75% | 75% | 75% | 90% | 0.77 | 0.009 |
| Two | 75% | 75% | 90% | 90% | 0.84 | 0.005 |
| All | 75% | 90% | 90% | 90% | 0.87 | --- |
| Additive | 75% | 80% | 83% | 90% | 0.74 | --- |
aPower Probability of correctly identifying at least one superior arm compared to standard
bFWER Family-wise error rate, the probability of erroneously declaring any non-superior arm(s) superior