Required sample size for a definitive trial to detect estimated effect of SAFE+TAU on ED use
| Dispersion parameter (k) | ~47% Reduction (from 3.4 to 1.8 visits) in 12 months compared with TAU | |
| N per group (α=0.05; 80% power, β=0.2) | N per group (α=0.05; 90% power, β=0.1) | |
| 0.17 | 123 | 164 |
| 0.5 | 191 | 255 |
| 0.69 | 230 | 308 |
| 1 | 293 | 393 |
| 1.21 | 337 | 451 |
Dispersion parameter taken from the adjusted NBR model in the pilot RCT (ie, k=0.69; 95% CI: 0.17 to 1.21); see table 6.
ED, emergency department; RCT, randomised controlled trial; SAFE, Seizure First Aid Training For Epilepsy; TAU, treatment-as-usual.