Table 2.
Median treatment effects estimated by the PMRM, MMRM, and the Cox proportional hazards modela
| Sample size | Study length (months) | Median effect | ||
|---|---|---|---|---|
| PMRM (% time delay, measured on the horizontal axis) | MMRM (% reduction in CDR-SB, measured on the vertical axis) | Cox (HR of progressing to dementia) | ||
| 200 | 18 | 19% | 23% | 0.84 |
| 200 | 24 | 20% | 26% | 0.82 |
| 200 | 36 | 20% | 25% | 0.82 |
| 300 | 18 | 19% | 23% | 0.85 |
| 300 | 24 | 19% | 26% | 0.83 |
| 300 | 36 | 20% | 25% | 0.84 |
| 400 | 18 | 19% | 22% | 0.85 |
| 400 | 24 | 20% | 26% | 0.83 |
| 400 | 36 | 20% | 25% | 0.83 |
| 500 | 18 | 19% | 22% | 0.85 |
| 500 | 24 | 20% | 26% | 0.83 |
| 500 | 36 | 20% | 25% | 0.83 |
| 600 | 18 | 19% | 22% | 0.85 |
| 600 | 24 | 19% | 26% | 0.83 |
| 600 | 36 | 20% | 25% | 0.83 |
| 700 | 18 | 19% | 22% | 0.85 |
| 700 | 24 | 19% | 25% | 0.83 |
| 700 | 36 | 20% | 25% | 0.83 |
aData are shown across scenarios with different sample sizes and study lengths (1000 simulations per scenario). The table shows the treatment effect at the final visit for the PMRM and the MMRM. The hazard ratio summarizes the treatment effect over the entire study duration