Table 2 Comparison of assumptions and sample size calculation for power parallel cluster and stepped wedge designs
Design | Sites | Assumptions | Participants | Alpha | Power | ICC | Detectable difference |
|---|---|---|---|---|---|---|---|
Initial parallel RCT | 6 | Adherence 58.6% No change in control group adherence | 540 total | 0.05 | 80% | 0.01 | 15.5% |
0.05 | 24.6% | ||||||
0.10 | 31.1% | ||||||
Options 1 and 2: Stepped wedge | 12 | Adherence 58.6% No change in control group adherence | 432 total | 0.05 | 80% | 0.05 | 19.7% |
0.10 | 19.8% | ||||||
0.20 | 19.1% | ||||||
Option 3: Final parallel RCT | 16 | Adherence 58.6% No change in control group adherence | 576 total | 0.05 | 80% | 0.05 | 18.0% |
0.10 | 22.5% | ||||||
0.20 | 28.8% |