Table A1.
Example power calculations to assess the potential impact of co-enrolment
| Modelling assumptions |
Estimated impact on trial X |
||||||
|---|---|---|---|---|---|---|---|
| Effect of aspirin on 5 year survival∗ | Participation rates in Add-Aspirin† |
5 year survival in trial X with co-enrolment |
Power (loss/gain in power) | Extra patients (OR follow-up) needed for 80% power | |||
| Control | Intervention | Control | Intervention | Difference | |||
| Trial X result is positive (5 year survival 55% control versus 45% intervention) in the absence of co-enrolment | |||||||
| 6% | 10% | 10% | 45.4% | 55.4% | 10.0% | 79.9% | 2 (1 month) |
| 20% | 20% | 45.8% | 55.8% | 10.0% | 79.9% | 3 (1 month) | |
| 30% | 30% | 46.2% | 56.2% | 10.0% | 79.8% | 5 (1 month) | |
| 10% | 15% | 45.4% | 55.6% | 10.2% | 81.5% | – | |
| 10% | 20% | 45.4% | 55.8% | 10.4% | 82.9% | – | |
| 15% | 30% | 45.6% | 56.2% | 10.6% | 84.3% | – | |
| 15% | 10% | 45.6% | 55.4% | 9.8% | 78.3% | 44 (3 months) | |
| 20% | 10% | 45.8% | 55.4% | 9.6% | 76.6% | 89 (5 months) | |
| 30% | 15% | 46.2% | 55.6% | 9.4% | 74.8% | 138 (8 months) | |
| 10% | 10% | 10% | 45.7% | 55.7% | 10.0% | 79.9% | 3 (1 month) |
| 20% | 20% | 46.3% | 56.3% | 10.0% | 79.8% | 5 (1 month) | |
| 30% | 30% | 47.0% | 57.0% | 10.0% | 79.8% | 7 (1 month) | |
| 10% | 15% | 45.7% | 56.0% | 10.3% | 82.4% | – | |
| 10% | 20% | 45.7% | 56.3% | 10.7% | 84.7% | – | |
| 15% | 30% | 46.0% | 57.0% | 11.0% | 86.8% | – | |
| 15% | 10% | 46.0% | 55.7% | 9.7% | 77.2% | 75 (5 months) | |
| 20% | 10% | 46.3% | 55.7% | 9.3% | 74.2% | 154 (9 months) | |
| 30% | 15% | 47.0% | 56.0% | 9.0% | 71.1% | 244 (15 months) | |
| Trial X result is negative (5 year survival 45% in both arms) in the absence of co-enrolment | |||||||
| 6% | 10% | 10% | 45.4% | 45.4% | 0.0% | ||
| 20% | 20% | 45.8% | 45.8% | 0.0% | |||
| 30% | 30% | 46.2% | 46.2% | 0.0% | |||
| 10% | 15% | 45.4% | 45.6% | 0.2% | |||
| 10% | 20% | 45.4% | 45.8% | 0.4% | |||
| 15% | 30% | 45.6% | 46.2% | 0.6% | |||
| 15% | 10% | 45.6% | 45.4% | −0.2% | |||
| 20% | 10% | 45.8% | 45.4% | −0.4% | |||
| 30% | 15% | 46.2% | 45.6% | −0.6% | |||
| Trial X result is negative (5 year survival 45% in both arms) in the absence of co-enrolment | |||||||
| 10% | 10% | 10% | 45.7% | 45.7% | 0.0% | ||
| 20% | 20% | 46.3% | 46.3% | 0.0% | |||
| 30% | 30% | 47.0% | 47.0% | 0.0% | |||
| 10% | 15% | 45.7% | 46.0% | 0.3% | |||
| 10% | 20% | 45.7% | 46.3% | 0.7% | |||
| 15% | 30% | 46.0% | 47.0% | 1.0% | |||
| 15% | 10% | 46.0% | 45.7% | −0.3% | |||
| 20% | 10% | 46.3% | 45.7% | −0.7% | |||
| 30% | 15% | 47.0% | 46.0% | −1.0% | |||
The table illustrates the potential impact of co-enrolment into Add-Aspirin on the power of a hypothetical study, trial X. Selected results are shown from models performed under a range of assumptions about the factors listed in Figure 2, including scenarios felt to illustrate the largest plausible impact on power.
Trial X: A hypothetical two-arm superiority randomised controlled trial of a new peri-operative chemotherapy regimen versus standard in gastro-oesophageal patients. Designed with 80% power to detect a 10% improvement (from 45% to 55%) in survival at 5 years, requiring 500 patients per arm. Patients who are disease free at the end of treatment may become eligible for Add-Aspirin.
A 6% survival benefit at 5 years is hypothesised in Add-Aspirin (gastro-oesophageal). Models are repeated for larger benefits to illustrate potential effects on power.
A range of participation rates are used to assess potential impact – actual rates are unlikely to reach 30% (limited by overlap in recruiting centres and recruitment periods, as well as trial X participants being ineligible or unwilling to participate in Add-Aspirin). Differences in rates between arms are also unlikely to be as large as illustrated here.