Table 3 Effect of misclassification levels on the required sample size in a hypothetical trial of anticoagulant versus antiplatelet in patients with cardioembolic stroke to demonstrate statistically significant result for the relevant outcomes, using N = 5000 permutations
Stroke aetiology misclassification level | Death | Recurrent stroke |
|---|---|---|
0% | 198 | 502 |
5% | 237 (20% increase) | 605 (21% increase) |
20% | 352 (78% increase) | 973 (94% increase) |