From: Accounting for centre-effects in multicentre trials with a binary outcome – when, why, and how?
Mean treatment effect | Type I error rate (%) | Convergence (%) | |
---|---|---|---|
Unadjusted for centre-effects | 1.00 | 4.7 | 100 |
Fixed-effects | 1.01 | 5.1 | 99.8 |
Random-effects | 1.00 | 4.8 | 99.4 |
GEE | 1.00 | 5.1 | 100 |
Mantel-Haenszel | 1.01 | 4.0 | 100 |