From: Accounting for centre-effects in multicentre trials with a binary outcome – when, why, and how?
tPA vs placebo (n = 98) | DNase vs placebo (n = 94) | tPA+DNase vs placebo (n = 98) | |
---|---|---|---|
Unadjusted for centre-effects | |||
Patients included – no. (%) | 98 (100) | 94 (100) | 98 (100) |
Odds ratio (95% CI) | 0.42 (0.09 to 1.87) | 2.72 (0.88 to 8.42) | 0.23 (0.04 to 1.29) |
P-value | 0.25 | 0.08 | 0.10 |
Fixed-effects | |||
Patients included – no. (%) | 66 (67) | 80 (85) | 81 (83) |
Odds ratio (95% CI) | 0.54 (0.12 to 2.56) | 3.34 (1.01 to 11.00) | 0.24 (0.04 to 1.36) |
P-value | 0.44 | 0.048 | 0.11 |
Random-effects | |||
Patients included – no. (%) | 98 (100) | 94 (100) | 98 (100) |
Odds ratio (95% CI) | 0.42 (0.09 to 1.87) | 2.72 (0.88 to 8.42) | 0.23 (0.04 to 1.29) |
P-value | 0.25 | 0.08 | 0.10 |
GEE* | |||
Patients included – no. (%) | 98 (100) | 94 (100) | 98 (100) |
Odds ratio (95% CI) | 0.33 (0.06 to 1.75) | 2.57 (0.76 to 8.66) | 0.18 (0.03 to 1.19) |
P-value | 0.19 | 0.13 | 0.08 |
Mantel-Haenszel | |||
Patients included – no. (%) | 30 (31) | 45 (48) | 35 (36) |
Odds ratio (95% CI) | 0.34 (0.05 to 2.15) | 3.59 (0.99 to 13.03) | 0.21 (0.02 to 1.98) |
P-value | 0.27 | 0.04 | 0.13 |