Table 4 The effective sample sizes and corresponding information fractions and power estimates from the four comparisons of newer treatments vs. low-dose NRT
From: Sample size and power considerations in network meta-analysis
Comparison (vs. low-dose NRT) | Effective head-to-head sample size | Effective indirect sample size | Total effective sample size | Information fraction | Statistical power | Statistical power to detect observed improvement* | Network meta-analysis odds ratio estimate (95% credible interval) |
|---|---|---|---|---|---|---|---|
Combination NRT | 1,664 | 1,691 | 3,355 | 53% | 66% | - | 1.05 (0.76 – 1.41) |
High-dose NRT | 3,605 | 2,211 | 5,816 | 92% | 88% | >99% | 1.32 (1.11 – 1.57) |
Buproprion | - | 7,707 | 7,707 | >100% | 95% | - | 0.99 (0.86 – 1.14) |
Varenicline | 720 | 3,625 | 4,268 | 68% | 76% | >99% | 1.38 (1.15 – 1.64) |