Table 3 |
|||||
Effect of various methods of handling missing response in trial | |||||
Methods of analysis | PIH in TXA and placebo group | ||||
Group | n/N | Rate | RD | RR | |
(person) | (%) | [95% CI] | [95% CI] | ||
1. Complete case analysis | TXA | 16/115 | 14% | −0.11 | 0.57 |
Placebo | 28/114 | 25% | [(−0.21)-(−0.01)] | [0.32-0.99] | |
2. Assuming poor outcome | TXA | 21/120 | 18% | −0.10 | 0.65 |
Placebo | 32/118 | 27% | [(−0.20)-(−0.01)] | [0.40-1.05] | |
3. Assuming good outcome | TXA | 16/120 | 13% | −0.10 | 0.56 |
Placebo | 28/118 | 24% | [(−0.20)-(−0.01)] | [0.32-0.98] | |
4. Extreme case favoring placebo | TXA | 21/120 | 18% | −0.06 | 0.74 |
Placebo | 28/118 | 24% | [(−0.16)-0.04] | [0.44-1.22] | |
5. Extreme case favoring TXA | TXA | 16/120 | 13% | −0.14 | 0.49 |
Placebo | 32/118 | 27% | [(−0.24)-(−0.04)] | [0.29-0.85] |
Yutthakasemsunt et al.
Yutthakasemsunt et al. BMC Emergency Medicine 2013 13:20 doi:10.1186/1471-227X-13-20