Table 6. The sensitivity analysis.
Intention-to-treat analysis | Complete case analysis | |||||
Estimate (95% CI) | p value | n 1, n 2 a | Estimate (95% CI) | p value | n 1, n 2 a | |
Modified Rankin Scaleb | 1.04 (0.67, 1.60) | 0.86 | 148, 141 | 1.07 (0.68, 1.67) | 0.78 | 126, 130 |
Barthel Indexb | 1.50 (0.94, 2.37) | 0.09 | 148, 141 | 1.50 (0.92, 2.45) | 0.11 | 114, 113 |
NEADL Scalec | 1.59 (−3.26, 6.45) | 0.52 | 148, 141 | 0.64 (−3.71, 4.99) | 0.77 | 93, 100 |
EuroQol EQ-5Dc | −0.01 (−0.10, 0.07) | 0.79 | 148, 141 | −.004 (−0.09, 0.10) | 0.94 | 110, 112 |
EuroQol EQ-VASc | −4.40 (−11.43, 2.63) | 0.22 | 148, 141 | −2.35 (−9.54, 4.84) | 0.52 | 81, 96 |
Numbers for oxygen and control groups, respectively;
odds ratio (reference category is oxygen);
mean difference (control–oxygen); CI = confidence interval.
Higher scores on the modified Rankin scale and on the Nottingham Extended Activities of Daily Living (NEADL) Scale are worse; higher scores on the Barthel Index and EuroQol are better.