Incorrect Statistical Measures and Typographical Errors

In the Original Investigation article entitled “Effect of Routine Low-Dose Oxygen Supplementation on Death and Disability in Adults With Acute Stroke: The Stroke Oxygen Study Randomized Clinical Trial”¹ published in the September 26, 2017, issue of JAMA, incorrect data were reported. In Table 1, the SI conversion factor should have read “To convert glucose to mmol/L, multiply values by 0.0555.” In Table 2, some values, reported as 99% CIs were actually 95% CIs. The correct 99% CIs for the Barthel ADL index should be 70.2 (68.2 to 72.2) in the continuous oxygen group, 71.1 (69.1 to 73.1) in the nocturnal oxygen group, and 70.9 (68.9 to 72.8) in the control group; for the Nottingham Extended ADL, 9.66 (9.29 to 10.02) in the continuous oxygen group, 9.54 (9.17 to 9.90) in the nocturnal oxygen group, and 9.77 (9.40 to 10.14) in the control group; for VAS for quality of life, 55.4 (53.8 to 57.1) in the continuous oxygen group, 55.7 (54.1 to 57.3) in the nocturnal oxygen group, and 55.5 (53.8 to 57.1) in the control group; for highest systolic BP within 72 hours, −1.96 (−3.48 to −0.44) in comparison 1; for highest diastolic BP within 72 hours, −1.10 (−2.06 to −0.15) in comparison 1; for highest temperature within 7 days, 0.01 (−0.03 to 0.04) in comparison 1; and for serious adverse events, 0.94 (0.78 to 1.13) in comparison 1 and 1.19 (0.96 to 1.47) in comparison 2. In Figure 3, there were 4703 patients in the continuous or nocturnal oxygen group who did not have congestive heart failure, and for patients in the continuous or nocturnal oxygen group on whom thrombolysis was performed vs not, the P value test for interaction was .40. Changes to these reported statistical measures do not affect the conclusions of this study. This article was corrected online.