Table 3.

Outcomes analysis^a

	Estimate (95% CI)	P value
Hospital mortality	1.23 (0.61 to 2.49)	0.562
Hospital-free days	−3.42 (−6.11 to − 0.74)	0.013
ICU mortality	1.27 (0.57 to 2.82)	0.559
ICU-free days	−2.07 (−3.36 to −0.78)	0.002

^a For linear and logistic regression, we used generalized estimating equations to account for non-intubated patients selected multiple times as matches. Estimates are odds ratios for mortality endpoints and values above 1 represent increased in odds of event due to early intubation within 24 h of ICU admission. Estimates for length of stay endpoints are for the increase in hospital or ICU-free days associated with early intubation within 24 h of ICU admission (estimates less than 0 indicate longer length of stay and thus, fewer hospital or ICU-free days). To account for uncertainty introduced by multiple imputation, analyses were run separately for each imputation and combined using methods to estimate the between and within sample