Skip to main content
. 2021 Feb 5;16(6):1593–1603. doi: 10.1007/s11739-021-02655-6

Table 2.

Association between corticosteroid use and in-hospital mortality

Methods a,b,c,d HR (95% CI) Corticosteroid versus Non-Corticosteroid p value
Multivariable analysis 0.67 (0.50–0.90) 0.01
Multivariable analysis adjusted for propensity score 0.67 (0.50–0.90) 0.01
Inverse probability weighting of pooling propensity score analysis 0.68 (0.52–0.90) 0.01
Propensity score matching analysis 0.67 (0.48–0.93) 0.02

HR hazard ratio, CI confidence interval

aMultivariable Cox regression analysis in the entire cohort, with stratification on race and azithromycin, was adjusted for age, gender, BMI, myocardial infarction, diabetes mellitus, hypertension, Charlson weighted score, number of ED visits prior to 6 months, initial vital signs, current medications, and laboratory tests on presentation

bMultivariable Cox regression analysis adjusted for propensity score was a multivariable Cox model with the same strata, covariates, and an additional adjustment for the propensity score in the entire cohort

cInverse probability weighting of pooling propensity score analysis was a multivariable Cox model with the same strata, covariates, and the stabilized weight based on the propensity score in the entire cohort

dPropensity score matching analysis was Cox regression in the matched cohort