Statistical methods and analysis
The cohort was stratified by their racial/ethnicity background in whites
or Caucasians, black or African American, Hispanic, and others to assess
baseline characteristics and outcomes comparatively. Sampling weights
provided by the NIS were used for national estimations. TEVAR incidence
rates were calculated for White, Black, and Hispanic patients using
race-specific US Census population data for each year
(Supplementary Table 2 ). Univariate comparison between groups
was performed with Pearson Chi2 test for categorical
variables. Continuous data, including patient age, hospital length of
stay, and total charges, were compared using the Kruskal-Wallis or ANOVA
after testing for normality using the Shapiro–Wilk test. Bonferroni
correction and Tukey’s multiple comparison adjustment prevented Type I
error. We have four groups which mean six pairwise; the significance
value of p used was less than 0.05/6=0.008 to reject the null
hypothesis. The primary outcome was in-hospital mortality. Secondary
outcomes included complications identified through ICD codes
(Supplementary Table 3 ) and resources utilization, such as
length of stay and hospital charges. A mixed-effects multivariable
logistic regression assessed the relationship between race and the
primary outcome, using hospitals as the random effect to account for
interhospital variability. Adjusting covariates were selected through a
correlation matrix. R (4.0.0 ’Arbor Day’) was utilized for the analysis.