Statistical Method:
All statistical analyses were performed using STATA (Version
16).[12] Weighting of patient-level observations was implemented to
obtain national estimates. Propensity matching was done by the Greedy
method. Multivariate regression analysis models were built by including
all confounders that were significantly associated with the outcome on
univariable analysis with a cutoff P-value of 0.2. Variables that were
deemed important determinants of the outcomes based on literature were
forced into the models. Binary outcomes such as index hospitalization
mortality and complications were compared by logistic regression. For
readmission comparison, we used the time to event cox regression
analysis. Demographic variables were compared using the Student t-test.
All p-values were two-sided, with 0.05 as a threshold for statistical
significance.