Inverse probability of treatment weights (IPTW) was performed to address imbalances between two study groups. Propensity weights were estimated using logistic regression with IXE vs ADA as the dependent variable and baseline variables as covariates (Table 3). The cohorts were considered balanced on a given variable if the standardized difference (Std. Diff) was ≤0.1. Statistical tests for significance were conducted using chi-square test for categorical variables and pooled two sample t-tests for continuous variables for weighted outcomes. A priori P value of <0.05 was considered statistically significant. The weighted probabilities of persistence, discontinuation, and switching were estimated using Kaplan-Meier curves and weighted log-rank test. Cox proportional hazards models were employed to estimate the risks of non-persistence, discontinuation, and switching. Logistic regression was used to estimate the odds of high adherence (PDC≥80%). Covariates used in the IPTW model were included in regression models as well. Effect sizes, hazard ratio (HR), and odds ratios (OR) were presented along with 95% confidence intervals (CI). Descriptive analyses were conducted using WPS version 4.2 (World Programming System, United Kingdom); IPTW and multivariable analyses were conducted using R version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria).
