Data Analysis
The outcome of interest wasfilling of EAI prescription after discharge from the ED. The variables of interest included the followingpatient characteristics: age, sex, race, ethnicity and insurance type. We re-categorizedrace and ethnicity into a single variable:Non-Hispanic white (NH-white), Non-Hispanic black (NH-black), Hispanicand other. Insurance status was categorized as commercial or Tricare, Medicaid-in-state, Medicaid-out-of-state, self-pay and other/unknown. Patient age was treated as a continuous variable. We used IBM SPSS Statistics for Windows version 26 (SPSS Inc., Chicago, Ill., USA) to perform descriptive statistics and logistic regression to measure associations between prescription fill rates and demographic factors.We performed bivariable logistic regression to assess for differences in prescription filling by patient demographics and performed multivariable logistic regression to adjust for potential confounding by all covariables. Regression models were fit to derive adjusted ORs with 95% CIs.