Statistical Analysis

We assessed overall frequencies (percentages) and means (SDs) for baseline patient characteristics using the Mantel-Haenszel χ² test for categorical variables and the t test for continuous variables. Crude and adjusted rates of ED visits or hospitalizations for hyperglycemic crises were calculated and presented as the total number of events per 1000 person-years among patients with type 1 diabetes or type 2 diabetes. Overall and annual adjusted rates of hyperglycemic crises were calculated using negative binomial estimates adjusted for age (during the year of the event), sex, race/ethnicity, region, and year. Adjusted rates of hyperglycemic crises for subgroups by age, sex, race/ethnicity, annual household income, and insulin use status (for type 2 diabetes only) were calculated using negative binomial estimates adjusted for age (at the index date), sex, race/ethnicity, and region. We tested for differences in the outcome by age, sex, race/ethnicity, income level, insulin use (in type 2 diabetes), and year using Wald tests.

Multivariable negative binomial regression models were also used to examine the association between hyperglycemic crises (dependent variable) and the independent variables outlined above. We evaluated hyperglycemic crises as the total number of events per each person, reporting results as incidence risk ratios (IRRs), 95% CIs, and P values. Person-years were used as an exposure to determine the estimated rates of and factors associated with hyperglycemic crises. Secondary analyses examined the outcomes of DKA and of HHS separately. Analyses were conducted using SAS Enterprise Guide, version 7.1 (SAS Institute Inc) and STATA, version 15.1 (StataCorp LLC). A 2-sided P < .05 was considered statistically significant.