Trends and Variations in Emergency Department Use Associated With Diabetes in the US by Sociodemographic Factors, 2008-2017.

Importance: Little is known about emergency department (ED) use among people with diabetes and whether the pattern of ED use varies across geographic areas and population subgroups. Objective: To estimate recent national- and state-level trends in diabetes-related ED use overall and by race and ethnicity, rural or urban location, and insurance status. Design, Setting, and Participants: This cross-sectional study of adults visiting the ED with a diabetes-related diagnosis used serial data from the Nationwide Emergency Department Sample, a nationally representative database, and discharge records from 11 state emergency department databases for 2008, 2011, 2014, and 2016 to 2017. Data were analyzed from March 16 to November 9, 2020. Exposures: Reported race and ethnicity, rural or urban location, and insurance status. Data were stratified to generate state-specific estimates. Main Outcomes and Measures: Rates of ED use for all-cause visits among adults with diabetes (all-cause diabetes visits) and visits with primary diagnoses of diabetes-specific complications. Results: A larger portion of all-cause diabetes ED visits (n = 32 433 015) were by female (56.8%) and middle-aged (mean [SD] age, 58.4 [16.3] years) adults with diabetes. Nationally, all-cause diabetes ED visits per 10 000 adults increased 55.6% (95% CI, 50.6%-60.6%), from 257.6 (95% CI, 249.9-265.3) visits in 2008 to 400.8 (95% CI, 387.6-414.0) visits in 2017. All-cause diabetes ED visits increased more for urban (58.3%; 95% CI, 52.5%-64.1%) and uninsured subgroups (75.3% [95% CI, 59.8%-90.8%]) than for their counterparts. Diabetes-specific ED visits (weighted number of 1 911 795) nationally increased slightly among all subgroups. State-specific ED use rates show wide state-to-state variations in ED use by race and ethnicity, rural or urban location, and insurance. On average across states, diabetes-specific ED use among Black patients was approximately 3 times (rate ratio, 3.09 [95% CI, 2.91-3.30]) greater than among non-Hispanic White patients, and among Hispanic patients, it was 29% greater (rate ratio, 1.29 [95% CI, 1.19-1.40]) than among non-Hispanic White patients. The mean rate of ED use among rural patients was 34% greater (rate ratio, 1.34 [95% CI, 1.26-1.44]) than among urban patients. The mean rates of ED use among patients with Medicaid (rate ratio, 6.65 [95% CI, 6.49-6.82]) and Medicare (rate ratio, 4.37 [95% CI, 4.23-4.51]) were greater than among privately insured adults. Conclusions and Relevance: This study suggests that disparities in diabetes-related ED use associated with race and ethnicity, rural or urban location, and insurance status were persistent from 2008 to 2017 within and across states, as well as nationally. Further geographic and demographic-specific analyses are needed to understand the sources of inequity.