Methods:
Public-use, all-payer ED data from non-federal, acute-care hospitals (2005–2014) were obtained for cross-sectional analysis. Due to substantial missing data, we used fully conditional specification multiple imputation with discriminant functions for age group, sex, race, and ethnicity. Multivariable logistic regression was used to examine asthma diagnosis (yes/no) among all ED visits and non-routine discharge (sent home vs. all else) among visits with asthma diagnosis. Primary independent variables were mental illness and the 3-digit zipcode of the patient’s residence. Covariates included demographics, payer type, and hospital characteristics.
Results:
During 2005–2014 there were 96,180,176 visits at 349 hospitals, and asthma diagnosis increased from 3.3% of ED visits in 2005 to 5.9% in 2014. However, asthma as a primary diagnosis decreased from 1.7% to 1.4% of ED visits. Among visits with asthma diagnosis (n = 4,419,629), co-occurring mood disorders increased from 2.1% in 2005 to 9.2% in 2014. Predictors of asthma diagnosis included attention deficit/conduct disorders [adjusted odds ratio (AOR) 1.41, 95% confidence interval (1.40–1.42)] and mood disorders [AOR 1.37, (1.36–1.37)]. Compared to Los Angeles, cities/areas most associated with asthma diagnosis were Richmond [zipcode 948, AOR 1.22 (1.20–1.24)], Bakersfield [933, AOR 1.21 (1.19–1.24)], and San Bernardino [924, AOR 1.20 (1.19–1.22)]. Ninety-six percent of ED visits with asthma resulted in routine discharge. Predictors of non-routine discharge included suicide/self-harm [AOR 4.74 (4.67–4.81)], schizophrenia [1.97 (1.94–1.99)], and mood disorders [1.35 (1.34–1.36)]. Areas associated with non-routine discharge included the Bakersfield vicinity [932, 1.29 (1.17–1.41)] and Ventura [930, 1.23 (1.10–1.38)].