Skip to main content
. Author manuscript; available in PMC: 2022 Feb 14.
Published in final edited form as: Psychiatr Serv. 2020 Apr 2;71(7):663–669. doi: 10.1176/appi.ps.201900407

Table 3:

Association between Race/Ethnicity and Treatment Type among 2008–2009 Medicaid Beneficiaries with Depression in 28 States and District of Columbia Using Multinomial Logistic Regression

Race/Ethnicity Model # OR 95% CI P-value
OR for Medication Alone vs. No Treatment
African American Unadjusted 0.54 0.53–0.55 <.001
Model 1 0.53 0.52–0.54 <.001
Model 2 0.48 0.47–0.49 <.001
Model 3 0.52 0.51–0.53 <.001
Hispanic Unadjusted 0.75 0.73–0.77 <.001
Model 1 0.75 0.73–0.77 <.001
Model 2 0.77 0.75–0.79 <.001
Model 3 0.71 0.69–0.73 <.001
Other Unadjusted 0.87 0.85–0.89 <.001
Model 1 0.80 0.78–0.82 <.001
Model 2 0.86 0.84–0.88 <.001
Model 3 0.84 0.82–0.86 <.001

OR for Therapy Alone vs. No Treatment
African American Unadjusted 2.03 1.79–2.31 <.001
Model 1 2.00 1.76–2.27 <.001
Model 2 1.98 1.75–2.25 <.001
Model 3 1.57 1.35–1.82 <.001
Hispanic Unadjusted 0.84 0.65–1.07 0.154
Model 1 0.85 0.66–1.08 0.178
Model 2 0.85 0.66–1.09 0.194
Model 3 1.07 0.82–1.39 0.619
Other Unadjusted 1.53 1.30–1.81 <.001
Model 1 1.40 1.19–1.66 <.001
Model 2 1.41 1.19–1.67 <.001
Model 3 1.45 1.20–1.75 <.001

OR for Medication & Therapy vs. No Treatment
African American Unadjusted 1.11 1.05–1.17 <.001
Model 1 1.09 1.03–1.15 0.003
Model 2 0.95 0.89–0.99 0.044
Model 3 0.76 0.71–0.81 <.001
Hispanic Unadjusted 0.67 0.60–0.74 <.001
Model 1 0.67 0.61–0.74 <.001
Model 2 0.70 0.63–0.77 <.001
Model 3 0.72 0.64–0.80 <.001
Other Unadjusted 1.02 0.95–1.10 0.603
Model 1 0.91 0.84–0.98 0.008
Model 2 1.00 0.93–1.08 0.995
Model 3 0.88 0.82–0.96 0.003

OR: Odds ratio; CI: confidence interval.

White people were the reference group in logistic regression models.

Model 1: sex and age group were adjusted. Model 2: sex, age group, and Elixhauser comorbidity were adjusted. Model 3: sex, age group, Elixhauser comorbidity, and zip code level variables (median personal income, proportion of black residents, proportion of Hispanic residents, proportion with less than high school diploma, and Gini coefficient) were adjusted.