Identifying Disparities in Behavioral Health Service Utilization Among People with Co-Occurring Substance Abuse and Mental Illness
The co-occurrence of substance use and mental health disorder, or comorbidity, has been widely documented. People living with multiple behavioral health issues require various services to ensure their well-being, but because of their multiple needs they tend to fall between cracks of separate service systems. This project aims to advance understanding of the service utilization of people with comorbidity (PWC).
The National Survey on Drug Use and Health (NSDUH) is an ongoing survey conducted in the 50 states and the District of Columbia. With the most recent data available (2010), this study investigates predictors of service receipt among PWC, and extends prior research by examining if the association between need for treatment and actual utilization is modified by demographic factors.
Method:
The NSDUH survey is based on a multistage, stratified sampling design. Computer-assisted interviews were used to gathered information from household residents across the United States. Only people age 18 years or older with both mental illness and substance dependence, as identified by “any mental illness and drug/alcohol dependence or abuse”, were selected for this study (n=2,441). Two questions about participation in mental health treatment and substance use treatment were used to create the dependent variable, behavioral health service utilization (BHSU), which was then dichotomized for analysis. Categorical predictors included gender, age, educational level, race, insurance resources, income level, high/low perceived and evaluated need for substance use treatment (PSU and ESU), and high/low perceived and evaluated need for mental health treatment (PMH and EMH). SPSS was used to conduct bivariate and multivariate logistic regression. Forward stepwise methods were used to select significant 2-way interaction effects. Log likelihood ratio tests were performed to analyze improvements in model fit. All significance tests were two-tailed, P<.05.
Results:
Less than half (41%) of the PWC received BHS in year preceding the survey. There were significant main effects of age, race, insurance, PMH and EMH on BHSU. Compared to people over 35 years old, younger people were less likely to receive treatment (odds ratio (OR) =0.30), non-Hispanic whites were more likely to receive treatment than Hispanics (OR=1.65), and people with insurance were more likely to be in treatment (OR=1.48). Higher PMH and EMH also increased the odds of BHSU (OR=1.50 and 2.4). The effects of PSU and ESU on BHSU were modified by gender and education, respectively. PSU has a greater effect on BHSU among female than males. ESU had a significant negative effect among people who had a high school degree or less but not among people with a college degree.
Conclusions and Implications
The findings suggest disparities among PWC are pervasive as gender, age, race, and insurance are associated with BHSU, and highlight policy-relevant questions related to health care inequalities. Clinical factors (EMH, PMH, ESU, and PSU) are also important predictors of obtaining BHS. Promoting self-awareness or identification of clinical symptoms by outreach or intervention programs that are sensitive to gender and education level factors may help improve access.