The goal of this study was to evaluate the impact of ethnic matching (client and clinician are of the same ethnic group), language matching (clinician is proficient in the language that the client prefers), and program involvement in treatment of minority patients on reducing in the use of emergency mental health service utilization. The authors’ premise was that lower levels of emergency room use implies that patients have greater capacity either to avoid crises of a magnitude sufficient to require an emergency intervention or to cope with them by alternative means.

All clients receiving outpatient mental health services for fiscal years 1987-1988 and 1989-1990 were included in the study. The number of participants in each fiscal year was 26,943 and 26,999 respectively. The participants were coded as African Americans (8% of the sample in 1987-88 and 10% of the sample in 1989-90), Hispanic (16% of each sample), Asian Americans (9% of each sample), and white (67% of the sample in 1987-88 and 65% of the sample in 1989-90). A majority of the patients were diagnosed with schizophrenia (32-35%), affective disorders (17-18%), or adjustment disorders (17-19%).

Participation in an ethnically-matched client-clinician relationship was associated with less frequent use of emergency services for African Americans, Hispanics and Asian Americans;
the effect was similar for all three ethnic groups (between .21 and .29 less visits per year). Participation in a language-matched relationship was also associated with reduction in use of emergency services for both Hispanic and Asian American clients (between .16 and .25 less visits per year). These analyses controlled for ethnicity, age, gender, diagnosis, adjustment status (GAS), education, employment status, marital status, presence of patient disability, the professional status of the clinician, and the gender of the clinician.

The investigators then constructed a variable to indicate ethnic matching alone, language matching alone, both ethnic and language matching, or neither ethnic non language matching. This variable was entered into a regression model that included all of the previously listed covariates, as well as the percent of the program’s patient population that was a minority ethnicity. Ethnic matching and language matching alone were each associated with a reduction of use of emergency services (between .06 and .13 visits per year), and both ethnic and language matching were associated with a reduction in emergency service use only in the 1989-90 fiscal year (a .11 day reduction). Additionally, African American clients were less likely to use emergency services in 1987-88 (a reduction of 0.02 visits) and Asian American clients were more likely to use emergency services in both years (an increase of 0.07 visits in 1987-88 and 0.35 visits in 1989-90) after controlling for matching and program involvement in treating minority patients. Emergency room use was also associated with age group 18-25, unemployment, lower overall adjustment (GAS), a diagnosis of dementia or alcohol/drug problems, and no physical disability. Having a clinician who held a professional degree was associated with avoidance of emergency care. More than one-third of the variance in emergency room use was explained by these models.

There are two main limitations to the analytic approach taken in this study. First, the data are cross sectional and therefore associations between improvement in patient functioning and either ethnic/language matching or program involvement in treating minority groups could not be assessed. Second, the authors did not discuss the validity of their premise that a reduction in emergency service use represented a desired outcome rather than, perhaps, a lack of access to needed mental health services.