(WITHDRAWN) County Mental Health System Characteristics and Use of Acute Mental Health Services

Background: Over the last fifty years, public mental health services have moved from the hospital into the community. Acute, hospital-based inpatient and emergency services play smaller but critical roles in service systems particularly for our most vulnerable citizens and racial/ethnic minorities. In addition to their disproportionally high cost, acute services are often viewed as stressful and traumatic for individuals in crisis. Unfortunately, little recent research has focused on the nature of those roles and how mental health systems use acute services. California's diverse, county-operated public mental health systems provide an excellent opportunity to explore relationships between local context and acute psychiatric service use.

Methods: This study used Medicaid MAX claims data from 2005-2013 for beneficiaries who used mental health services, received diagnoses of mood, anxiety, bipolar, or psychotic disorders, and lived in one of 24 diverse California counties. General estimating equation models were used to examine relationships between acute service use and county characteristics (e.g. demographics and economic conditions) and county mental health system variables (e.g. county spending on mental health) adjusting for beneficiary demographics, psychiatric diagnoses, county, and year.

Results: When considering county mental system characteristics, greater availability of mental health professionals was associated with less acute service use (Z=-7.49, p<.0001), while higher per capita spending on public mental health care was associated with more acute service use (Z=4.90, p<.0001). When considering county contextual factors beyond the mental health system, larger population (Z=-3.46, p=.0005) , higher population density (Z=-16.43, p<.0001), and a larger proportion of racial and ethnic minority residents (Z=-9.45, p<.0001) were associated with less acute service use, while a larger proportion of residents living in poverty (Z=10.16, p<.0001) and poorer mental health in the general population (Z=5.63, p<.0001) were associated with more acute service use.

Conclusions: These findings suggest that county contexts impacts use of acute psychiatric services. Acute service use is more common in counties with more limited resources in terms of mental health care providers and population income, as well as counties with smaller and more dispersed populations--factors that may limit availability of community-based services. Two findings require further exploration at the county level. The association between more racially/ethnically diverse populations and lower acute service use could reflect attenuation of previously documented overrepresentation of minority group members in acute services or underutilization of services. The relationship between higher per capita public mental health spending and more acute service use also warrants further exploration.