Racial factors associated with spatial clustering of mental health service utilization

Objective: The objectives of the study were to identify the racial factors associated with unusually high or low spatial “hotspots” of public sector mental health service utilization in Philadelphia. Method: The analysis was done using mapping and spatial data analysis. Address data from the adult Medicaid eligible population of over 200,000 and service users of ~50,000 in 2005 was geocoded and the geographical coordinates of the residences were used for the analysis. A K-function analysis for spatial clustering was employed to detect hotspots of utilization. The calculation provides for the average count of users in the area around each grid point. It then compares the observed cluster to the expected cluster based on a random scenario. The analysis used ArcView (ESRI) for spatial data visualization and Matlab (MathWorks) for K-function analysis. Results: The resulting maps clearly showed the existence of several hot-spots of under utilization of outpatient service use after controlling for the eligible Medicaid population in that area. The maps indicated that (i) proximity to a service provider appears to increase service use, and (ii) service use is lower in areas where the African American population is higher, based on census data. Further research is needed on reasons why variation in service use is occurring in certain sites, such as transportation networks, provider availability and quality of services.