How are caseload and service utilisation of psychiatric services influenced by distance? A geographical approach to the study of community-based mental health services

Introduction The aim of this study was to assess how the caseload and the utilisation of community-based mental health services is influenced by distance and to socioeconomic characteristics. Methods Spatial and statistical analyseswere conductedwith a sample of 12,347 patients,with ICD-10 psychiatric diagnosis, who had at least one contact with psychiatric services in Verona, Italy, between 2000 and 2006. Three types of mental health facility were considered: acute inpatient wards, outpatient clinics, and community mental health centres (CMHC). To measure distance and accessibility, the locations of static mental health facilities and patients’ homes were geocoded. Data were organised in a spatial database, which included census blocks, catchment areas locations, road network graphs, patients’ and facilities’ locations. In order to calculate travel distances, patients’ and facilities’ locations were connected to the road network. Accessibility was modelled by using the Network Analyst Service Area Function and 13 Service Areas were created around all facility locations, by measuring distances along the street network. For the epidemiological analyses, patients and census block centroids were linked to the service areas by using spatial join techniques. Epidemiological and utilisation analyseswere performed for each type of setting. Results The facilities were not equally located in the catchment areas. Of particular significance, rural areas appear to be poorly served by mental health services. The distance decay effect exists, with different trends for the three types of facility. The caseload (number of patients using services) decreased with increasing distance; at a distance of 10 km, there was a decrease of 80, 60 and 85%, respectively, for CMHCs, inpatients wards and outpatients clinics. From the Poisson regression models, distance was significantly correlated (p value\0.0001) with service use. Also univariate analyses showed a statistically significant association between distance and caseload for each type of setting (p value\0.05), with a decrease in service use for each service area increase in distance (1.5%for acute inpatientwards, 2.0%forCMHC, and 2.1% outpatient clinics). By adding other predictors in the Poisson regression models, these percentages increased. Conclusions Further studies are needed to evaluate the influence of other factors, such as environmental variables, that may influence the use of mental health services