Multifacility utilization by the chronically mentally ill in the Department of Veterans Affairs

Abstract

The objectives of this research are (1) to quantify the multifacility utilization patterns ("shared care") for a selected subset of patients with a chronic mental disorder, (2) to examine the patient characteristics correlated with several observed utilization patterns, and (3) to determine facility groups through cluster analysis, based on multifacility readmissions. Patients were identified from the VA's hospital discharge abstract system based on a diagnosis of schizophrenia in any discharge during the two-year study period. All discharges for this cohort during the period were analyzed. Essentially, three different groups of patients from this cohort were found. The vast majority have one or two hospitalizations and in the process are seen at one or maybe two facilities. However, two small minorities are hospitalized many times. One small group confines these hospitalizations to relatively few facilities while the other travels from facility to facility around the country. These two groups present very different clinical management problems from each other as well as from the majority. Since the distribution of travel distance between facilities remained similar for all types of patients (median distance was about 120 miles), facilities were clustered based on the number of linkages they shared. By use of six different measures of the disjointedness of shared care, clustering produced a reasonable number of meaningful groups which account for a majority of possibly uncoordinated shared care. We feel that these results and this approach can be useful for administrators, policymakers, and researchers. By looking simultaneously at the utilization patterns of its clients and their characteristics (e.g., age) as well as the utilization patterns within groups of facilities, any multifacility health system can characterize its shared care and identify those clients and facilities most in need of intervention, resources, and future study.