Cost-Effectiveness of Intensive Psychiatric Community Care for High Users of Inpatient Services
Background This 2-year experimental study evaluated the effectiveness and cost of 10 intensive psychiatric community care (IPCC) programs at Department of Veterans Affairs medical centers in the northeastern United States.
Methods High users of inpatient services were randomly assigned to either IPCC or standard Department of Veterans Affairs care at 6 general medical and surgical hospitals (n=271 vs 257) and 4 neuropsychiatric hospitals (n=183 vs 162). Patient interviews every 6 months and national computerized data were used to assess clinical outcomes, health service use, health care costs, and non–health care costs.
Results There was only 1 significant clinical difference between groups across follow-up periods: IPCC patients at general medical and surgical sites had higher community living skills. However, at the final interview, IPCC patients at general medical and surgical sites showed significantly lower symptoms, higher functioning, and greater satisfaction with services. Treatment with IPCC significantly reduced hospital use only at neuropsychiatric sites (320 vs 513 days, P<.001). Total societal costs, including the cost of IPCC, were lower for IPCC at neuropsychiatric sites ($82454 vs $116651, P<.001), but greater at general medical and surgical sites ($51537 vs $46491, P<.01). When 2 sites that incompletely implemented the model were dropped from the analysis, costs at general medical and surgical sites were $38 lower for IPCC (P=.26).
Conclusions At acute care hospitals, IPCC treatment is associated with greater long-term clinical improvement and, when fully implemented, is cost-neutral. At long-stay hospitals treating older, less-functional patients, it is not associated with clinical or functional improvement but generates substantial cost savings. Intensive psychiatric community care thus has beneficial, but somewhat different, outcome profiles at different types of hospitals. (Authors)
Type of Resource: