BACKGROUND/RATIONALE:
Since its implementation in fiscal year 2000, Mental Health Intensive Case Management (MHICM) has been the VA's flagship case management program for severely impaired persons with serious mental illness. However, given the striking decline in VA psychiatric inpatient length of stay and the rapid expansion of the MHICM program, it is important to reassess its impacts.
OBJECTIVE(S):
Using data from a national cohort of VA enrollees with serious mental illness (SMI), the objectives of this study are to: (1) assess the average effects of MHICM on psychiatric inpatient utilization, partial hospitalization days, nursing home days, and costs during the 12 months after MHICM enrollment; and 2) estimate the relationship of the total numbers of psychiatric inpatient days occurring in the year prior to MHICM enrollment to the effects of MHICM on services utilization and costs; 3) assess the possible effects on services utilization and costs of specific changes to the high hospital use criterion (>30 psychiatric inpatient days or at least 3 psychiatric inpatient stays), such as requiring fewer prior inpatient days.
METHODS:
Using data from the VA National Psychosis Registry and the MHICM Program registry, we identified all patients who enrolled in MHICM between FY01-04 and a sample of 28,204 patients who were MHICM-eligible in this period but did not enroll (MHICM-eligible non-enrollees). MHICM eligibility criteria required >30 days or >=3 episodes of psychiatric hospitalization, a diagnosis of schizophrenia or bipolar disorder, and living within 60 miles of a VA hospital. MHICM enrollees (N=2,102) were propensity score matched to 2,102 MHICM-eligible non-enrollees. Ordinary Least Squares (OLS) regression analyses and instrumental variables (IV) analyses of matched data were used to assess effects of MHICM initiation.
FINDINGS/RESULTS:
Enrollment into MHICM was significantly (P<.05) negatively related to distance from the nearest MHICM team, having been diagnosed with a substance use disorder, homelessness, age and male gender, and was positively related to psychiatric inpatient days during the prior year. After matching, matched non-enrollees closely resembled (abs(t)<1.10) MHICM clients on all study covariates, which included inpatient and outpatient mental health services use in the prior 12 months, demographics, homelessness, service connection, and substance use diagnosis. In OLS analyses, MHICM initiation was associated with a 24% reduction in psychiatric inpatient days (22.6 vs. 29.8 for non-enrollees; P<0.001), a 16% increase in partial hospitalization days (12.6 vs. 10.9 for non-enrollees; P=.055), and a 64% reduction in nursing home days (7.6 vs. 21.4 for non-enrollees; P=.115). In instrumental variables analyses, MHICM initiation was associated with a 40% reduction in inpatient days (18.0 vs. 29.8 for non-enrollees; P=.042), a 120% increase in partial hospitalization days (23.9 vs. 10.9 for non-enrollees; P<.001), and a 44% reduction in nursing home days (12.0 vs. 21.4 for non-enrollees; P=.115). Reductions in inpatient utilization days following MHICM initiation were positively related the client's number of inpatient days prior to entry into the program.
IMPACT:
The MHICM program continues to support severely impaired veterans' independence from institutionalized care, and reduces costs among persons who otherwise would be dependent on institutions. Although MHICM teams usually are thought to provide all of clients' outpatient care, many MHICM clients also rely on partial hospitalization programs. Unimpeded access to partial hospitalization services may be critical to ensure that all MHICM clients are fully supported.
PUBLICATIONS:
Journal Articles
- Gold KJ, Kilbourne AM, Valenstein M. Improving care for patients with serious mental illness. American Family Physician. 2008; 78(3): 314-5.
