Who Terminates From ACT and Why? Data From the National VA Mental Health Intensive Case Management Program Somaia Mohamed, Ph.D. Robert Rosenheck, M.D. Timothy Cuerdon, Ph.D.
Objective: One of the original principles of assertive community treatment (ACT) is that treatment should be time unlimited. Although termination is not uncommon in ACT, it has not been empirically studied. This study examined termination from a large program based on ACT. Methods: This study used national data from the Department of Veterans Affairs Mental Health Intensive Case Management program modeled on ACT to compare veteran characteristics, patterns of service delivery, and early clinical changes among veterans who terminated early (less than one year) and later (one to three years) with those of veterans had not terminated after three years. Bivariate comparisons and multinomial logistic regression analyses were used to identify factors associated with early and later termination. Results: Among 1,402 veterans enrolled in fiscal years (FY) 2002–2004, 16% terminated early, 26% terminated later, and 57% had not terminated after three years. Compared with those who had not terminated, those who terminated early showed higher suicidality scores, and participants who terminated early and those who terminated later were less likely to have a diagnosis of schizophrenia and were more likely to have lower quality of life at entry. Stronger differentiating effects were observed for program participation. Those who terminated received less intensive services during the first six months of participation and had a weaker therapeutic alliance. Although participants who terminated early showed more violent behavior at follow-up than the other two groups, there were no other differences in early clinical changes. Conclusions: Rates of both early and later termination were substantial, and less active participation was a stronger predictor of termination than either patient characteristics or clinical changes. A diagnosis of schizophrenia was associated with continued treatment. Further research is needed to determine the impact of termination on longer-term outcomes. (Psychiatric Services 61:675–683, 2010)
A
long-standing operating principle of assertive community treatment (ACT) is that services should be time unlimited. The Dartmouth ACT fidelity scale lists
long-term service delivery as a fidelity indicator, such that the highest score on this item is for programs that provide services indefinitely (1). The principle of “continuous long-term
Dr. Mohamed is with the Department of Psychiatry, Department of Veterans Affairs (VA) Northeast Program Evaluation Center, 950 Campbell Ave., West Haven, CT 06516 (e-mail:
[email protected]). Dr. Rosenheck is with the VA New England Mental Illness Research, Education, and Clinical Centers, West Haven, Connecticut. Dr. Cuerdon is with the Office of Mental Health Services, Department of Veterans Affairs, Washington, DC.
PSYCHIATRIC SERVICES
o ps.psychiatryonline.org o July 2010 Vol. 61 No. 7
service” originated with research findings that even after 14 months of treatment, participants who terminated experienced subsequent clinical deterioration and increased hospitalization (2). However, the assumption that such services should be time unlimited may limit access to ACT, because staff resources may remain committed to clients who can successfully transition to other services. Despite the continuing evolution of community practices for people with severe mental illnesses, there have been no recent studies supporting time-unlimited service delivery, and a small number suggest that safe transitions are possible (3–5). We are unaware of publications that have presented basic information on the proportions of patients who discontinue ACT treatment either early in treatment or after more extensive involvement or on clinical correlates of early or late termination. It is unclear whether persons who terminate from treatment have more or less severe problems, make more or less use of services initially, or show clinical improvement or deterioration. Further consideration of this long-standing ACT practice of timeunlimited treatment would require data on postdischarge outcomes, but the design of such studies would be informed by descriptive data on the natural process of termination. Since 1987 the Department of Veterans Affairs (VA) has developed a large national program based on the ACT model. This program, referred to as Mental Health Intensive Case Management (MHICM), currently 675
serves over 7,000 veterans annually. The MHICM program conforms to the ACT program criteria pertaining to human resources, organizational boundaries, service delivery, and substance abuse treatment (6). We used administrative data from the MHICM program to compare experiences of veterans who terminated services before one year or between one and three years of admission with those of veterans who continued beyond three years. Through these comparisons we seek to present basic descriptive information on a littlestudied but potentially important aspect of ACT.
Methods The MHICM program The MHICM program serves veterans with severe and persistent mental illness and significant functional impairment who are inadequately served by standard outpatient care and who typically have high past levels of hospital use. MHICM services are intended to follow the ACT model (6,7). Data sources In this study we included veterans enrolled in the MHICM program from FY 2002 to FY 2004 who had complete baseline and service delivery forms during the first six months of participation (N=1,402). Sociodemographic characteristics, diagnoses, clinical status, and community adjustment were documented on standardized forms completed at the time of admission. Data on patterns of service delivery were obtained from structured semiannual summaries completed by MHICM staff for each veteran after six months of participation, at the time of termination, or both. Termination was documented by clinicians at the time it occurred, and such data were available through the end of FY 2007, ensuring at least three years of potential participation. A veteran is considered to have terminated (and case managers complete a termination documentation form) when a veteran is judged to no longer need intensive case management, when all efforts to engage a veteran in the program have failed, when the veteran has moved 676
out of the team’s catchment area, or when he or she is placed in a longterm care facility, such as a nursing home. In this article termination thus refers to a clinical judgment by the case manager that the veteran will no longer be participating in the program for any one of several reasons. It does not preclude future participation, nor does it reflect the source of the decision (case manager, veteran, or some other party). It is entirely an indicator of present expectation of future behavior. Once a veteran has terminated from the program, all MHICM services discontinue and a new referral is required to reenter the program. Detailed reasons for termination were not documented during the period covered by this study, but the clinicians’ impression of reasons for termination was coded for a limited number of categories after FY 2005 (for example, moved away, did not need further treatment, moved to nursing home or other institution, was uncooperative, or refused services despite clinicians’ concerted efforts to engage the veteran). Veterans who died were excluded from the analyses presented here. Additionally, an outcome assessment was completed after six months, allowing comparison of clinical change for a subset of participants for whom data were available (N=913, 65%). The institutional review boards of the VA Connecticut Healthcare system and Yale Medical School approved this study. Measures Sociodemographic measures documented age, gender, race, education, marital status, employment, income, disability status, and assignment to a payee or fiduciary to manage the veteran’s finances. Clinical measures included the diagnosis assigned by the treating clinician at the time of enrollment. The Brief Psychiatric Rating Scale (8) was used to measure symptom severity, and the Brief Symptom Index (BSI) Global Severity Index (GSI) (9) was used to measure subjective distress. Selected questions from the Addiction Severity Index documented alcohol and drug use in the previous 30 days (10). PSYCHIATRIC SERVICES
Lifetime duration of both VA and non-VA psychiatric hospitalizations was also coded. Additionally, self-reported rating scales were used to assess suicidality (range 0 to 5, with higher scores indicating more serious suicidality—for example, requiring hospital care) and violence (range 0 to 5, with higher scores indicating more violent behavior) (11) in the 30 days before enrolling in the program and at follow-up. Community adjustment was evaluated with the Global Assessment Scale (GAS) (12), and selected items from the Lehman Quality of Life Inventory (13) assessed satisfaction with life in general, with living arrangements, with friendships, and with family relationships. Role functioning was assessed with the Instrumental Activities of Daily Living (IADL) (14). Veterans were also asked whether they had been arrested or spent a night in jail for any reason during the previous six months. To assess treatment processes, case managers completed semiannual structured summaries of MHICM services provided to each veteran. We present information from the first of these assessments documenting the frequency of face-to-face contacts with the veteran in the community (scored from 1, once a month or less, to 5, three times per week); the percentage of face-to-face contacts that took place in the community (scored as 1, 0%–20%; to 5, 81%–100%); total time spent with the veteran weekly (1, less than 30 minutes per week, to 5, more than 120 minutes per week); and both distance from MHICM staff offices to the veteran’s home (1, less than one mile, to 8, greater than 50 miles) and travel time from MHICM staff offices to the veteran’s home (1, less than one minute, to 6, more than 60 minutes to drive). General types of services that were provided during the six-month period (for example, rehabilitation, psychotherapy, crisis intervention, management of psychiatric medications, screening for medical problems, and provision of substance abuse treatment, housing, and vocational support) were documented with a series of dichotomous variables. Case manager and veteran ratings of therapeu-
o ps.psychiatryonline.org o July 2010 Vol. 61 No. 7
tic alliance were assessed using a 7point Likert scale with questions about the nature of the clinical relationship adapted from the Working Alliance Inventory (15). Participants were also asked to rate their general satisfaction with VA mental health services. Clinicians were required to document termination when it occurred, using a structured progress report. Termination was said to have occurred when all efforts to engage the veteran had failed or when the veteran moved or was placed in a longterm care facility, such as a nursing home, or expressed a desire to discontinue this service. Reasons for termination have been documented since FY 2005 and thus are available only for more recent samples. Analysis We compared sociodemographic characteristics, clinical characteristics, and community adjustment of MHICM clients who terminated early (before one year), who terminated later (one to three years), and who did not terminate during the study period (more than three years). Chi square analysis was used to test the significance of differences in categorical measures, and analysis
of variance was used for continuous variables. To decrease risk of spurious associations because of the large number of variables and sample size, the level of statistical significance for bivariate comparisons was set at p
