Impact of Flexible Duration Assertive Community Treatment: Program Utilization Patterns and State Hospital Use.

The Program of Assertive Community Treatment (PACT) model of care developed in Madison, Wisconsin has been emulated by many mental health providers in cities and towns across America, in Canada and abroad. Evaluation studies conducted in the past 25 years have demonstrated that this approach can significantly reduce the utilization of inpatient psychiatric care by people with serious mental illness (Bond, Miller, Krumwied, & Ward, 1988; Stein & Test, 1980; Witheridge, Dincin, & Appleby, 1982). A study of six assertive community treatment (ACT) programs in northeastern Indiana found that psychiatric hospitalizations were reduced by one third and inpatient bed-days reduced by one half (McGrew, Bond, Dietzen, McKasson, & Miller, 1995). There is little doubt that the model can be exported successfully to other locales.

Articles describing positive outcomes by programs that have adapted the PACT model for special populations are also available. Witheridge (1990) explains the successful development of the Thresholds Bridge to help people with the most serious forms of mental illness. Morse et al. (1997) found ACT services superior to a brokered case management approach for people who were homeless and had serious mental illness (SMI). Meisler, Blankertz, Santos, and McKay (1997) evaluated ACT for people who were homeless and suffered from co-occurring severe psychiatric and substance use disorders. While abstinence levels and social benefits were not high, the authors did report high rates of retention in treatment, housing stability and community tenure.

Evidence of the PACT model's effectiveness has also led to large scale implementations of ACT services by local and state mental health authorities. Because the model is so dependable, ACT services were strongly encouraged and largely adopted by all of the 18 service sites participating in the Access to Community Care and Effective Services and Supports (ACCESS), a national research project funded by the Center for Mental Health Services (Winter & Calsyn, 1999). During a four-year intake period, the ACCESS service sites have successfully engaged more than 7,000 homeless people with co-occurring mental illness and substance use disorders (Rosenheck, Lam, & Chinman, 1999).

With response to this model so favorable and the demand for replication strong, the National Alliance for the Mentally Ill (NAMI) sponsored a "how to" manual to guide practitioners in the start-up of PACT-style programs (Allness & Knoedler, 1998). Early in this book, the reader is introduced to the concept of "continuous, long-term services" which is described as one of five unique characteristics of a PACT program. Supporting this principal are results from the original PACT clinical trial and early results from an on-going, random longitudinal evaluation of 122 young adults with schizophrenic disorders. During the first study, people discharged after 18 months of services returned to their previous levels of hospital use in the following 14 months. Early results of the longitudinal study show that PACT enrollees consistently spend less time in hospitals and jails, experience less time homeless, and show other positive psychosocial gains as compared control subjects (Test, Knoedler, Allness, Kameshima, & Burke, 1994).

The literature has carried a "continuity of care" theme forward over time. Witheridge (1991) lists "time-unlimited" services as one of 12 "active ingredients" of the Thresholds Bridge model in Chicago. Providing long-term services became a core measure in an ACT fidelity scale developed by Teague, Bond, and Drake (1998). Higher scores are given to programs that provide services indefinitely and lower scores are given to those that time limit care.

While long term service has been a popular battle cry, the reality of limited mental health funding and competing mandates of funding agents has forced some ACT providers to compromise on length of service. In contrast to the negative impact of time-limited service reported in the original PACT study, one evaluation found good results. McRae, Higgens, Lycan, and Sherman (1990) examined the effect of terminating service after an extended period of the treatment. After five years of intensive case management, subjects were transferred to a community mental health center (CMHC) operated by the same agency. Hospitalizations during the ensuing two years increased but did not reach significant levels while contacts with the CMHC increased significantly. The authors concluded that the relative stability of the subjects following transition to mainstream services and the ability to offer intensive case management to a larger number of people over time was a win-win situation.

Dincin et al. (1993) described Thresholds Bridge Southwest, a program that also worked to transition its ACT clients in order to help a greater number of people. However, the relative short study period did not allow a comparison of outcomes as a function of time. This paper makes such a comparison by examining long term outcomes of the 64 people in the Dincin cohort.

Background

Development of the Thresholds Bridge Model

The first Thresholds ACT program was started in 1978 as a collaboration between the Illinois Department of Human Services, Office of Mental Health (OMH) and Thresholds, a private non-profit psychosocial rehabilitation center in Chicago. With funding from the National Institute of Mental Health, the 3-year mandate was to locate and engage 50 people who had frequent state hospital admissions in a north-side neighborhood. The OMH generated a list of the most frequently hospitalized people from the target area and instructed the project staff to systematically search for the highest users of inpatient days.

The staff assigned to the new Thresholds Bridge initiative traveled to Madison, Wisconsin to learn about the PACT approach. Many PACT ideas were adopted wholesale and other features were adapted to fit the complex urban service system in Chicago. Initial results were encouraging. Hospital utilization was cut in half for 41 people who completed a full year of services.

This success led to replications in other Chicago neighborhoods. Thresholds also developed ACT teams for people who are hard of hearing or deaf and have SMI and people who are homeless with SMI. The impact on these populations was consistent with the first project and by the mid-eighties a Thresholds Bridge "model" had emerged. Twelve hallmarks of this approach cited by Witheridge (1991) reflect the PACT influence and its Chicago roots. Important attributes of the model included an explicit mission to prevent the use of psychiatric hospitals, an exclusive reliance on in vivo methods, a robust staff to member ratio, the use of a team approach and tenacious advocacy across system boundaries. The importance of its "no close" policy was also stressed as a core feature.

Pressure to Change to the Thresholds ACT Model

In 1986, overcrowding at one of the state hospitals prompted a funding initiative for new community-based programs. By this time, evaluations of Thresholds Bridge programs had proven that the model could lower the hospital admissions and bed-day utilization by program participants. However, the OMH wanted the new project to show an impact on bed days at the catchment area level. Faced with limited new dollars, Thresholds and OMH designed a flexible duration ACT model. The project would abandon the "no close" policy but maintain a high staff to member ratio by providing intense services over a shorter period of time.

The Community Area

The target community is an old neighborhood named Back of the Yards, so called due to its proximity to the stockyards. Prior to the introduction of the ACT team, a city-operated CMHC was the only mental health provider. The ethnic composition of the area has varied little over time. The total population is about 140,000 and the ethnic mix is 58% Caucasian, 30% Hispanic, and 12% African American. Socioeconomically, the area is largely middle to lower middle class. The numerous factories and warehouses in the area are the primary employers. Housing is primarily single family homes; there are no supportive housing programs in the area.

Overview of the Resulting ACT Project

A six person ACT team was created. As in other similar programs operated by Thresholds, all staff shared responsibility for the team's caseload, including the project supervisor who dedicated half of his time to direct service. This project differed from other Thresholds programs because this team would be responsible for evaluating all area residents admitted to the state hospital and would be participating in discharge planning for people volunteering for ACT. Therefore, one staff position was dedicated to hospital-based duty. This staff assignment left 4.5 positions for community-based ACT interventions. Because the team strove to maintain a 1:10 staff to member ratio for ACT supports, the program capacity was 45 people.

A specific limit for ACT tenure was not applied to the participants. Instead, the project staff were expected to constantly evaluate the need for continuing service. A review panel of high level administrators from the OMH, Thresholds, the state hospital, and the CMHC aided the process of targeting people for transition to mainstream services at the CMHC. The panel met monthly during the first year and bi-monthly for the duration of the study. This review process kept pressure on the program to serve the highest number of people possible.

The approach to helping a new member leave the state hospital and avoid readmission was identical to that employed by other Thresholds ACT programs. In fact, a veteran of other Bridge programs supervised the team. The typical member received regular home visits from ACT staff. Program interventions focused on both the concrete aspects of community living and requisite psychiatric care elements. Examples of community living supports include help to secure entitlement funding, acting as representative payee, performing money management tasks, and finding and keeping housing. Psychiatric care typically included connecting people to a psychiatrist, picking up prescriptions, monitoring compliance to prescriptions, and facilitating patient/doctor communication.

Concurrent with ACT interventions at the person level was an ongoing brainstorming at the program level. This brainstorming identified opportunities to facilitate the transition to mainstream services. The ACT staff worked with CMHC staff and doctors to establish firm linkages, often continuing concrete supports for a period after CMHC services had begun.

A Philosophical Change

Faced with a high demand for a limited resource over an extended period, this program had to continuously search for ways to make scarce ACT slots available to new referrals. Departing from an agency policy to provide services on a "time-unlimited basis" (Witheridge, 1991), this program redefined its commitment to people by promoting the concept of "once a member, always a member." To keep the relationship viable, inactive members were invited to holiday celebrations and other member gatherings. In addition, members, their families, and other important support agents were strongly encouraged to alert the ACT team if difficulties arose. In all cases, full ACT support could be instantly reinstated if problems surfaced.

Method

State hospital records for the year prior to ACT program intake and for each of the 10 years following intake were examined for each of the subjects. To analyze hospital utilization rates over time, each subject functioned as his or her own control. Admissions and bed-days for the year prior were compared to the same statistic for the second, fourth, sixth, eighth and tenth years. A new electronic records system assured that all state hospitalizations were captured, even hospitalizations for people who traveled outside of the original community area. When members died, only complete years were used in the analysis.

After this review of the hospital utilization data, the information was segregated according to history of ACT utilization. Client records were reviewed to calculate the number of ACT service episodes and the length of the episodes. Three subgroups emerged as a result of this scrutiny. The first was a "never closed" group. These people received uninterrupted ACT services for the entire 10-year period because the program staff never considered termination of ACT support to be a practical option. The second group was the "multiple service episode" group. These people were transferred out of ACT and later reopened one or more times. The third group was the "single service episode" group. These people were transferred and did not return to ACT before the end of the study period.

At the end of the period of study, telephone interviews were conducted to confirm the status of the people not active in the program. Member vignettes were prepared to illustrate how ACT supports functioned for people from the three subgroups.

Results

Subject Mortality

Eleven of the 66 subjects died, two in the second year of the study, two in the third, two in the fourth, one in the fifth, two in the seventh, and one in the tenth. Because two members died before two-year mark, their data were dropped from all comparisons. Subsequent deaths resulted in a declining number of subjects, with 60, 59, 56 and 55 people for years 4, 6, 8 and 10 respectively.

The mean age at time of death for these members was 44.7(SD=12.9) years. There were no suicides. The causes of death were accidental drug overdose (2), cirrhosis of the liver (1), traffic accident (1), cancer (1), and heart attacks or heart failure (6).

State Hospital Utilization Over Time

Table 1 presents hospital use during the year before ACT intake and for even numbered years after enrollment. The reduction in the number of hospitalizations was significant at all comparison points, as was the reduction in the number of days hospitalized. Further analysis revealed that 25% of subjects (n=16) never returned to the state hospital. Of a possible 640 post-enrollment years, 438 (68%) were hospital free; 89 (14%) were years in which subjects used fewer hospital days as compared to the year before ACT; 66 (10%) were years in which more hospital days were used; and 47 (7%) were years subsequent to member death.

Table 1

State hospital utilization by ACT members during a 10 year period.

                     Times Hospitalized      Days Hospitalized

Period               Mean         SD         Mean   SD

Year before (n=64)    2.4         1.59       79.3   76.6
Year 2 (n=64)(a)      0.4         0.89       28.3   70.2
Year 4 (n=60)(b)      0.6         1.05       26.2   68.5
Year 6 (n=59)(c)      0.3         0.68       28.7   86.8
Year 8 (n=56)(d)      0.2         0.46       27.0   79.9
Year 10 (n=55)(e)     0.1         0.59       22.9   77.7

(a) For times hospitalized difference from Year before intake, t = 9.268, df = 63, p [is less than] .0001. For days hospitalized difference from year before, t = 4.437, df = 63, p [is less than] .0001.

(b) For times hospitalized difference from Year before intake, t = 7.647, df = 59, p [is less than] .0001. For days hospitalized difference from year before, t = 4.690, df = 59, p [is less than] .0001.

(c) For times hospitalized difference from Year before intake, t = 8.841, df = 58, p [is less than] .0001. For days hospitalized difference from year before, t = 4.006, df = 58, p [is less than] .0001.

(d) For times hospitalized difference from Year before intake, t = 9.990, df = 55, p [is less than] .0001. For days hospitalized difference from year before, t = 4.329, df = 55, p [is less than] .0001.

(e) For times hospitalized difference from Year before intake, t = 9.121, df = 54, p [less than] .0001. For days hospitalized difference from year before, t = 3.770, df = 54, p [is less than] .0001.

ACT Service Episodes

The 64 members included in the long-range study accounted for a total of 125 service episodes. Twelve people (19%) fell into the never closed group and had continuous ACT services throughout the study period. Seventeen people (27%) were transferred and did not return to ACT and constituted the single episode group. The remaining 35 members (55%) made up the multiple episode group. Subjects in the latter group had an average of 2.8 service episodes with one person experiencing six episodes over 10 years. No significant differences in age, race, gender, psychiatric diagnosis, co-morbidity, or the prior hospital utilization rate were found between these groups.

ACT Utilization as Compared to Hospital Utilization

Table 2 presents key statistics for program utilization and state hospital utilization for the entire cohort and for the three subgroups. As compared to the mean scores for the entire cohort, never closed subjects averaged fewer hospital admissions and fewer hospital days; single episode members averaged fewer admissions and more hospital days; and multiple episode members averaged more admissions and more hospital days.

Table 2

Mean scores for ACT utilization and state hospital utilization over 10 years.

                                   ACT         State Hospital

Service Group              Episodes   Months   Admissions    Days

Total cohort (n=64)          2.0       49.4       3.6       258.9
Never closed (n=12)          1.0       94.1       3.0        56.9
Single episode (n=17)        1.0       17.9       2.5       314.4
Multiple episodes (n=35)     2.8       49.4       4.3       301.1

Status of Subjects at End Date

Of the 55 members still living at the end date, 22 (40%)were actively enrolled in the ACT program. Twenty-four people (44%) were residing within the geographical boundaries of the catchment area, 29 (53%) were living in other Chicago area communities, and two (4%) were living out of state.

Forty-four (80%) were living independently in the community, five of whom were gainfully employed and maintaining a linkage to psychiatric services. By contrast, eight people (15%) were in skilled care facilities for mental health or medical reasons, 2 (4%) were in the state hospital, and one person was incarcerated for grand theft auto.

Discussion

The first year reduction to state hospital utilization, originally reported by Dincin et al. (1993), continued at significant levels during the balance of this 10-year study. As compared to their pre-intake year, program members used 66% fewer days during each subsequent year.

The results of this study support the value of continuous ACT services and the feasibility of time-limited interventions as well. In agreement with Stein and Test (1980), the greatest reduction in this study was recorded for those people whose ACT support was not interrupted. Similar to McRae et al. (1990), who concluded that intensive community support programs need not go on forever, people carefully moved to mainstream services did maintain significant reductions in state hospital utilization rates. However, a third important option emerged; ACT provided on an intermittent basis over time can also be effective. Vignettes help illustrate these central points.

Paul is a program member whose ACT support was continuous throughout the study period. He is a 31-year old, single, white male with a diagnosis of schizophrenia, paranoid type. Paul has a history of 10 state hospitalizations and irregular contact with community-based services. When residing with his mother, he would collect, but not cash, Social Security checks; would miss doses of his medications, eventually discontinuing them; and would completely unravel over a period of months.

After intake, the ACT team became Paul's payee, helped him find and maintain housing, and remained directly involved in his relationship with a psychiatrist and the management of his medications. Multiple attempts to withdraw elements of this support package failed. Thus, from time to time, Paul was readmitted at the state hospital. However, compared to the six admissions and 111 bed days used in the 12 months before intake, he experienced six admissions for 167 days during the following 10 years. This reduction is about 85%.

Near the other end of the spectrum, Beth is an example of the careful withdrawal of ACT service and longevity of linkage to mainstream services. Beth is a 27-year old, African American, mother of one who was diagnosed bipolar, manic type. She had a history of two state admissions and resistance to help post discharge. Initial ACT engagement was tenuous and would have failed if not for her desire to maintain custody of her daughter. Along these lines, the ACT team became a vehicle for linkage to a psychiatrist which the supervision plan of the Department of Children and Family Services required. ACT staff became a buffer between Beth and the psychiatrist with whom she argued vociferously.

Over the course of four years, Beth's relationship with the doctor became positive. She was able to complete mandated parenting classes, find employment, and move into an apartment in the suburbs of the city. After two admissions for 38 days in the state hospital during the year prior to intake, Beth had two admissions for 37 days over 10 years. This reduction is 90%.

Leah is an example of a transition that quickly floundered. A 32-year old, married, Hispanic mother of three, Leah had a diagnosis of schizophrenia, paranoid type. Her prior pattern of state hospital use included 13 admissions and little or no contact with the CMHC after discharge. Complicating matters were strong language and cultural barriers. With some bilingual staff, the ACT program linked her to the CMHC and focused on medication education and support. After three years without hospitalization and with a solid link to the CMHC, the ACT outreach was withdrawn. Two months later, Leah was re-engaged by ACT after missing two consecutive doctor's appointments. She was not closed from ACT again. As compared to one admission and 82 bed days in the year before intake, Leah experienced nine admissions for 424 days the following decade. This reduction in days is 48%.

Sam is an example of a how multiple service episodes actually facilitated a continuous care scenario. A 32-year old, African American male, Sam was diagnosed with bipolar disorder, hypomanic, and had a history of tour state hospitalizations. He steadfastly denied a mental illness and would accept ACT membership strictly on his terms. On three occasions, Sam discontinued medications and asked that workers "close my case." But the real surprise was that despite his rock solid denial of illness, Sam proved very willing and able to initiate contact with the team whenever symptoms emerged. While hospitalizations did occur, staff believe that many bed days were saved. Compared to one admission for eight days in the prior year, Sam had six admissions for 75 days in the following 10 years. Although, statistically, this reduction in days is only a 6%, it was achieved with only 29 months of active outreach services. At the close of his first decade, Sam is in the midst of his fifth outreach service episode.

Together, these cases illustrate how the philosophy of "once a member, always a member" and the decision to operate with a permeable program boundary lowered the thresholds at which staff were willing to risk helping members transition away from "active outreach." With this risk-taking came unexpected outcomes. Sam proved to have strengths never imagined by the staff and Beth managed to steer clear of huge hurdles despite fears to the contrary. Moreover, Leah demonstrates that if help can be immediate, this advantage does not come at a high price to those who are not as ready as ACT staff had predicted. Finally, Paul proves that uninterrupted treatment is possible in this environment as well.

Conclusion

Review of program utilization patterns and state hospital utilization rates revealed that 80% of the first 64 members served by this ACT program were moved to mainstream services without significant negative impact on the hospital utilization rates. About two thirds of people returned to the ACT program for a second or subsequent service episodes and one person had as many as six distinct ACT interventions. For the service system with high demand for ACT services but scarce resources, carefully moving people to mainstream services by an ACT team with a permeable boundary is a way to balance a mission to help people with SMI over extended periods with the system need to reach the greatest number of people. Investigation into the composition of support networks that kept people living in the community with little or no further ACT interventions could shed more light on this subject.

Author's Note

The authors would like to acknowledge Sandy Scala and Eileen Niccolai for their contributions to this manuscript.

References

Allness, D.J., & Knoedler, W.H. (1998). The PACT model of community-based treatment for persons with severe and persistent mental illness. Arlington, Virginia: NAMI Anti Stigma Foundation.

Bond, G.R., Miller, L.D., Krumwied, R.D., & Ward, R.S. (1988). Assertive case management in three CMHCs: a controlled study. Hospital and Community Psychiatry, 39:411-418.

Dincin, J., Wasmer, D., Witheridge, T.E, Sobeck, L., Cook, J., & Razzano, L. (1993). Impact of assertive community treatment on the use of state hospital inpatient bed days. Hospital and Community Psychiatry, 44, 833-838.

McGrew, J.H., Bond, G.R., Dietzen, L., McKasson, M., & Miller, L.D. (1995). A multisite study of client outcomes in assertive community treatment. Special section: Assertive community treatment. Psychiatric Services, 46, 696-701.

McRae, J., Higgins, M., Lycan, C., & Sherman, W. (1990). What happens after five years of intensive case management stops? Hospital & Community Psychiatry, 41, 175-179.

Meisler, S., Blankertz, L., Santos, A., & McKay, C. (1997). Impact of assertive community treatment on homeless persons with co-occurring severe psychiatric and substance use disorders. Community Mental Health Journal 33(2), 113-127.

Morse, G.A., Calsyn, R.J., Klinkenberg, W.D., Trusty, M.L., Gerber, E, Smith, R., Tempelhoff, B., & Ahmad, L. (1997). An experimental comparison of three types of case management for homeless mentally ill persons. Psychiatric Services 48(4), 497-503.

Rosenheck, R., Lam, J.A., Chinman, M.J. (1999, March). Lessons learned in ACCESS: Clinical process and client outcomes. Paper presented at the grantee meeting of the Center for Mental Health Services, Access to Community Care and Effective Services and Supports, Bethesda Maryland.

Stein, L.I., & Test, M.A. (1980). Alternatives to mental hospital treatment: Conceptual model, treatment program and clinical evaluation. Archives of General Psychiatry, 37, 392-397.

Teague, T.B., Bond G.R., & Drake, R.E. (1998). Program fidelity in assertive community treatment: development and use of a measure. American Journal of Orthopsychiatry, 68(2), 216-232.

Test, M.A., Knoedler, W.H., Allness, D.A., Kameshima, S., & Burke, S.S. (1994, June). Seven years of continuous community treatment of young adults with schizophrenia: Results and implications. Paper presented at 11th International Symposium for the psychotherapy of Schizophrenia, Washington, D.C.

Winter, J.P., & Calsyn, R.J. (1999, March). Case management implementation project: Interim report. Paper presented at the grantee meeting of the Center for Mental Health Services, Access to Community Care and Effective Services and Supports, Bethesda Maryland.

Witheridge, T.F., Dincin, J., & Appleby, L. (1982). Working with the most frequent recidivists: a total team approach to assertive resource management. Psychosocial Rehabilitation Journal, 5: 9-11.

Witheridge, T.F. (1990). Assertive community treatment: a strategy for helping persons with severe mental illness to avoid rehospitalization, in Psychiatry Takes to the Streets: Outreach and Crisis Intervention for the Mentally Ill. Edited by Cohen, N.L. New York, Guilford.

Witheridge, T.F. (1991). The "active ingredients" of assertive outreach. New Directions for Mental Health Services, no 52: 47-64.

Daniel Wasmer, M.S., Thresholds Assistant Director, 4101 North Ravenswood Avenue, Chicago, IL. 60613.