bipolar disorders), the acute phase involves psychosis and disorganized or bizarre behavior. Hospitalization and medication are the appropriate treatment. The stable phase involves several key impairments: vulnerability to stress, deficient life skills, difficult interpersonal relationships, and inability to generalize skills learned in one setting (such as the hospital) to another (the community). Stein and Santos devised ACT to target precisely these impairments via continuous support and the teaching of life skills in the immediate contexts of patients’ lives. ACT also delivers pharmaceutic treatment directly to people’s homes in order to prevent relapse.
Stein and Santos endorse the now dominant paradigm of contemporary American psychiatry. Mental illnesses are brain diseases: discrete, organic malfunctions, each with its characteristic signs, symptoms, course, disabilities, and appropriate medications (see Luhrmann 2000 and Lakoff 2005). They write their ACT handbook as clinicians first and foremost, not critics of deinstitutionalization. The revolving door pattern is simply bad medical practice that treats the acute episode but neglects the stable phase of a chronic but manageable disease. Treatment during the remission phase is critical because what happens between episodes affects the severity and frequency of relapse. This rationale anchors their entire vision of Assertive Community Treatment. To help clients with community living, ACT teams must be multidisciplinary, comprising social workers, nurses, psychologists, psychiatrists, vocational specialists, and experts in drug and alcohol treatment. To prevent relapse and hospitalization, they must be ready to intervene around the clock, seven days a week. They must take ultimate responsibility for providing what clients need, when and where they need it, and not force them into 9-to-5 schedules and office appointments. In practice, this means that ACT staff members themselves deliver as many services as possible. When forced to broker services with outsiders, they must monitor and coordinate the relevant agencies. The 10-person ACT team is the fixed point of responsibility that ensures clients receive all necessary services from the fragmented nonsystem of public mental health.
The instructions for everyday work and the organization of ACT teams embody the biopsychiatric view of severe mental illness. Because predictors of relapse are subtle and unique to each person, clinicians must continually scrutinize clients’ lives to discern the symptoms, stressors or lifestyle changes that precede psychotic breaks or disorganization. Because the needs of clients can change quickly, everyone on the team must know every client: case managers share details of their own caseload with the entire group during morning staff meetings. Because of their poor transfer of learning among people with severe mental illness, clinicians must deliver services in the real contexts of clients’ lives: shelters, meal sites, clients’ apartments, and workplaces. Borrowing a metaphor from laboratory science, the ACT literature champions in vivo as opposed to in vitro services. Because of clients’ disorganization, anxiety, and poor coping skills, clinicians must act assertively to help clients benefit from treatment. If a client does not answer the phone, staff members drive out to his apartment. If he is not at home, they comb the neighborhood to find him. Clinicians continue to seek out clients even if they refuse contact for months on end. Finally, this model assumes that serious mental illness is chronic disease with lifelong impairments. Allness and Knoedler (2003: 397) provide sample discharge criteria, but they warn that clients should not be discharged for “traditional reasons” like needing less care or because their problems are too complex.
Assertive Community Treatment is the single most well-known model of mental health services that appeared after deinstitutionalization. A cadre of researchers continues to refine the approach, apply it to new populations, identify its core ingredients, and measure whether ACT teams on the ground adhere to the model’s standards.9 Even as a pilot program, it won the Gold Achievement Award from the American Psychiatric Association in 1974 (Dixon 2000). In 1996, the National Alliance on Mental Illness—the leading mental health lobby in the United States—officially endorsed ACT and began to pressure federal and local officials to make it available nationwide (Allness and Knoedler 2003: xi). In a coordinated effort, the federal Substance Abuse and Mental Health Services Administration (SAMHSA) proclaimed ACT one of six approved evidence-based treatments for severe mental illness and devised an implementation toolkit for clinicians and patients. It has won recommendation from the U.S. Surgeon General’s Report on Mental Health; it has its own professional association in the United States and currently influences clinical practice internationally (Burns et al. 2001).10
The diffusion of ACT illustrates a key tension in community psychiatry: between the need for improvisation and practical knowledge, on the one hand, and the influence of formal models and regulation on the other. ACT began as an experimental response to the failures of deinstitutionalization. It was nurtured by psychiatrists and social workers with relatively elite academic standing. The treatment norms they devised were operationalized, published in manuals, and passed on to lower-status frontline workers. The model was implanted across the United States in settings with fewer resources and dramatically different social problems than a Midwestern college town. The ongoing implementation of ACT, of course, transforms daily practice despite attempts to ensure fidelity to the original principles, on the part of program evaluators and mental health policy makers. In the end, the infinite adaptations of the model and the top-down attempts to standardize and control clinical work place significant strains upon frontline staff.
Some of the adaptations took place at the organizational level. By 1995, several states had reformed their public mental health services in line with the ACT model (Deci et al. 1995). Local mental health administrators received consultation and training from the ACT founders, demonstrated cost savings, and persuaded state legislators and mental health departments to scale up services. Newly organized teams inevitably differed from the ideal model program. After all, they were not constructed de novo, but instead by retraining people already employed in the mental health system. To build up new teams, the ACT consultants enlisted outreach workers and alcohol counselors who actually knew very little about chronic psychiatric illness or the community support model. ACT planners took staff from the shrinking state hospitals or from traditional case management programs where clinicians had caseloads of up to 60 clients (instead of the ACT recommendation of 10). The earliest ACT teams easily drifted away from the original standards; clinicians continued to see the majority of clients in their offices during normal business hours or simply did not try to engage with people who refused contact. In response, ACT planners encouraged work teams to distance themselves from the daily operation of the surrounding agency. Teams started to have separate staff meetings and record-keeping systems, to develop their own network of housing options, and generally to develop their own organizational identity.11
The sheer expense of ACT also influenced how treatment ideals got translated into workable programs on the ground. Like so many community medical services established in the United States since World War II, ACT programs faced a crisis at the end of the demonstration period (cf. Starr 1982: 370). State mental health authorities turned immediately to Medicaid as the new source of funding, and their strategy had important long-term effects. First of all, Medicaid reimbursements cannot actually pay the entire cost of a full-fledged ACT program. Under Medicaid’s lower funding, a bottleneck appeared in the public mental health system, with doctors and social workers pleading for the few open spots. In 1992, therefore, step-down programs—less intensive, with fewer services and higher caseloads—were created for high-functioning clients. At the same time, administrators tightened admission criteria to reserve ACT slots for the most severely disabled and most often hospitalized people. The shift to Medicaid essentially retooled ACT programs to admit the most difficult clients and to encourage the healthier ones to leave sooner. Clinicians found their caseloads filling up with sicker and more disabled individuals.
Secondly, the ACT model fits poorly with the basic Medicaid categories for reimbursement, because it combines health and social services in a single delivery technology (Lewin Group 2000). For example, delivering in vivo treatment may involve several trips until a case manager finds her client, but most of this travel cannot be reimbursed. Teaching people how to shop or write a household budget falls entirely outside Medicaid’s definition of medical treatment. The recently implanted ACT teams came under pressure to tailor their services to the balky logic of Medicaid. This problem was solved in a
