Managed Behavioral Health Care in the Public Sector: Will It Become the Third Shame of the States?

In the 1970s federal legislation and the desire to contain costs that were consuming an inordinate proportion of the U.S. gross national product launched managed approaches to health care. The generic term "managed care" is used to describe health care delivery that incorporates mechanisms to monitor and authorize service utilization. Managed care programs replace traditional fee-for-service arrangements with a negotiated, capitated payment for all program enrollees. Managed care arrangements in the private sector now cover the majority of Americans, and they recently have made inroads into publicly funded health care.

Managed care also permeates behavioral health care—that is, mental health and substance abuse services. According to some, public-sector managed behavioral health care is market oriented, efficient, and driven by accountability and thus can significantly improve service delivery. However, others have expressed grave concerns about the potential negative effects on quality of care and access to care, fearing that the profit motive will reduce or eliminate care for those least able to provide for themselves. Nearly 50 years ago, the era of deinstitutionalization produced similarly conflicting responses.

The purpose of this paper is to compare the similarities of deinstitutionalization and public-sector managed behavioral health care, in the belief that lessons from the past can guide the changes the mental health system is undergoing today. After an overview of the growth of managed behavioral health care and a historical analysis of deinstitutionalization, we compare the antecedents, benefits, and negative consequences of both movements. Finally, from lessons learned about deinstitutionalization, we offer recommendations for public policy as managed behavioral health care approaches are further developed in the public sector.

The growth of managed behavioral health care

In the past decade, the number of covered lives, or enrollees, in managed behavioral health care organizations has risen dramatically, nearly doubling from 1993 to 1996 (1). In 1999 a total of 79 percent of Americans with health insurance were enrolled in managed behavioral health care organizations (2). Many managed care organizations provide enrollees with mental health services through a carve-out vendor, or subcontractor, who in some cases also assumes insurance risks for these services.

As with managed physical health care, the inception of managed behavioral health care was driven by economic issues—insurers' and employers' growing concerns that mental health and substance abuse services were consuming ever-larger proportions of health care dollars (3,4). With the introduction of managed behavioral health care, this tide has been stemmed (4), so that mental health and substance abuse costs have progressively decreased as a percentage of health care expenditures (5).

Reports thus far indicate success not just in controlling costs but also in increasing access to care and achieving other positive effects. For example, a recent seven-year analysis of managed behavioral health care utilization in plans of 30 employers with nearly 60,000 employees found increased overall use of mental health services, greater provision of within-network care, and reduced long-term costs for behavioral and medical care (6). The researchers concluded that the mechanisms of managed care—prospective and concurrent clinical review, substitution of benefits, individualized treatment planning, provider networks, fixed rates of reimbursement, and use of intermediate levels of care—can be effective in containing costs over the long term. Cost-benefit analyses now include factors such as patients' improved functioning, improved work performance, and decreased disability and indirect costs along with the more traditional direct costs (7,8).

With its economic success and in anticipation of enrollment saturation in the private sector (1), managed behavioral health care has moved into the public sector. Medicaid funding has been increasingly applied to community-based treatments for persons with long-term, severe mental illnesses and now supports about one-third of the cost of such treatment (9).

The resultant rise in state Medicaid budgets is creating fiscal crises for some state governments. Consequently, many states have developed or are planning to implement managed behavioral health care for Medicaid and indigent clients with mental illness, substance use disorders, or both (10). In the mid-1980s, the federal government began approving waivers for states to develop new market arrangements for Medicaid. As of January 1999, a total of 47 states were redesigning their Medicaid programs, up from 14 in 1996 (11). In 1998 a total of 54 percent of Medicaid recipients were enrolled in some form of managed behavioral health care (1).

However, clinicians and analysts express concerns about the applicability of managed care approaches to public mental health care recipients and their problems because of historical differences in treatment of behavioral and somatic health disorders. Differentiating factors include the stigma of mental illness, the lack of established treatment effectiveness, and the higher rates of undiagnosed and untreated mental illness (12). Hogan (13) has noted that "the public mental health system is the only substantial, disorder-specific treatment system in existence today." It is also the only remaining state-government-supported treatment system for a specific set of disorders.

The appropriateness of managed care models for publicly funded mental health and substance abuse services has also been questioned. Whereas the standard optimal insurance model provides coverage for catastrophic events after first-dollar coverage, current carve-outs for mental illness and substance abuse often increase access to entry-level care but constrain intensive or frequent services. Application of this model to publicly funded consumers, for whom cost-sharing is less of an option, raises serious questions about the model's relevance (14).

Such concerns hark back to deinstitutionalization and to the tragic and inhumane consequences experienced by psychiatric patients and their families because of a policy that was conceptually flawed and bankrupt in its implementation. As we move forward in the era of managed behavioral health care, we need to ask whether this shift in policy and practice will achieve the intended results or, like deinstitutionalization, will produce more negative effects.

The deinstitutionalization movement

Deutsch's 1948 book, The Shame of the States (15), contained articles originally written for the newspaper PM on the abuses found in 20 state mental hospitals. The exposé documented the outrageous conditions in words and pictures: patients crowded into day rooms; patient census two to three times above bed capacity; old hospital buildings that were vermin and rat infested and condemned; patients malnourished and unclothed; and patients needlessly restrained or exploited as free labor, working 12 to 14 hours a day. The book also described physician staffing ratios as low as one doctor to 800 patients and ward staffing at less than 20 percent of recommended levels. The result was not just patient warehousing but also risks of accidents, physical abuse, and premature death.

The Shame of the States undoubtedly contributed to the deinstitutionalization movement, which began less than ten years later. Deinstitutionalization, usually perceived as a failure, has been labeled the second shame of the states (16). As originally formulated, deinstitutionalization was broadly conceptualized and borne of a noble policy intent: to improve the situation of persons with serious mental illness by replacing total institutions with community-based services, using small facilities and neighborhood environments (17). A 1977 General Accounting Office report defined deinstitutionalization as follows: "The process of 1) preventing both unnecessary admission to and retention in institutions; 2) finding and developing appropriate alternatives in the community for housing, treatment, training, education, and rehabilitation of the mentally disabled who do not need to be in institutions; and 3) improving conditions, care, and treatment for those who need institutional care" (18).

In practice, however, deinstitutionalization was translated simply to mean reductions in the census of public mental hospitals. Its magnitude can best be appreciated through statistics. The resident population of state and county mental hospitals had increased steadily since the early 1900s and peaked in 1955 at 558,922 patients, or half the patients in all hospital beds in the United States (19). By 1980, the number of residents was one-quarter of its previous high. By the mid-1980s, the inpatient census had dropped to 115,000 patients (19).

Interestingly, the number of public mental hospitals stayed fairly constant during this period. While overall census decreased by about 5 to 10 percent a year, admissions actually rose through 1970 and then declined (20). The major effect of deinstitutionalization was on the number of beds per state hospital and the length of stay. Although this fact is not routinely recognized, deinstitutionalization continues today, with downsizing and closure of state hospitals.

Comparing antecedents of both movements

The process of deinstitutionalization was driven by a confluence of social forces, both conservative and liberal. The factors most frequently identified are dissatisfaction with existing mental health care, consumerism, social science research and ideologies, improved treatment technologies, economic burden, and new kinds of funding (19,21,22,23). In this discussion, we describe each factor and then examine whether similar factors are operating in today's managed behavioral health care environment.

As this discussion and later analysis of the impact of deinstitutionalization and of managed care suggest, we believe that managed behavioral health care as it is now developing in the public sector is in danger of becoming the third shame of the states. This concern is especially strong for patients with severe and persistent mental illness.

Factors contributing to deinstitutionalization

Dissatisfaction and consumerism. An important driving force behind public support for deinstitutionalization was undoubtedly a long history of dissatisfaction with mental institutions. Deutsch's exposé (15) was congruent with personal accounts of mistreatment, such as Mary Jane Ward's The Snake Pit (24), later made into a motion picture. The 1950s and 1960s were an era of general concern about civil rights and of initiatives supporting consumerism. Victories in litigation involving minorities bolstered interest in the rights of other disadvantaged groups, such as the mentally ill (23).

Social science research and ideologies. Adding to more empathic public interest in mental illness were ideologies and social science research of the day. Popularized works by Erving Goffman, Thomas Szasz, and R. D. Laing promoted the belief that mental illness is a myth constructed by society to marginalize or reject people who are different and that psychiatric diagnoses function to reify and legitimate the existing social order (25).

Deinstitutionalization was also supported by liberal and humanist ideologies, through the New Left and the 1960s counterculture. The former perceived mental illness as a failure to adjust to societal demands for repression, and the latter celebrated mental illness for its perceived similarity to drug-induced states, thought to yield a higher reality (16).

In this antipsychiatry movement, mental illness, if it existed at all, was thought to be created by the social institutions designed to cure it. Eliminating mental hospitals would therefore eliminate mental illness. Psychological theories recognizing the influence of parenting practices on child development and adult health also contributed to the rejection of illness and of hospital treatment models. The identification of psychiatric trauma due to "shell shock" among World War II veterans further supported beliefs that environmental factors were prime contributors to mental illness (19,25). According to the mental hygiene movement, early interventions with families, schools, and communities could prevent mental illness so that in the future hospitals would not be needed.

Improved treatment. Another major factor linked to deinstitutionalization was improved treatment technologies. The discovery of medications to treat psychotic conditions is popularly thought to be the major cause of deinstitutionalization. However, Johnson (21) pointed out that wide-scale use of neuroleptic drugs had an economic as well as a clinical basis. That is, entrepreneurial pharmaceutical companies invested heavily in marketing strategies aimed at getting state legislatures to increase hospital drug budgets. Medication-based treatments were attractive to fiscal conservatives, because they promised to reduce institutional costs. Mechanic (26) has concluded that although the introduction of neuroleptic drugs in the 1950s was not the cause of deinstitutionalization, it is unlikely that social forces could have reduced state hospital censuses without these drugs.

Economic burdens and new funding. The most significant factor in the process of deinstitutionalization was probably economic. Before deinstitutionalization, the financial burden of operating public mental hospitals was borne mainly by state governments. By 1955, state psychiatric institutions were consuming politically indefensible portions of state revenues—for example, 38 percent of New York State's budget (16). State officials were in the unenviable position of spending more and more money on a "service" that was rejected as inhumane by the public and as unnecessary by experts.

At the same time, new federal funding initiatives emerged. When patients were discharged from state hospitals to small residential facilities or to nursing homes, programs like Supplemental Security Income (SSI), Social Security Disability Insurance (SSDI), Medicaid, and Medicare enabled states to shift the funding burden to federal sources. Grob (25) has noted that hospital reductions after the introduction of new psychotropic drugs were actually quite modest; the major impact on hospital censuses followed the passage of Medicaid and Medicare legislation in 1965.

Factors contributing to managed behavioral care

Many factors that contributed to deinstitutionalization have resurfaced with visible influence on current mental health policy.

Dissatisfaction with mental health services. Since the 1980s there has been a growing belief that community-based care is failing to meet the needs of persons with serious mental illness. For example, as many as one-third to one-half of the homeless population have long-term mental illness (27), as do 6 to 15 percent of persons in jails and 10 to 15 percent of those in prisons (28). Because of inadequate community mental health services, severely mentally ill people are increasingly referred to the criminal justice system (29,30), a situation that has fostered jail diversion programs to channel those with mental illness to the mental health system (31). This new "institutional circuit" for persons with mental illness consists of stays in homeless shelters, jails, or prisons alternating with short-term psychiatric hospitalizations (32).

Advocates have raised awareness of abusive treatment of mental patients in residential and inpatient psychiatric care. Unnecessary deaths resulting from illegal use of restraints reportedly number 50 to 100 a year, but may be drastically underestimated (33). The failure of postdeinstitutionalization mental health policies has been reiterated in a fairly constant stream of media accounts of violence committed by persons with severe mental illness. A recurring feature is that many of the offenders were allegedly under the care of, or at least known to, local mental health professionals.

Improved technology to treat mental illness. Ironically, in spite of public attitudes deeming treatment of severely mentally ill people a failure, there is renewed optimism that technological advances in biochemistry and psychosocial rehabilitation are positively affecting their lives. Although conventional neuroleptic medications were useful in the treatment of schizophrenia and other psychotic disorders, they left up to 30 percent of patients without clinical improvement (34). Pervasive side effects limited patient compliance, worsening long-term relapse rates.

In contrast to these older medications, a new generation of psycho-pharmaceuticals has been developed: clozapine in 1989 and risperidone in 1994, followed by olanzapine, quetiapine, and ziprasidone. Besides minimizing side effects, these new atypical antipsychotics promise effective treatment for more patients, with a broader spectrum of efficacy (35). With these advances, combined with newer selective serotonin reuptake inhibitors for depression, the market for psychotropic drugs quadrupled between 1994 and 1998 (36).

Biochemical advances led the National Institute of Mental Health (NIMH) to declare the 1990s "the decade of the brain." But at the same time, psychosocial interventions have been affirmed as necessary and effective components of community integration for persons with severe mental illness (37,38), mediating the clinical effectiveness of medication (39). The modality of assertive community treatment has established itself as an effective intervention and is now replicated nationally (40). Positive outcomes have been demonstrated for other rehabilitative techniques, such as clubhouses (41), case management (37), psychiatric rehabilitation (42), supported education (43), and supported employment (44). Further, evidence is increasing that mental health treatments can generate higher productivity and income for their recipients (45) while offsetting expenditures in general medicine (46,47) and acute care (48,49).

Expanded consumer and family involvement. Movements involving mental health service recipients and their families have substantially expanded in the past two decades. From 1988 through 1992, NIMH funded 13 three-year, consumer-operated demonstration projects. As of 1996, a total of 35 state mental health agencies offered direct funding for at least one consumer-operated program (50). At present we see a visible and increasing emphasis on consumerism in psychosocial rehabilitation programs, on administrative requirements for person-centered planning, and on legislative initiatives for rights protection and advocacy. For instance, 29 states have statutory or regulatory mandates for consumer participation in program planning (50).

Social science research. Research has now identified the comprehensive community-based services fundamental to meeting the needs of persons with severe mental illness. Community support systems and a new understanding of severe mental illness support the possibility of recovery (38). Replacing the view of severe mental illness as a chronic condition with little variation, research now suggests that the outcomes of long-term care for disabilities such as schizophrenia differ along a continuum from incapacity to complete recovery (51,52).

Budget concerns. The 1980s witnessed a noteworthy increase in public skepticism about government spending and a corresponding decrease in state and federal funding for human services. Reflecting goals to improve U.S. competitiveness in a global economy, concerns arose about escalating expenditures for health care as a proportion of government and employer costs and about expansion of mental health and substance abuse benefits as a proportion of national health expenditures (9,53).

By 1990, health care expenditures constituted about 12 percent of the gross national product (10). Between 1987 and 1992, Medicaid spending rose 12 percent annually—five times the rate of inflation (9,10). Expenditures for Medicaid increased from 8.1 percent of state budgets in 1987 (54) to 20 percent in 1993 (55), a proportion larger than that typically spent on higher education or law enforcement (56).

New funding arrangements. Concerns about rising costs for public mental health services have led to a movement to privatize care in an effort to control costs while ostensibly ensuring access to and quality of services. On the assumption that market forces can ensure efficiency, economy, and quality for public services, Medicaid has been restructured to reflect managed care arrangements used by private insurers. Moving from traditional Medicaid fee-for-service models to federal and state capitation funding limits the financial risk of the insurer—that is, the government—and shifts the cost burden to private and local sources, such as contractual managed care organizations, local governments, and families.

A major concern about privatization under Medicaid is that private insurers will focus on reducing costs rather than on meeting the needs of disadvantaged and chronically ill patients (56,57,58). Managed care arrangements have been shown to reduce expenditures for mental health services (59,60), which means that the previous costs were unnecessary, that they are being absorbed by other sources—for instance, families, private charities operating shelters, or local governments paying for jails—or that the needs are not being met.

It is an unfortunate coincidence that the current growth in knowledge about new and effective treatment technologies is occurring in a policy environment in which the primary goal is cost containment. Jackson (57) charges that while provision of health care to the most needy and disadvantaged was previously acknowledged as a public responsibility, government has now transferred this obligation as a business matter to the commercial marketplace.

Outcomes of deinstitutionalization

In response to the previously described social and economic concerns about mental illness and institutional care, Congress created the Joint Commission on Mental Illness and Health. Its 1961 report, Action for Mental Health, recommended upgrading state hospitals to therapeutic levels, increasing psychiatric treatment in general hospitals, and developing community mental health centers (CMHCs) (61) to divert persons with mental illness from hospitals and provide aftercare for those discharged yet incompletely recovered. The legislative response was the Mental Retardation Facilities and Community Mental Health Centers Construction Act of 1963.

However, the act and its associated regulations were conceptually flawed, emphasizing primary prevention and lacking adequate mechanisms or funding to improve conditions in state hospitals. The act also ignored the role of the states in monitoring mental health care. State authorities were thus still without legitimate means to address the problems of state hospitals.

Benefits of the movement

Before addressing the short-term failures of deinstitutionalization, which are readily visible, we will outline its long-term positive consequences, which are not often identified.

Over time, the use of outpatient care and inpatient psychiatric treatment in community hospitals increased (19). Deinstitutionalization eventually contributed to the development and expansion of innovative modalities such as psychiatric rehabilitation, clubhouse programs, and assertive community treatment; to more humanistic treatment through protection of rights; and to a social movement exemplified by advocacy activities of former patients, self-help groups, and consumer-run programs.

The era of deinstitutionalization saw a major shift away from public mental hospitals; examples of well-planned state hospital closures, accompanied by exemplary community-based treatment, can be found (62,63,64). For many patients, treatment was more appropriate and more responsive, preventing some of the secondary disabilities and iatrogenic conditions previously associated with "chronic" mental illness.

In fact, long-term data on mental health service utilization from 1970 to 1986 reflect the intended effects of deinstitutionalization. In 1986 only 24 percent of episodes were in inpatient facilities, compared with 77 percent in 1970; the number and size of public mental hospitals had decreased; and their resident population had been reduced by two-thirds (65). Finally, despite allegations of poor or unsafe community treatment, consumers almost uniformly prefer community residence to hospitalization (66). Mechanic (19) concludes, "Despite its failures, the CMH movement vastly improved living options for mental patients, and many patients lead more satisfying lives than they did in the past…. When all was said and done, these were no small achievements."

Negative outcomes and unintended effects

Although some analysts have rightly identified benefits of deinstitutionalization, most of the mental health field has seen it as a failed policy, or as a failure to implement policy. Many unintended, negative results have been attributed to deinstitutionalization: patients discharged before they were ready, transferred to inappropriate sources, or refused admission in order to decrease hospital use. The burden of care for these individuals consequently has been borne by other sectors, allegedly resulting in homelessness, transinstitutionalization, criminalization of mental illness, and increased family burden.

Homelessness. Homelessness has supposedly increased because released patients with no place to go and little capability to care for themselves are discharged to homeless shelters or end up living on the streets. A review of several studies concluded that a significant number of patients in mental institutions, particularly state mental hospitals, were or had been homeless and that a large proportion of discharged patients became homeless (67). Although behavioral problems obviously increase vulnerability to homelessness, only after deinstitutionalization was the presence of these problems so highly correlated with homelessness.

Transinstitutionalization. The phenomenon of transinstitutionalization, the movement of people between institutions, is alleged to affect older residents and would-be patients. They have been transferred directly to nursing homes, or diverted from state hospitals to nursing homes, through federal Medicaid or Medicare funding, and receive inadequate care for their serious psychiatric problems. For example, in 1950 a total of 40 percent of institutionalized elderly persons were in mental hospitals and 20 percent were in homes for the aged and dependent; by 1980, these proportions were 10 percent and 50 percent, respectively (68). Other older persons with mental illness have been sent to poorly staffed board-and-care homes, funded through SSI (69).

Criminalization. Criminalization of mental illness occurs when individuals with mental illness, unable to receive needed treatment through hospitals, disturb communities with bizarre behaviors or with illegal acts necessary for survival, such as loitering, stealing food, and breaking and entering to obtain shelter. Their resulting "treatment" is jail or prison (32,70). The lifetime prevalence of psychiatric disorders among incarcerated populations has been reported as 62 to 80 percent (71,72).

Family burden. Finally, family burden affects families that, in the absence of resources for needed treatment for their ill relatives, have no choice but to care for their relatives themselves or turn them away. Care increases the family's own stress and economic vulnerability, as usually no additional assistance is received from public authorities (73).

These processes reflect shifts by state mental health departments of the costs of operating state mental hospitals. Cost shifts to federal revenues have led to transinstitutionalization, cost shifts to local governments have generated homelessness and criminalization, and cost shifts to personal or private sources have led to family burden.

The largest effect, however, has been on quality of care. Rather than receiving appropriate attention in a hospital, patients live by themselves, stigmatized in the community, unable to obtain jobs, and lonely and isolated. Such conditions exacerbate symptoms, producing revolving-door phenomena.

Ironically, advocacy groups' attempts to protect patients' rights may have exacerbated negative outcomes by limiting treatment options even further. For instance, especially in the 1970s, a cohort of young public-interest attorneys focused on eliminating civil commitment, ending unnecessary detention, and upholding the right to refuse treatment. Some observers believe that outpatient commitment laws, such as Kendra's Law in New York State, represent rights protection and deinstitutionalization gone too far. To others, such laws reflect a reversion to criminalization as money is spent for police personnel to track, locate, and arrest persons with psychiatric disorders rather than for adequate residential and other community-based treatment and rehabilitation programs.

The impact of managed behavioral health care

As with deinstitutionalization, experiences with public-sector managed behavioral health care are mixed. Some empirical evidence of its successes, failures, and uncertainties exists, but the systematic, longitudinal study required to draw substantive conclusions is still lacking. We report available literature on the benefits and disadvantages of public-sector managed care as it has been implemented to date.

Benefits and potentials

Process effects. For more than 20 years, managed care proponents have contended that integrating primary care with other specialty services reduces deficiencies in access to care that occur under fee-for-service payment systems. Indeed, under managed care, initial access to diagnostic and treatment services and frequency of referral from primary care to specialty mental health services have improved for enrollees (12,74). The gate-keeping function, computer technology, and flexible benefits structure of managed care can enhance accountability and coordination of services (75).

In traditional fee-for-service models, clinicians usually avoid spending their time on anything but face-to-face contacts with patients—the only reimbursable services (76). However, in a capitated payment model, health systems may use a variety of techniques to maximize effective use of clinicians' time, such as telephone monitoring and referrals to community resources. This flexibility for providers may be supported by an expanded benefits package, giving patients access to more appropriate levels, types, and sites of care. For patients with complex illnesses and multiple comorbidities, flexibility in coverage and access can mean a vast improvement in quality and continuity of care (74,75).

In Hall and Beinecke's survey (77) of consumers and family members, both self-selected respondents and respondents affiliated with the National Alliance for the Mentally Ill, those enrolled in managed care reported better coverage. Specifically, residential care, crisis services, home care, office visits, family psychoeducation, and intensive case management were all more likely to be covered.

Outcomes. Empirical evidence abounds that initial and short-term cost savings accrue to private-sector groups insured under health maintenance and preferred provider arrangements compared with traditional indemnity plans. Recent reports from public-sector agencies implementing managed behavioral health care plans also support this trend. "Administered appropriately, [managed behavioral health care] can provide quality care at reasonable cost," according to an Institute of Medicine publication (12).

The cost savings are derived from several sources. In Massachusetts nearly half the savings generated in a public-sector plan resulted from a managed care organization's ability to extract discounts on traditional fees from providers (78). Organizations with large enrollments can also negotiate favorable prices on the basis of promised volumes; case studies indicate that some savings result from simple price reductions (4). Most savings come from substituting less costly outpatient or noninstitutional care for expensive inpatient services (79). Reducing unnecessary institutionalization can also benefit patients by decreasing psychological costs and improving their motivation and potential for rehabilitation, assuming alternative treatments are available.

Theoretically, through managed care the delivery of mental health services can become more evidence based, with research results translated into practice standards and ultimately into increased effectiveness. Currently, little consensus on psychiatric practice guidelines exists (75,80,81). However, for conditions for which guidelines and protocols are available, managed care can provide an infrastructure, through information systems, to make the protocols administratively feasible. For example, guidelines for depression could be the basis for automated prompts in medical charts for medication-related visits.

Treatment effectiveness could also be increased through managed care's potential to support innovative service delivery. The private sector is often viewed as less bureaucratic, less vulnerable to political pressures, more able to implement innovative programs, more able to change nonproductive services and practices, and more likely to dislodge entrenched institutional and professional barriers to improved care. The private sector has shown in some cases that managed care's flexibility can allow creative use of resources to manage difficult patients throughout the course of their illness (74,75).

Proponents of managed care also highlight consumer choice as a strength. Especially in network-based managed care organizations, information about participating providers is available to consumers, who can then choose among providers under contract to offer services at the same price to the consumer (19). Overall, managed care can provide an organized framework and appropriate standards to decrease inconsistent mental health treatment and the chaotic and fragmented nature of mental health care. The elements of choice and market competition can promote more accountability and increase providers' orientations toward outcome.

Negative effects

Unfortunately, the empirical and anecdotal evidence of the negative effects of public-sector managed care is as compelling as the evidence for its benefits.

Decreased access to care. Managed care may prevent access to care or reduce care either indiscriminately, for all groups, or by discriminating against certain groups, such as those with severe mental illness. In some states, substitution of managed care for traditionally delivered public services has caused confusion or produced a double standard of care (74,82). Managed care programs operating in state Medicaid programs may or may not include the protections originally intended to ensure availability of services to clients. State policy changes made to increase or decrease reimbursement for some services have led to treatment decisions not in patients' best interests. Medical necessity, a concept drawn from the traditional insurance literature, can be used to delay or preclude access to care (82).

Research indicates that Medicaid clients are subject to differential levels of treatment in the medical care system. Research suggests similar delays in treatment for patients with severe mental illness. When managed care is applied to public-sector clients with severe mental illness, traditional medical models can result in delays in treatment or exclusions of vital services. Vulnerable and disabled populations are potentially most at risk from the failures in the managed behavioral health care market (5).

In Hall and Beinecke's survey (77) of consumers and family members, respondents enrolled in managed care plans reported better coverage, as noted, but they also reported more difficulties with access, more grievances and complaints filed, and more unresolved appeals. Mechanic and McAlpine (5) reported reductions in mental health care across the board for individuals in the public system, regardless of diagnosis. They noted that irrespective of level of illness, the intensity of services was relatively unrelated to patients' prior time in treatment or the severity of their symptoms. Other studies found that treatment duration was often shorter in the public sector than in the private sector for those who had similar levels of illness, suggesting deliberate skimping or insufficient access (80,83).

Reductions in treatment may also be a function of contracting care to vendors through carve-out arrangements. Contracting methods and terms used for indemnity medical insurance are not necessarily appropriate for managed behavioral care of severe mental illness. Introducing gatekeepers into service delivery can change access to and continuity of care for individuals for whom constancy in treatment is paramount (19); those with chronic illnesses are not well served by disjointed or short-term arrangements.

Furthermore, contracts for delivering public-sector behavioral health care are often negotiated for only one year, which is often too short a period to ensure long-term provider commitments. Interfering with relationships with patients' usual providers can put patients at higher risk of falling out of the system altogether (84). If a vendor can delay care and reduce services under a capitated, short-term plan, the vendor can profit in spite of unmet client need. Ultimately, the public sector is responsible for any care that must be delivered, which calls into question whether contracting mechanisms are suitable for the individual client or for the public good.

The contracting method also affects stability of enrollment, through vendors' implicit or explicit selection of risk. Issues of risk selection have always been paramount in the managed care debate. Plans that offer comprehensive high-quality services, or services targeting less healthy or more needy patients, risk attracting more costly enrollees. By measuring potential enrollees' health status and predicting their future service use, plans can be selective about who enrolls. By limiting access to care once patients are enrolled, plans can encourage disenrollment of expensive patients.

In fact, patients in employer-sponsored managed behavioral health plans who have severe psychiatric disorders are significantly more likely to disenroll than those with less severe problems (85). In Los Angeles, a public mental health managed care plan had disproportionately higher rates of disenrollment of non-English-speaking clients and those with a diagnosis of schizophrenia (84).

Deficiencies in quality, appropriateness, and outcomes of care. The efficiencies and cost-effectiveness touted by managed care proponents may be negated by current models. Contracting with mental health practitioners separately from medical care providers, as in carve-out plans, adds administrative costs at several levels, and those costs are often passed through to the purchaser (12), which can be the plan, the employer, or the employee. In plans that require the vendor to assume risk and costs, the vendor may try to reduce administrative costs to improve profit margins. One way of doing so is to minimize the numbers of providers in the network or to contract with smaller provider groups, which often lack the specializations needed to treat complex cases like severe mental illness.

The potential for cost-shifting and profit-making is determined by the contractual terms between the payer and the managed care organization. Thus the incentives and disincentives built into these agreements are particularly important (75). The populations most vulnerable to serious and persistent mental disorders are often at risk due to failure to recognize their special circumstances or failure to treat them soon enough or at the appropriate level of care with specialty services. Economic incentives to provide less-intensive services and for shorter periods place those who need long-term services at risk of missed diagnoses, more complications, longer hospital stays, and poor recovery (12,83).

Recent case studies have illustrated measurably poorer outcomes from managed behavioral health care for persons with more serious disorders. In Massachusetts, managed care programs for persons with severe mental illness did decrease the number hospitalized, but they also increased the number of admissions and the length of stay for those admitted (78). A Utah study compared outcomes for clients with a diagnosis of schizophrenia who were enrolled in a managed mental health carve-out with those who received traditional fee-for-service care under Medicaid (86). All clients improved; however, those in the carve-out showed significantly less improvement, and the differences were greatest for those who had the worst mental health status at baseline.

Sometimes intensity of care may be reduced too much. In one study, patients with the largest reductions in inpatient length of stay had a significantly higher risk of readmission within 60 days (5). Other studies have also found deleterious effects for those with severe mental illness (12,87).

With regard to costs, savings gained from managed care do not appear to be directed toward improved care for persons with serious mental illness or for public clients. Shifting enrollees from a behavioral health plan to a medical plan, a medical contract, or a nonmedical system such as housing or welfare can reduce potential costs to the behavioral health plan and transfer costs to other payers, including the consumer and the community (5,12,74,75). In attempts to reduce pharmacy costs, some publicly financed managed care programs have constrained formularies, excluding the newer, more costly psychopharmaceuticals and promoting generic substitutes and lower-cost alternatives. With little evidence of clinical substitutability or effectiveness, this approach limits treatment alternatives (12).

The fallacy of choice. Consumer choice is considered a potential strength of the managed care model. However, it may mean relatively little in terms of plan selection, provider selection, or treatment options for individuals with mental illness, for several reasons. First, the chance to choose among health plans is available primarily to the employed. For public-sector beneficiaries, most state-sponsored behavioral health plans offer only one plan and one network of providers.

Further, the aspect of choice is only as good as the quality and transmission of information about the provider network and the quality and breadth of the network. Some behavioral health plans offer limited networks of providers, or networks heavily weighted with master's-level providers rather than doctoral-level professionals or psychiatrists. Public-sector networks are often restricted to providers who are willing to provide services at the Medicaid contracted amount, usually below market-based fees, further constraining enrollees' provider selections.

In addition, information about treatment may be geared to those who have greater capacities to understand and translate data—capacities that are often compromised among persons with mental illnesses (5). Culturally appropriate care may not be available in the provider networks offered to indigent mentally ill clients, or it may be financially inaccessible (88). Mental health consumers are also less likely to be able to accurately assess future treatment needs.

Mentally ill consumers as individual purchasers seem poorly positioned to determine what clinical conditions and treatments should be included in pooled-risk arrangements or to influence how plans are administered to make them cost-effective and socially beneficial (80). When we also consider the fact that individuals with severe mental disorders are not often direct purchasers, it is clear that espousing choice as a positive attribute of public-sector managed behavioral health care can be a farce.

The wrong theory and the wrong population. An analysis of public-sector managed care for persons with severe mental illness leads to the question of whether it is the wrong model or whether it is applied to the wrong population or both. Are adequate services being delivered to the most vulnerable of populations in an efficient, cost-effective way?

Managed care operates within an entrenched American system in which health insurance is tied to employment, except in certain instances in which medical care is provided as part of welfare or disability systems. This tie to private enterprise or to political, state-specific policies places certain populations at higher risk, regardless of any improved efficiencies under managed care. Under health insurance tied to employment, the majority of enrollees are relatively healthy, not those most vulnerable to severe mental illness. Managed care techniques of gatekeeping and utilization management are most suited to reducing unnecessary services for those with mild or moderate mental illness (83), while individuals with severe mental illness are typically underserved (5).

The employer's role in determining availability of medical care also muddies the relationship between managed care structures for managing and delivering services and the benefits packages available to subscribers. The most common benefits structure within managed care is counterintuitive to standard insurance theory—that catastrophic or high-end services should be covered by insurance, promoting consumer cost-sharing for the frequently used, low-end services and financially protecting consumers from the more costly but rare services. In most behavioral health plans, managed care covers the low-frequency, low-intensity patient; benefit limitations reduce coverage for high-frequency and high-need users.

Managed care may be applying these employer-based models to the wrong population. Under the new models, the most vulnerable individuals, and those for whom insurance should offer protection, remain at risk. Medicaid programs have a preponderance of persons who need high-intensity, specialty mental health care. The prevalence of psychiatric disorders in Medicaid programs is often four to five times higher than in the private sector; in employed populations, mental health services are used by less than 10 percent of the population (14,89).

In many ways, managed care for behavioral health services is a blunt instrument. Its incentives do not easily translate to mentally ill individuals whose care is publicly financed and who are often both poor and without supports. Managed care theoretically could increase flexibility in delivering needed benefit packages, but short-term contracts and emphasis on reduced costs provide little incentive for doing so. A system in which choice is limited, care is managed to decrease costs, and continuity is threatened is particularly troublesome for individuals with socially stigmatized, poorly understood illnesses that have traditionally been treated outside of standard medical practice settings. Whether such an approach will be an improvement over deinstitutionalization remains to be seen.

The mistakes of deinstitutionalization

Deinstitutionalization presents a complicated story of causes and effects, but we can now recognize some of the broad mistakes of the movement. They involved inadequate planning and monitoring and insufficient focus on client-centered services (19,23,90).

Lack of adequate planning

Mechanic (26) concluded that deinstitutionalization was premised on two central assumptions: that mental illness exists on a continuum, and that outpatient care is always more effective than hospitalization. Although these assumptions were certainly testable, they were instead accepted on faith. Indeed, in the early 1960s, when CMHCs were being proposed as state hospital alternatives, serious consideration was given to conducting pilot demonstration studies to evaluate whether deinstitutionalization would work (3). However, the idea was rejected because the research would take too long, results would probably not be definitive, and the favorable climate supporting mental health system changes would slip away.

Instead of being driven by facts and figures, deinstitutionalization was driven by congruent ideologies: that mental illness is a myth, or at least a social construction, and that mental illness is produced largely by environmental forces and is therefore preventable. Without any counterevidence, the rhetoric of community care developed a momentum of its own, producing excesses of ideology rather than thoughtful planning. The ideology was compatible with budget concerns in that community care was seen as a way to reduce costs, or at least to shift the cost burden from states to other funders.

Thus ideologies turned into polemics. The need was to empty state hospitals, to save money, and to affirm particular ideologies. Specific mechanisms and structures linking community-based alternatives to state hospitals were basically nonexistent.

Deinstitutionalization needed an integrated and meaningful federal policy on treatment of mental illness. Kiesler (90) has repeatedly censured the lack of any consistent national policy in mental health. Despite the significance of deinstitutionalization, it was really not a policy—it just happened. It represented basically an all-or-nothing approach to change: closing hospitals and changing laws versus making gradual changes in practice and relying on existing services as defaults for hospital treatment. Mental health law reforms for deinstitutionalization were made in a single-minded manner, following the "rule of the instrument" (91)—that is, when you give a person a hammer, everything he or she sees will need pounding (92).

Inadequate monitoring and accountability

From its initiation, deinstitutionalization lacked a clear focus of responsibility and authority. Thus it should not have been surprising that policy implementation did not match policy intent. Alternatives to institutions had few advocates or constituencies. The political process had insufficient checks or balances to monitor what the legislation and funding actually supported.

The first effect was legislation that created a community mental health system that was not focused on the original problem. Although CMHCs were required to include "aftercare" and alternatives to institutionalization, these services were not monitored for their effects on state hospitals, and thus they had little or no impact. Local programs required oversight to ensure congruence with legislative intent. However, the federal government provided neither sufficient funding nor authority for timely audit or corrections and no monitoring role for the states. Local citizen constituencies focused on the prevention aspects of the community mental health initiatives. Critically absent was an integrated, meaningful national policy on mental illness.

Lack of focus on client-centered services

As noted, deinstitutionalization legislation focused on the wrong population to achieve the original congressional intent. Federal CMHC legislation and the services it supported mainly emphasized prevention of mental illness in "normal" or, at best, at-risk community populations.

Deinstitutionalization was also based on the wrong theory for the target population. That is, to the extent that planning occurred, it was based on simplistic ideas about system operations and assumption of linear effects: that reducing hospital beds would have continuing, linear, and direct effects on expenditures. This simplistic perspective was fed by the belief that the mentally ill population was homogeneous, either because of the patients' illness or because of the institutions that housed them (93).

In retrospect, analysis of the data then available would have clearly indicated that this homogeneity did not exist. Initial hospital census reductions from 1956 to 1965 reflected release or shortened stays for a group of higher-functioning patients, and the process went quite well (20). However, later reductions involved less functional patients who often had co-occurring medical or substance abuse conditions and lacked family assistance for living arrangements and community support.

To be viable, mental health planning needed to be client focused and to take into consideration the diversity of the severely mentally ill population. Patients had different disorders and needs, showed various levels of disability and capability, and were at different points in their illness trajectories.

Further, as problems of mental illness are often significantly compounded by poverty and racism, the patient's family and community context should also have been incorporated into planning. Context includes family support, availability and quality of mental health services in the geographical area, community stigmatization of mental illness, access to jobs and educational opportunities for rehabilitation, and availability of public and private funding for medications and treatment (26). Instead, service plans were driven by dogma and self-interest.

Recommendations

The lessons learned from deinstitutionalization, along with empirical data and experiences from private-sector managed care, can and should be applied to public-sector managed behavioral health care. We suggest applications for planning, monitoring, and client-centered services.

Planning

Enough evidence is available to suggest that if the move to deinstitutionalization had been planned carefully and thoughtfully, history may have told a different story. But the movement had few phased-in programs, and even fewer demonstration projects. Goals were poorly specified, and plans for evaluating process or outcomes were rare. In many cases, the move to public-sector behavioral health care has also been inadequately planned in the midst of rapid, crisis-oriented implementation due to budget shortfalls and political pressures.

Early adopters of Medicaid managed care—states that adopted the approach before 1996—were particularly prone to mistakes, having had little managed care experience (55). The National Academy for State Health Policy analyzed experiences of the four states with the longest histories of Medicaid managed care for physical health (94). One critical element for successful operations was experience in measuring performance and setting goals before implementing the system—both to build baseline data and to hold contractual plans accountable. A well-functioning plan needs substantial lead time to prepare provider organizations and to build the infrastructure for information processing and reporting, utilization review, formulary control, and access to comprehensive care (5,55).

After general planning must come specific technical preparations for implementing a capitated plan, which requires experience in contracting, rate setting, insurance, and risk adjustment. Expertise in these elements is critical to operating, as well as evaluating and monitoring, a managed care program.

The accountability lacking in deinstitutionalization and in many managed care plans can be pegged to the absence of well-constructed risk adjustment measures, which would allow states and program administrators to assess the seriousness of disorders among their clientele and the appropriateness of the services delivered. Unfortunately, research on risk adjustment for mental illness is in its infancy, but efforts are under way to improve the predictability of risk adjustment measures for contracting purposes (19,54,95). Blended models of capitation and fee for service are gaining credence as methods for promoting incentives to treat serious mental illness yet deliver comprehensive services efficiently (80,96,97).

Contracts for care must have a long enough duration to promote coordinated, innovative care and to allow treatment plans to develop and come to fruition. Measures of contract performance should include both process and outcome indicators. Utilization review measures have substantial validity, having been applied to both inpatient and outpatient medical care and more recently expanded to incorporate behavioral health care. Measurement of short- and long-term outcomes can be based on clinical and social sciences research; studies are being funded on treatment efficacy, effectiveness, and costs for general and special populations, including those vulnerable to reduced access and care (81). Managers can then implement measurement systems in which data are collected at baseline and beyond and analyzed often enough to result in changed protocols.

Public responsibility and monitoring

Beyond the planning required to ensure successful programming and operations, responsibility for the program and its outcomes must be clearly assigned. Given the goals inherent in public-sector programming for individuals with severe mental illness, authority and responsibility must reside with the public purchaser. Contracting for services does not alleviate responsibilities for quality, access by intended beneficiaries, or costs; the public safety net must be maintained (5).

There is some evidence that states may be contracting with managed behavioral health care organizations on the basis of political considerations (10). One study of six states found difficulties in developing and implementing performance-based contracts for publicly funded mental health services (98). A study of public-sector managed physical health care in Florida identified lack of leadership and systems to handle state responsibilities as major impediments to implementation (55).

Government authorities should provide strong regulatory guidance and oversight, both of which were absent in the era of deinstitutionalization. All relevant levels of government should be involved in monitoring and assessing quality of care (12). This approach of course means that data necessary for monitoring must be available (5). The groundwork for evaluation must be laid by public-sector agencies—at several levels, and well in advance of the implementation of managed mental health care (99). Experiences of the Federal Employees Health Benefits Program indicate that employers, plan sponsors, and policy makers must all be involved in overseeing the adequacy of coverage, benefit levels, and performance requirements (80).

Relying on individual-choice mechanisms to influence quality and costs of managed behavioral health care is not likely to work well for consumers with severe mental illness. Thus policy makers and plan sponsors will need to assume a much stronger role on consumers' behalf in benefits structure, plan performance requirements, flexibility in plan management, and funding levels of competing plans. Policies focused on mandating the structure of benefits are not enough, as they leave open to managed care providers many other ways to affect coverage (100).

The responsibility of the public sector goes beyond technical, programmatic oversight and extends to the government's role as protector of the vulnerable. With regard to public-sector managed care for serious mental illness, the government plays an especially crucial role in providing beneficiaries with adequate health care, a safe environment, and the freedoms other citizens expect.

Multifaceted, client-centered programs

Given the complex and comprehensive needs of individuals with serious and persistent mental illness, the lessons of deinstitutionalization lead to multidimensional recommendations for behavioral health care in the public sector. The focus must be client driven at individual and macro levels. It must incorporate consumer and family education, measures of satisfaction, shared governance of care systems, and comprehensive insurance structures that support quality of care and quality of life (12).

Funding should be available to meaningfully involve consumers, family members, and their advocates in concrete planning, evaluation, and governance activities at local and state levels. The National Association of State Mental Health Program Directors' study of Medicaid managed care in states found that successful outcomes were attributable to extensive efforts to solicit public input, identify priorities, and develop public and legislative support (101). Health care purchasers must be made responsive to preferences of consumers and families and ensure their meaningful participation (12). Resources on advocacy strategies for developing managed behavioral health plans are available (102), as are reports on best and worst practices in private-sector managed mental health care (103).

Public-sector managed care organizations have significant advocacy responsibilities for the public-sector population with mental illness. They should be obliged to create additional acceptable and effective care alternatives, as a limited number of models now exist (88,101). History indicates that it is easy to reduce utilization and costs of mental health services by administrative fiat. Because individuals with severe mental illness are less able to speak for themselves, indicators of high-quality care and satisfaction are needed as counterchecks (5). More mental health services research to monitor and improve operations is also needed.

Other roadblocks to this recommendation for client-centered programs come from federal regulations, which can be barriers to consumers' community integration, rehabilitation, and recovery. For instance, Hogan (13) has noted that for assertive community treatment, which is an innovative and demonstrably effective model of intensive case management, Medicaid will pay for some services but not others, such as housing and job supports. It will pay for some clients but not all, because of their periodic or permanent changes in eligibility. Mechanic (26) has proposed a revision of SSI and SSDI disability determination policies so that subsistence benefits are provided without discouraging rehabilitation.

Integrated models of primary care and behavioral health services can also promote comprehensive, holistic care, using wraparound services and including housing and vocational rehabilitation services to promote full functioning (12). Carve-out arrangements in public programs should contain explicit contractual language about the availability and intensity of other needed services often considered not medically necessary, such as psychosocial rehabilitation and housing (5)—particularly because for-profit organizations will not otherwise have incentives to consider overall, long-term outcomes when operating under short-term contracts.

The Surgeon General's 1999 report on mental health concludes that we know what works (2). We also have a clearer, although imperfect, idea of what incentives help ensure accessible, high-quality, and comprehensive services. However, the report does not detail how managed care supported by public programs can fit into the complex world of serious mental illness. Although care can be mandated, quality will not automatically be assured (5). The public sector has a responsibility to guarantee the rights of individuals with serious mental illness. These rights must include high-quality, effective, individualized, and client-responsive services.

Conclusions

Public mental health services before, during, and after deinstitutionalization have had multiple, serious problems. Managed behavioral health care could be a welcome mechanism for systems improvement.

However, while managed care may hold out promise, there are significant reasons for caution. Thus greater attention to policy and procedures for managed behavioral health care for the public sector is critical to ensure that this method of service delivery does not become the third "shame of the states."

And what will happen now? Rochefort (104) and Armour (105) have suggested that policy making for severe mental illness runs in cycles, coming to the fore when the plight of affected individuals periodically penetrates societal consciousness. The hallmark of each "mental health reform" has been a new environmental approach to treatment and an "innovative" locus of care—historically, care has moved from almshouses and jails to moral treatment in asylums to the mental hygiene movement and the psychopathic hospital to community mental health systems (20).

Each cycle of reform began with the premise that early treatment through the chosen modality would prevent the personal and societal problems associated with mental illness. Each approach was championed as a generic solution for mental illness, later to be proven useful only for acute or milder forms of illness, not for long-term, serious psychiatric disabilities. Each reform cycle flourished for a few years, but ultimately each produced limited effects, no effects, or harmful effects for patients with "chronic" psychiatric conditions.

Durham and LaFond (106) have reminded us that "mental health policy is shaped…by overarching values and social forces at work throughout society." Thus the liberal era of the 1960s and 1970s, which emphasized individual freedom and fairness to individuals, also saw great reductions in the institutionalization of persons with mental illness and the noble intent to better their lives in the community.

The 1980s and 1990s brought a shift to neoconservatism as a new world economy threatened American society. Individual responsibility was emphasized, as was the need to restore stability, order, and traditional American values and institutions. The results for individuals with serious mental illness are confusing and frightening: increased criminalization of patients to control aberrant behaviors; a safety net with giant holes, ever widening on the ill-founded belief that these individuals can take personal responsibility for their condition; and, finally, expanded implementation of managed behavioral health care in the public sector. This expansion is based on the assumption that all adults can make rational choices about their health care needs and health providers, and that severe mental illness does not de-level the playing field; nor does poverty, racism, or disability. However, in the 1990s we also witnessed the increasing activism of consumers and family members. As Mechanic (19) has noted, "The case for equal treatment of psychiatric disorders with other medical problems has gained greater credibility and public support."

Now that we have observed a pattern, maybe we can avoid or minimize the mistakes of the past. Perhaps this comparative analysis of deinstitutionalization and public-sector managed behavioral health care will enable advocates, consumers, and concerned policy makers to amass and better use knowledge so that public-sector managed behavioral care does not become the third shame of the states. One hopes that implementation of managed behavioral health care for those with serious, long-term mental illness will develop purposefully and wisely, to the betterment of consumers and society.

Dr. Mowbray is professor at the University of Michigan School of Social Work, 1080 South University Avenue, Ann Arbor, Michigan 48109-1106 (e-mail, [email protected]). Dr. Grazier is associate professor in the University of Michigan School of Public Health, Health Management, and Policy, and Dr. Holter is assistant professor at the School of Social Work.