Implementing Supported Employmentas an Evidence-Based Practice

As a result of more than two decades of research, we know a great deal about improving outcomes and enhancing the recovery process for persons with severe mental illness by providing effective mental health services. Unfortunately, the implementation of interventions that have been shown to be effective by research, termed here evidence-based practices, lags considerably behind the state of knowledge. Individuals with severe mental disorders such as schizophrenia are unlikely to receive treatment with basic evidence-based practices in routine mental health settings (1). Implementation of evidence-based practices must overcome many obstacles, some generic and some specific to a particular evidence-based practice. Nevertheless, the field of mental health services is slowly committing itself to providing research-based services as the foundation of care (2).

In this paper, the first of several on specific evidence-based practices for persons with severe mental illness, we discuss supported employment, a recent approach to vocational rehabilitation that has proved to be consistently more effective than traditional approaches. Our goals are to familiarize clients, families, clinicians, administrators, and mental health policy makers with supported employment; to review the findings and limitations of current research; and to discuss implementation issues, including availability, barriers, and strategies. Because several recent reviews of research on supported employment already exist (3,4,5,6,7), our intent is to provide information that is accessible to stakeholder groups other than researchers.

Supported employment

Supported employment is a well-defined approach to helping people with disabilities participate as much as possible in the competitive labor market, working in jobs they prefer with the level of professional help they need. According to the federal definition, supported employment means "competitive work in integrated work settings…consistent with the strengths, resources, priorities, concerns, abilities, capabilities, interests, and informed choice of the individuals, for individuals with the most significant disabilities for whom competitive employment has not traditionally occurred; or for whom competitive employment has been interrupted or intermittent as a result of a significant disability" (8).

Although the federal definition of supported employment includes reference to transitional employment, that is, temporary community job placements, the two are very different, both conceptually and in practice (9). Many agencies offer both, and when they do, practitioners understand them to be different approaches; transitional employment is seen as a step toward supported employment (10). We do not discuss transitional employment in this paper.

Although many supported employment principles have been espoused for decades (11), these ideas crystallized in the 1980s through the efforts of a national network of educators, who concluded that sheltered workshops isolate people with developmental disabilities from mainstream society (12). This network was successful in changing federal regulations on the types of services funded by the federal-state vocational rehabilitation system.

By 1987 supported employment had attracted attention in the psychiatric rehabilitation field (13). As adapted for this population, supported employment programs typically provide individual placements in competitive employment—that is, community jobs paying at least minimum wage that any person can apply for—in accord with client choices and capabilities, without requiring extended prevocational training. Unlike other vocational approaches (4,14), supported employment programs do not screen people for work readiness, but help all who say they want to work; they do not provide intermediate work experiences, such as prevocational work units, transitional employment, or sheltered workshops; they actively facilitate job acquisition, often sending staff to accompany clients on interviews; and they provide ongoing support once the client is employed.

Supported employment programs are found in a wide variety of service contexts, including community mental health centers, community rehabilitation programs, clubhouses, and psychiatric rehabilitation centers (10,15,16). Although the evidence suggests that supported employment is optimally effective only when clients concurrently receive adequate case management, it is not necessarily limited to a specific service model such as assertive community treatment.

The most comprehensively described supported employment approach for people with severe mental illness is the individual placement and support model (17,18). We do not view this approach as a distinct supported employment model. Instead, it is intended as a standardization of supported employment principles in programs for people with severe mental illness, so that supported employment can be clearly described, scientifically studied, and implemented in communities. In fact, a survey of 116 supported employment programs throughout the United States found that these programs generally follow principles of the individual placement and support model (19).

Effectiveness of supported employment

To understand the context of the current review, several points from the broader vocational literature are critical. First, interventions that do not target job placement directly have very little impact on employment outcomes (20). Second, many vocational approaches to helping people with severe mental illness gain employment have been developed over the past half century. Few have been evaluated rigorously; those that have been examined in controlled trials have yielded disappointing results (4,14,21,22).

Quasi-experimental studies. To date, three quasi-experimental studies have evaluated day treatment programs that converted their rehabilitation services to supported employment. Drake and colleagues (23) studied a rural New Hampshire community mental health center that developed a supported employment program to replace the day treatment services. A natural experiment compared the conversion site with a nearby site, which continued its day treatment along with traditional brokered vocational services. The competitive employment rate increased substantially at the conversion site, whereas the rate was unchanged at the comparison site. Moreover, adverse outcomes such as hospitalization, incarceration, and dropouts did not increase at the conversion site.

Clients, their families, and mental health staff had favorable reactions to the conversion, although a minority mentioned loss of social contact as a drawback (24). Interestingly, many clients who did not find work also reported that they benefited from the change because they discovered satisfying activities outside the community mental health center.

Replacing day treatment with supported employment also led to cost savings (25). Given the success of the initial conversion, the second site subsequently converted to supported employment with similarly favorable results (26). In a second study involving the downsizing of a day treatment program in a small city, clients who transferred to a new supported employment program had better outcomes than those who remained in day treatment (27).

A third study compared two Rhode Island day treatment programs that converted to supported employment with one that did not (28), with similar findings. Others have also reported successful conversions of day treatment to supported employment programs (29). These evaluations demonstrate that supported employment can be implemented in a cost-effective manner in real-world settings with a broad range of clients with severe mental illness, not just a select group who sign up for supported employment.

Randomized controlled trials. A 1997 review (3) summarized the findings of six randomized controlled trials comparing supported employment with a variety of traditional vocational services for people with severe mental illness (30,31,32,33,34,35). All six studies reported significant gains in obtaining and keeping employment for persons enrolled in supported employment. For example, a mean of 58 percent of supported employment clients achieved competitive employment at some time over a 12- to 18-month period, compared with 21 percent of the control group, who received a range of alternative vocational interventions, including skills training, sheltered work, and vocational counseling as steps toward competitive job placement. Control subjects received what providers in their communities believed to be best practices in vocational rehabilitation.

Other competitive employment outcomes, such as time employed and employment earnings, also favored supported employment clients over those in control groups. A meta-analysis of these studies reached very similar conclusions, noting that the findings were robust (5,6).

Recently, data collection was completed for the Center for Mental Health Services Employment Intervention Demonstration Program (36). Eight sites in this project used randomized controlled trials to evaluate the effectiveness of supported employment. Reports of findings from this multicenter trial are expected over the next year.

Two sites have reported preliminary experimental findings. In Hartford, Connecticut, Mueser and associates (37) compared individual placement and support with two established vocational approaches. One was a psychiatric rehabilitation center using transitional employment, and the other was a brokered approach using a combination of sheltered workshops, government set-aside jobs, and individual placements. Meisler and colleagues (38) compared an individual placement and support program working within an assertive community treatment team with usual vocational services in a rural community in South Carolina. The control group was assigned to a well-respected rehabilitation center with long-term contracts providing numerous government set-aside jobs.

Findings from both studies replicated the previous findings of large differences in competitive employment outcomes favoring supported employment over traditional approaches. Even with protected jobs—transitional employment and set-aside jobs—factored in, supported employment clients in both studies still had better employment outcomes.

Many of these studies have also examined nonvocational outcomes, such as rehospitalization rates, symptoms, quality of life, and self-esteem. Studies rarely have found any experimental differences in nonvocational outcomes favoring clients enrolled in supported employment programs over those in comparison programs. In other words, the group effects for supported employment programs appear to be restricted mainly to competitive employment outcomes, at least for the relatively brief follow-up periods in the studies reviewed. However, neither has any research suggested any adverse effects from participation in supported employment programs. Rehospitalization rates are unaffected by participation in supported employment, contrary to the belief that the stress of work might lead to higher relapse rates.

Although enrollment in a supported employment program itself does not lead to improved nonvocational outcomes, clients who actually engage in competitive work do experience improvements in self-esteem and in control of symptoms, compared with clients who do not work or work minimally (39,40).

Cost considerations are a core issue in decisions to implement psychiatric services. Supported employment services are labor intensive. Annual cost per supported employment participant is around $2,000 to $4,000 (25,41). These figures are similar to those for traditional vocational services (42). Clients enrolled in supported employment programs sometimes use fewer mental health services, notably day treatment, suggesting a cost offset (25,43,44,45).

Critical components

Reviewers seeking to identify empirically validated principles of supported employment have reached similar conclusions (7,46,47,48,49). Certain components are almost always present in successful vocational programs. They are generally found in the supported employment programs evaluated in the eight randomized controlled trials summarized above. The following components are predictive of better employment outcomes:

• The agency providing supported employment services is committed to competitive employment as an attainable goal for its clients with severe mental illness, devoting its resources for rehabilitation services to this endeavor rather than to day treatment or sheltered work. Numerous studies indicate that this element is common in successful programs (23,26,27,28,29,33,34,49,50).

• Supported employment programs use a rapid job search approach to help clients obtain jobs directly, rather than providing lengthy preemployment assessment, training, and counseling. The evidence in this area is strong, with two randomized controlled trials focusing specifically on this variable (30,51), plus five randomized controlled trials in which this component was a critical difference between study conditions (32,33,34,37,38). A randomized controlled trial evaluating a vocational approach involving extended classroom training before job placement yielded employment outcomes similar to those of a control group referred to the state vocational rehabilitation office for vocational services (52).

• Staff and clients find individualized job placements according to client preferences, strengths, and work experiences. Several correlational studies support this conclusion (49,53,54,55,56).

• Follow-along supports are maintained indefinitely. Correlational findings from four different research groups indicate that this component is an important one (31,57,58,59).

• The supported employment program is closely integrated with the mental health treatment team. The experimental evidence is consistent with this conclusion even though this variable has not been studied in isolation (31,32,33,35,37,38,59,). This principle is also supported by a strong theoretical rationale (60). However, despite its strong evidence base, it is not universally practiced (19).

Together these principles serve as a foundation for evidence-based guidelines for providing effective supported employment services. In one statewide survey, programs rated high in implementing these principles had better employment outcomes (unpublished data, Becker DR, 2000). A number of specific program elements—for example, reasonable caseload size, diverse employment settings, assertive outreach, and benefits counseling—are found in most supported employment programs (15), but the association between these elements and better employment outcomes has not yet been established. Further research is needed to clarify the critical ingredients of supported employment, which will lead to modifications, refinements, and additions.

Limitations of the evidence

Client factors

The most consistent finding from the supported employment literature has been the absence of specific client factors predicting better employment outcomes. Diagnosis, symptoms, age, gender, disability status, prior hospitalization, and education have been examined, and none have proved to be strong or consistent predictors (30,32,33). Notably, a co-occurring condition of substance use has not been found to predict employment outcomes (61,62,63).

Although a work history predicts better employment outcomes in supported employment programs, supported employment remains more effective than traditional vocational services for clients with both good and poor work histories (28,32,33). We speculate that the professional assistance provided by supported employment programs at every stage of the employment process compensates for client deficits in a way that less assertive vocational rehabilitation approaches do not. Consequently, the extensive literature on client predictors of work outcomes among people with severe mental illness who either have had little vocational assistance or have been enrolled in traditional vocational programs (48) may be largely irrelevant for supported employment programs.

Randomized controlled trials of supported employment have been conducted in settings with significant numbers of Caucasian (30,31,32,59), African-American (33,38), and Latino (37) clients. Although more replications are needed, all the evidence to date suggests that the greater effectiveness of supported employment compared with traditional vocational services is generalizable to both the African-American and Latino populations. Within-study comparisons of employment rates for different ethnic groups have been hampered by small sample sizes, so we cannot yet determine whether supported employment is equally effective for all ethnic groups within a specific setting.

We may make our best progress in understanding the role of ethnicity in supported employment programs by combining results across studies using meta-analytic techniques and through qualitative studies (64,65,66). We know anecdotally that culture and language pose significant barriers to providing supported employment in some populations.

Not all clients benefit from supported employment. For example, in community mental health centers converting day treatment programs to supported employment programs, some clients do not have employment as a current goal; not surprisingly, these clients usually do not work. But even among clients who express an interest in working, a sizable proportion are not working at any given time. We need to develop effective strategies for these clients. Helping clients decide whether supported employment is right for them also is critical. Informational sessions explaining beforehand how supported employment works improve clients' ability to make informed decisions about participating, thereby potentially reducing dropout rates (67,68).

Community and economic factors

Supported employment has been implemented successfully in many different types of communities. Programs in rural areas are no less successful than those in urban areas (49,50). One counterintuitive finding is that economic conditions apparently do not have a potent influence on employment rates for a supported employment program (50,69,70,71). Catalano and colleagues (69) have speculated that an economic theory of labor markets applies here. The primary labor market, comprising professional and semiprofessional jobs, shrinks during economic recessions. The secondary labor market, which includes entry-level jobs in the service industry, is more elastic and less vulnerable to economic downturns. Supported employment programs find jobs mostly in this secondary labor market, where jobs are usually available. However, the aforementioned studies examined a relatively restricted range of unemployment. The findings may not be generalizable to communities where the unemployment rate is very high (35).

Job opportunities available to clients with severe mental illness are often restricted because of the clients' limited work experience, education, and training, and consequently most supported employment jobs are unskilled (3,72). Half of all clients leave their supported employment positions within six months (3), although nondisabled workers in these occupations also have high turnover rates (73). Moreover, most supported employment positions are part-time. Clients often limit work hours to avoid jeopardizing Social Security and Medicaid benefits (48,74). A continuing challenge for supported employment programs is helping clients capitalize on educational and training opportunities so that they may qualify for skilled jobs and develop satisfying careers (72).

Program factors

Specific details about the best ways to provide supported employment services have not been adequately researched. Issues include the role of disclosure of mental illness in finding and keeping jobs, the range, location, timing, and intensity of supports provided to clients (57,75), and the nature of coworker and supervisor supports (76). The relationship between supported employment services and medication issues has not been well studied despite its assumed importance (77).

Long-term outcomes of supported employment also have not been widely studied. Programs that remain engaged with their clients over time, respond to clients' expressed wishes, and sustain an approach that integrates clinical and rehabilitation services are those we believe have the best outcomes over time. However, with few exceptions (30,58,59), most randomized controlled trials do not have follow-up information beyond two years. Much longer follow-up periods are needed to determine whether sustained commitment can yield favorable outcomes for more clients.

Implementation barriers

Access to supported employment

Sixty to 70 percent of people with severe mental illness would like to work in competitive employment (78,79), yet 85 percent or more of those in public mental health systems are not doing so (78,79,80,81,82). Most prefer competitive employment to sheltered workshops (83) and day treatment (30,84). However, most clients lack access to employment services of any kind. Less than 25 percent of clients with severe mental illness receive any form of vocational assistance (1,85,86), and only a fraction of these clients have access to supported employment (87). In some states, supported employment programs are now commonly found in community mental health centers, but their capacity falls far short of the need (19,50). A further question concerns the quality of available programs. Not surprisingly, it is mixed (15,49).

Barriers to implementation of high-quality programs exist at many levels—within federal, state, and local governments and program or clinic administrations, among clinicians and supervisors, and in the collaboration with clients or families. The remainder of this paper is devoted to discussing the barriers to implementing high-quality supported employment programs and offering suggestions based on experience for overcoming them.

Government barriers

Historically, the federal-state vocational rehabilitation system has been the primary funding source for employment services. However, federal funding for vocational rehabilitation has never been sufficient to serve more than a tiny proportion of the population in need (88). Moreover, many observers have expressed doubts about whether this funding has been used wisely. Vocational rehabilitation expenditures apparently have been disproportionately devoted to administration and to assessment and other preemployment activities (89). Compounding the problem is the fact that persons with severe mental illness fail to complete the vocational rehabilitation eligibility process twice as often as people with physical disabilities (90). Nevertheless, vocational rehabilitation agencies continue to allocate minimal funding for supported employment services (91).

Public funding for mental health is a second source for financing supported employment services. Unfortunately, community mental health centers historically have allocated only a tiny proportion of their budgets to vocational services (85). Since the 1980s, most states have amended their Medicaid state plans to cover community mental health services under the optional rehabilitative services provision, which permits a broad interpretation of the range of reimbursable interventions.

Vocational training is among the few services statutorily excluded from Medicaid reimbursement. However, evidence-based components of supported employment, such as ongoing supportive counseling in home and community-based settings, team meetings, psychiatrist involvement in rehabilitation planning, and assisting clients in developing job opportunities, are all Medicaid-reimbursed rehabilitative services that states may cover. Yet most state Medicaid plans include unnecessary limitations on covered services when they involve vocational activities. Given the increasing proportion of total funding of community mental health services that Medicaid expenditures represent, misinterpretation of federal Medicaid policy results in a major barrier to supported employment service access.

Fee-for-service systems of reimbursement for units of service, regardless of outcomes, have created incentives to perpetuate services that are not evidence based, such as day treatment (92). Some commentators have concluded that financing of supported employment programs within managed care systems will not be any easier (93).

The fragmentation of supported employment funding has also resulted in separation of services. Historically, supported employment services have been brokered—that is, offered at an agency separate from the community mental health center (16)—even though we now know that this approach is counterproductive (47,60). Even supported employment programs that are located in community mental health centers often are not closely integrated with mental health treatment teams (19), despite strong evidence that such integration is vital for success. In Indiana, a separate role for follow-along specialists created by separate funding sources has contributed to discontinuity of services (94).

Directors of state mental health departments can have a critical leadership role in promoting supported employment services. In the 1980s, Ohio's decision to pursue case management and housing as top priorities led to critical improvements, but this decision sacrificed the development of employment services by relegating it to a secondary goal (82). Some states have adopted a "range of vocational options" (95), leading to a proliferation of diverse—and untested—models, whereas other states have invested major resources in specific models that are not evidence based. Still other states have taken the stance that supported employment is not the business of the state mental health agency. Moreover, most states do not systematically monitor client outcomes, precluding the development of objective methods for rewarding successful employment programs.

Program administrators

From an administrator's perspective, common barriers include finding money to finance start-up and ongoing program costs, managing organizational change, and coping with political ramifications of change in the community. Administrators often do not provide the leadership for the adoption of innovations, even when they are evidence based. Administrators who do not have information about evidence-based practices may not value their outcomes or believe that they are possible (49). Administrators, especially those who received training and professional experience in an earlier era, may hold negativist attitudes about the feasibility of work—for example, "Schizophrenia is a chronic disease with little hope of recovery…work is a source of unnecessary stress."

If administrators are unwilling to consider change, it is unlikely that practitioners will. Poor management practices constitute another obvious barrier to implementation of evidence-based practices (96). Agencies that are driven by crises and chaos often have leaders and supervisors who have not established a system of careful treatment planning that is related to clients' desires and needs.

Clinicians and supervisors

Like administrators, clinicians often view clients as too unmotivated to work (97) and often underestimate the need for vocational services (98,99). Many practitioners lack adequate information and skills to staff supported employment programs (100,101,102).

Resistance to change is a barrier in any organization. In the mental health field, professional identities are defined by what practitioners do—methods employed, program name, and the like—or by their discipline, not by the outcomes sought. Program changes sometimes are introduced as externally imposed ideas rather than resulting from a process that includes the participation of the clinicians and supervisors, who are ultimately responsible for implementing the desired change (103). In such circumstances, practitioners perceive change efforts as a criticism and devaluing of their work.

Another common barrier concerns inadequate resources. Staff members cannot implement supported employment programs effectively if they do not have enough time to carry out their duties or if supervisors give them conflicting messages about the scope of their responsibilities. For example, when employment specialists are assigned additional job duties that are not vocational, they are distracted from the employment effort.

Clients and families

Clients and family members often do not have accurate information about supported employment. Sometimes clients are discouraged from considering employment by well-meaning clinicians and family members who believe that the stress associated with work outweighs the benefits. Instead, they are directed to day programs. Clients often believe that returning to work automatically compromises their eligibility for Social Security and Medicaid benefits. Families may not be given information on how to support a family member's work efforts, or they may not be considered part of the team or support network.

Strategies for implementation

Although we have more systematic information about barriers to evidence-based practices than we do about strategies to overcome them, some approaches for implementing evidence-based practices have been identified (104,105).

Government efforts

At the state level, a first step is to set clear outcome priorities. Next, systematic assessment of employment outcomes is absolutely essential. State mental health authorities must remove organizational and financial barriers to the development of supported employment programs, as has been done in New Hampshire (50), Vermont (unpublished data, Dalmasse D, 1998), Rhode Island (106), and Kansas (49). In both New Hampshire and Rhode Island, state mental health and Medicaid agencies joined to request that the Health Care Financing Administration allow reimbursement for supported employment services aside from direct interventions to teach job skills. Their requests were approved, thereby enabling Medicaid financing to greatly increase clients' access to supported employment services.

Recent federal legislation—the Medicaid buy-in program authorized by the Balanced Budget Act of 1997 and the Ticket to Work and Work Incentives Improvement Act of 1999—has permitted state governments more flexibility in establishing Medicaid eligibility, with the intent of reducing barriers to employment posed by the potential loss of Medicaid benefits (107). Some states—Oregon and Minnesota, for example—have implemented new policies expanding Medicaid coverage to allow more liberal income and resource thresholds for people with disabilities who work.

State mental health authorities have had success in providing direct incentives to local systems for meeting employment goals. In Ohio, participating systems doubled their employment rate when incentives were instituted (82). In New Hampshire, the competitive employment rate for community mental health center clients with severe mental illness has increased from 7 percent to 37 percent since 1990, when the state began emphasizing competitive employment in contracting (50). State vocational rehabilitation agencies in Alabama, Oklahoma, and Pennsylvania have initiated "results-based funding" for supported employment, which similarly rewards agencies for performance (108,109). Some caution is necessary, because unless designed carefully, such incentive systems may encourage enrolling clients with the fewest needs.

Incentives are not enough, however. The state agencies should also take the leadership in providing technical assistance by forming partnerships with leading research and training centers with appropriate expertise, as have those of New Hampshire (110), Rhode Island (106), and other states. Kansas, Indiana, New Jersey, and New York City have established supported employment technical assistance centers to help local programs implement and monitor supported employment services.

Building consensus among stakeholders is another element in the adoption of evidence-based practices. The National Association of State Mental Health Program Directors has issued a position statement on employment and rehabilitation for persons with severe psychiatric disabilities that identifies state mental health agencies as having a responsibility to influence vocational rehabilitation and other state employment agencies to collaborate to improve access by persons with severe mental illness to competitive employment (111). Accepting this mandate, Rhode Island's state mental health agency has involved the state's Medicaid and vocational rehabilitation agencies in funding supported employment. Funding for consensus-building activities related to exemplary practices is available through the Community Action Grant Program of the Center for Mental Health Services.

Efforts of program administrators

High-achieving organizations concentrate energy and resources on specific outcomes and reduce distractions to those outcomes (112). Important elements of leadership include articulating desired outcomes and practices for achieving them, building an organizational structure and culture that will facilitate implementing evidence-based practices, designing systems to monitor evidence-based practices and client outcomes, hiring staff with appropriate attitudes and skills, establishing group supervision or other methods of collaboration, creating employee evaluation procedures that emphasize evidence-based practices and employment outcomes, and providing rewards for high performance in those areas (49,113).

Supported employment programs are most successful in agencies that make a total commitment to competitive employment without diluting their focus and resources with traditional forms of vocational programming (49,50). A similar pattern is found in the developmental disability field, where supported employment has failed to develop its full potential because many agencies have viewed supported employment as an "add-on" service while maintaining large sheltered workshops (114).

As noted above, community mental health centers have been successful in converting day treatment programs completely to supported employment. Because this redeployment of resources has the advantage of cost savings in addition to acceptance by important stakeholders, it is a very appealing strategy. Consumer-run services can play a role in meeting the social needs of unemployed clients after conversion from day treatment to supported employment (115).

Monitoring the fidelity of program implementation is critical for implementing evidence-based practices (116). Accordingly, researchers, state planners, program directors, clients, and family members are increasingly emphasizing fidelity. The Individual Placement and Support Fidelity Scale (117), a 15-item instrument that assesses the implementation of critical ingredients of supported employment, is one such tool in the public domain. Although it was designed for use by assessors who are familiar with the critical ingredients of the model, its simplicity permits its use by nonresearchers.

Adequate reliability has been found in a field test using site visits by pairs of assessors who interviewed staff, studied charts, and observed program activities (117). The Individual Placement and Support Fidelity Scale clearly differentiates supported employment programs from other vocational approaches, suggesting that it can be used to determine whether a program actually is implementing supported employment (19). More comprehensive scales measuring supported employment implementation also have been field-tested (15,94).

Efforts of clinicians and supervisors

Agencies that successfully adopt supported employment appear to share a set of common elements (49,113,118). Successful programs give staff the resources they need to do their job well. This also means that the agency itself must be well managed in other areas and must provide high-quality case management services. Supervisors need to provide clear vision, organize services into a multidisciplinary team structure, and focus on outcomes rather than service units and paperwork (119).

Community mental health centers successfully adopting an innovation usually have at least one key change agent who champions the innovation (120). The change agent must have sufficient authority to implement change. When introducing supported employment, the change agent identifies respected frontline practitioners who can help lead the implementation effort. They in turn recruit other staff to join in the planning and development of the new program so that all staff will feel ownership of the program.

Adequate training and ongoing supervision are critical to give staff the skills to implement the practice (118). Guidelines, training manuals, and videotapes are important tools for ongoing monitoring and transmission of the culture of the supported employment program (118). Another critical element is expert consultation through site visits and telephone conference calls. Implementation is facilitated by having staff—not just employment specialists but also administrators, clinicians, and supervisors—visit exemplary supported employment programs.

Efforts of clients and families

Clients and families are well aware of the need for vocational services (89,98,121) but need to know what good services look like and how to advocate effectively in legislation and funding decisions. They can have influence over setting standards and ensuring adherence to those standards at the state, program, and client levels. Clients and family members should seek membership on advisory boards at all levels. They can collaborate with state officials to fund supported employment programs and to establish standards based on evidence-based practices and have them incorporated in licensing standards, requests for proposals for grant funds, and so on. At the program level, they can demand that entrance criteria for supported employment be based on a client's desire to work and not on symptoms or work history. They can also participate in designing supported employment programs. On an individual client level, they can argue for client choice and services that match evidence-based practices.

Conclusions

The emerging evidence base on supported employment is clear and consistent, with improved employment outcomes across many different types of settings and populations. In addition, most supported employment approaches described in the literature converge on a set of critical components.

One key remaining task is to overcome implementation barriers to make supported employment services available on a widespread basis. No other vocational rehabilitation approach for people with severe mental illness has attained the status of evidence-based practice despite a half century of program innovation and informal experimentation by many psychiatric rehabilitation programs. Proponents of other vocational approaches either have failed to empirically investigate their methods or have failed to find strong evidence. It is also true that many vocational program approaches that are not effective continue to be widely practiced.

Beyond implementing supported employment, we must continue to refine and improve our model to reach a wider spectrum of the population and to help clients not only find and keep paid community jobs but also to develop long-term careers.

Acknowledgments

Work on this paper was supported by National Institute of Mental Health grants MH-00842 and MH-00839, Center for Mental Health Services contract 280-00-8049, and a grant from the Robert Wood Johnson Foundation.

Dr. Bondis affiliated with the department of psychology at Indiana University-Purdue University Indianapolis, 402 North Blackford Street, Indianapolis, Indiana 46202 (e-mail, [email protected]). Ms. Becker and Dr. Drake are with the New Hampshire-Dartmouth Psychiatric Research Center in Lebanon, New Hampshire. Dr. Rapp is affiliated with the School of Social Welfare at the University of Kansas in Lawrence. Mr. Meisleris with the department of psychiatry at the Medical University of South Carolina in Charleston. Dr. Lehmanis with the department of psychiatry at the University of Maryland in Baltimore. Dr. Bell is with the department of psychiatry at Yale University in New Haven, Connecticut. Dr. Blyleris affiliated with the Center for Mental Health Services of the Substance Abuse and Mental Health Services Administration in Rockville, Maryland.