Training the leaders... of tomorrow.


Persons with a severe mental illness (SMI) are individuals with serious and long-term mental disorders that impair their capacity for self-care, interpersonal relationships, work and schooling. It is estimated that approximately 200,000 of the SMI population are homeless. These individuals tend to be relatively young, unmarried, poor and with low prospects of long-term gainful employment; some also have substance abuse problems, are HIV positive, and find the transition from living on the street to living indoors exceedingly difficult.

The housing needs of the SMI homeless can be divided into two parts: the need for decent, affordable housing; and the need for supportive assistance. Based on our current state of knowledge, there is nothing to suggest that either severe mental illness or homelessness requires unique physical structures or dwellings, or unique approaches to make housing affordable. Thus, new or unique housing policies are not required to address the first part of this population's housing need.

By contrast, the need for supportive assistance is so compelling that it is likely to require new approaches to housing policy. This conclusion is based on three factors: (a) the manifestations of severe mental illness--symptoms, behaviors, and functional impairments--do not disappear when a severely mentally ill homeless individual moves into a housing setting; (b) the variations in the manifestations of mental illness and the unpredictability of episodes of decompensation or movement into the active phase of illness; and (c) the consensus among those who have developed and managed housing for SMI individuals that the availability and continuity of appropriate services is key to successful housing experiences. Thus, a fundamental housing policy issue is ensuring that the service needs of residents are met.

A review of research on housing strategies for assisting SMI individuals reveals many knowledge gaps. At this writing, research has neither demonstrated what attributes or factors are critical to an individual's capacity to live independently, nor described the types of residential alternatives that are most effective for individuals within the SMI population, nor identified the specific attributes of individuals that are systematically associated with, or predict, the type(s) of residential setting that will be best, nor come to an agreement on the most appropriate way to conceptualize and measure effectiveness.

There are two primary implications of these information gaps for housing policy development. First, early policies should be viewed as experiments or demonstrations, designed at the outset to achieve two goals: delivering effective housing assistance to the SMI homeless, and providing essential information, through careful evaluation studies, upon which these policies can be improved or new policies developed. Second, all stakeholders should have realistic expectations regarding what these early policies can accomplish, and the likely need for mid-course corrections.

Some consistent observations can be gleaned from the existing body of research, however:

  1. Among studies of psychiatric inpatients, one consistent finding is that the greatest risk of homelessness occurs immediately following hospital discharge. Thus, if homelessness is to be prevented among this group of the SMI population, hospital discharge planning must include explicit housing arrangements and service supports that will be available to the individual as he or she re-enters the community.
  2. Heterogeneity and diversity of the SMI population requires a range of housing and service options. The fact that these individuals share in common the presence of a severe and persistent mental illness should not be misinterpreted to mean that they also share identical tastes, preferences or needs. Policy, therefore, should foster interventions that can be tailored to individual needs.
  3. Unpredictability of illness episodes requires continuity of care. Mental health experts currently believe that the individualized case management approach is an effective way to achieve such continuity.
  4. Housing and services must be linked as neither alone is sufficient.
  5. Conceptualizing and measuring "outcomes" or "effectiveness" of supportive housing is complex; the choice between alternative approaches may have to be based on other criteria, at least in the short run.
  6. Key housing policy issues have not been addressed by past research. These include such topics as the strengths and weaknesses of mixing the SMI homeless with other population groups in the same residential setting; the potential problems of social isolation associated with a scattered-site, Section 8-type certificate strategy; or the relative merits of services being "built in" rather than "layered on."

Housing initiatives for SMI individuals across the nation typically subscribe to one of two different philosophies on how best to meet the housing needs of the SMI population: the "level of care" or "residential continuum" approach; and the "independent housing" or "housing as housing" approach. The concept guiding the former approach is to accommodate the heterogeneity of need by establishing a continuum of residential settings which varies in the level and intensity of staff supervision and program structure. In contrast, the second approach views housing "as a place to live, not a place to be treated," and as a need that remains constant over a person's lifetime (in contrast to service needs which vary over time) (Ohio Department of Mental Health (ODMH), 1988).

Most states have subscribed to the "level of care" approach to housing policy. But few provide a full residential or care continuum. During the last several years, however, there appear to be increasing instances of consideration, experimentation or implementation of examples of the "housing as housing" perspective.

Other findings from a review of existing approaches include the following:

  1. States have historically assumed the primary responsibility for filling the housing and related service needs of SMI individuals. However, housing and services are developed and delivered locally, requiring local capacity in both the private sector (e.g., housing development expertise) and the public sector (e.g., local mental health authorities). Federal initiatives should therefore be two-pronged: they should take account of the expertise and resources that currently exist at the state and local levels, but should also provide incentives to build capacity in supportive housing.
  2. Residential program staff often have little or no professional training. Some observers also believe that case management services for the SMI population need to be upgraded and increased across the full gamut of housing settings. Both of these factors raise concerns about the quality of current interventions.
  3. Quality assessment procedures for services to SMI persons need to be developed and implemented.
  4. While there are examples of successful efforts by public-private partnerships in many cities, to date these efforts have typically operated on a very small scale, assisted few individuals, and required time and expertise to package multiple and complex sources of financing. Many have also taken advantage of unique circumstances in their particular location, thereby raising questions about replicating the approach elsewhere. But more important still, combining all of these efforts together does not produce a comprehensive strategy that reaches--or has the potential to reach--all of the SMI homeless.
  5. Both housing and supportive services for the SMI population are considered to be grossly under-funded. Furthermore, in contrast to the state hospital system that typically has a single, reliable and continuous source of funds, supportive housing and community-based services for SMI individuals do not. Finally, supportive housing for other vulnerable groups such as the mentally retarded and the developmentally disabled are reported to be funded at higher and more consistent levels than initiatives for the SMI population. Taken together, these factors raise the specter that the fundamental problem is neither the absence of appropriate "models" of supportive housing nor gaps in service but rather gaps in resources to provide supportive housing to the numbers of SMI individuals who need it.

The main public sector sources of funding for supportive housing for the SMI population include state supplements to the Supplemental Security Income (SSI) program, two optional programs under Medicaid (Targeted Case Management, and Rehabilitative Services), the Social Services Block Grant, the HUD Section 811 program (formerly Section 202/8), and the McKinney Act, including the Projects for Assistance in Transition from Homelessness (PATH) program and the Shelter Plus Care program. But most of these programs provide only a portion of the funds necessary to develop and operate a supportive housing setting. Medicaid, for example, is a potential source of funds for case management, but explicitly cannot be used to cover housing costs. SSI payments for special residential settings, even in the most generous states, may be adequate to cover operating costs but are not designed to underwrite new development nor to defray any significant capital costs. There are also significant budgetary constraints on all o these programs, which affects both the depth of subsidies and the numbers of individuals who can be assisted.

As a result, the typical developer of supportive housing for SMI persons can expect to have to piece together bits of financing from multiple sources that, ironically, may not even include these mainstream programs. While weaving this patchwork of funding may challenge creativity, it is not an efficient method for achieving the goal of assisting large numbers of SMI homeless expeditiously and effectively.

There are four main challenges or lessons for housing policy that emerge from this review:

  1. Given our current knowledge, there is no single, best model for developing or operating supportive housing for the SMI homeless. As noted above, it may be that the main gap is resources, not ideas. But unless systematic studies of alternative approaches are undertaken, hard evidence to answer this question will never exist.
  2. Meeting the housing needs of the SMI homeless will require a rethinking of the boundaries of federal housing policy. For example, housing policy has a stake in assuring the access, quality and continuity of appropriate services--not just quality dwelling units. And second, if all of the SMI homeless are to be reached, HUD will need to reconsider its policies regarding individuals typically ineligible for housing assistance. A prime example are substance abusers. HUD will not house individuals who engage in illegal activities in their dwellings. But perhaps, by working with appropriate service agencies, HUD can assist individuals who are involved in substance abuse rehabilitation programs. Two recent initiatives, Shelter Plus Care and the Supportive Housing Demonstration Program, and one newly proposed initiative--Safe Havens--move in these two directions.
  3. Federal housing policy will need to find the right balance between requiring adherence to federal guidelines and standards, on the one hand, and encouraging state and local flexibility, on the other.
  4. Current approaches to financing supportive housing for the SMI homeless are inefficient and inequitable. One strategy to resolve these problems is to undertake a thorough review of all relevant policies and programs so that specific sources of overlap and inconsistency can be revealed and modified. A more radical alternative would be to start afresh with a new comprehensively funded program specifically targeted on this most vulnerable of populations. This zero-based programming approach would have at least two positive effects: it would reorient scarce resources--money, time, energy, creativeness--toward direct assistance to the SMI population and development of additional supportive housing units; and it would replace the current maze of programs and funding sources with a clear, consistent and coherent supportive housing policy for the SMI homeless.