Housing First for Long-Term Shelter Dwellers with Psychiatric Disabilities in a Suburban County: A Four-Year Study of Housing Access and Retention

This report provides preliminary outcomes from a suburban county’s Housing First implementation project, the goal of which was to reduce homelessness for the county’s chronically homeless population. Almost two years after Housing First was introduced into the county, results indicate that participants assigned to Housing First were placed in permanent housing at higher rates than the treatment-as-usual group, who had received the county’s standard array of services. Twenty months was a long enough period of time for participants in the control group to have successfully advanced along the continuum through treatment and transitional housing to reach permanent, independent housing. However, the data do not demonstrate such progress, revealing instead that most participants in the control group still had not reached the endpoint of permanent, independent housing. Meanwhile, housing retention rates show that the majority of clients in both Housing First agencies were able to end years of homelessness and were assisted in preventing a return to homelessness. Over the course of almost four years, 68% of participants who entered housing through Housing First were able to keep it without having to satisfy requirements for treatment and sobriety. Pathways’ retention rate after four years, which was just below 80%, is noteworthy, especially given the sample’s chronic homelessness and high rates of shelter recidivism and co-occurring substance use disorder. As compared to Pathways—an agency with several years of experience providing Housing First services—the lower rate of housing retention for the Consortium agencies may be reflective of some of the challenges that existing providers face when shifting their services towards Housing First.

Implementation of a Housing First approach in new locales and within existing agencies can encounter significant barriers. Such implementation difficulties were reported by Felton (2003) who documented the experiences of key stakeholders during the implementation of this “Housing First” intervention in the county. Housing First challenges traditional provider-consumer relationships by requiring clinicians and other service providers to relinquish authority in prioritizing consumers’ needs and goals. Further, it requires them to shift perspectives from an emphasis on mental health and substance use needs towards a greater appreciation of the housing needs of consumers (Meschede 2004). It requires providers to reconsider their beliefs about the capabilities of formerly homeless persons with mental illness, particularly with regard to their ability to maintain independent housing when offered comprehensive, but flexible, supports. Additionally, existing agencies and providers may be ill-prepared for the programmatic and systemic changes implied by adopting a Housing First approach that greatly reduces the need for shelters or other transitional housing programs. These providers may have very practical concerns, such as losing valuable shelter contracts, and by extension, jobs for shelter staff.

Providers new to Housing First must also be aware of ways in which their practices may deviate from some of the essential features and philosophy of Housing First. The Consortium’s lower retention rate suggests that their discharge policies may not reflect the practice of separating housing from treatment. It is important to continue to provide services through housing loss and to assist consumers in finding new housing when they experience difficulties in one building or neighborhood, or upon their discharge from hospital or clinic-based treatment. Providers shifting to Housing First services must, therefore, be especially observant of the need to keep clinical matters separate from housing matters and to ensure that a clinical crisis results in the consumer receiving intensive clinical services, not being evicted from housing.

The disparity between the Housing First programs in the ratio of clients housed to those outreached/engaged suggests that the agencies used two different approaches to enrolling participants and placing them in housing. The large number of participants engaged by the Consortium may suggest that these agencies were extremely rigorous in their efforts to screen-out ineligible applicants. Also possible, however, is that the Consortium’s selective enrollment was the result of clinicians turning down participants who were eligible but whom they did not consider appropriate for immediate placement in permanent housing. New Housing First providers may still be reluctant to work with consumers who are traditionally considered “difficult to house.” One of the principles of Housing First is to target consumers who have had difficulty accessing traditional services and to then sequentially accept these consumers on a first-come, first-served basis. Providers who are shifting towards Housing First services must, therefore, be mindful of a long held but erroneous bias that equates psychiatric symptoms or substance use with an inability to maintain housing. Given that the Consortium had lower rates of housing retention despite carrying out a more extensive selection of consumers also reinforces the fact that housing providers and clinicians are not able to successfully predict which consumers among a chronically homeless pool of applicants will be able to successfully maintain housing.

With regard to implementing a Housing First approach based on a scattered-site, community integration model in a suburban or rural locale, service providers may encounter several challenges that could require slight modifications to the model. With regard to staffing, if sufficient resources for a full-scale ACT team operating within one agency are lacking, or if the number of consumers to be served is small, programs can create smaller sub-ACT teams that maintain low caseload ratios but must broker some services from agency or community providers. Another variant successfully implemented in some cities consists of ‘composite teams’ comprised of several staff members but each from a different agency (e.g., a mental health expert from the local mental health clinic, a substance abuse specialist from the drug treatment program, and a case worker from the shelter). Programs may also employ intensive case managers who provide essential supports to consumers and then broker other specialty services. Housing First staff working in suburban or rural areas must also traverse greater distances to visit consumers. Staffing patterns may shift in order to address this issue: for example, teams may need more staff or teams may need to be redundant with regard to staff with essential specialties, such as nursing. This would require greater financial resources for staff and consumer travel, purchase of extra vehicles, additional staff, or reimbursing staff for use of their personal vehicles.

With regard to housing stock, affordable housing may be in more limited supply in non-urban areas and studio or one-bedroom apartments may be in short supply. In these areas consumer choice may be restricted to the limited options available and consumers may have to choose between renting a room in someone’s home, sharing a house with one or two other consumers, or waiting until an individual unit is found. Additionally, rural and suburban areas do not offer the same level of anonymity to consumers as do urban areas, but they may be advantageous in countering the loneliness and alienation that some consumers experience living alone in large cities (Yanos et al. 2004). Finding the right type of affordable housing in non-urban areas, therefore, may be more difficult and time-consuming, requiring greater up-front efforts and investment in securing apartments and other options for potential consumers, as well as implementing an effective public relations and education campaign about the program for community members and key stakeholders. Although tight housing markets or lengthy distances for staff to travel may spur some programs to consider placing consumers in shared or congregate settings, it is important to note that, unless such arrangements are explicitly based on the consumer’s preference, they represent a significant deviation from the Housing First model. Housing First programs are successful because they offer consumer-driven solutions to end homelessness, and in most cases, consumer choice dictates provision of scatter-site independent apartments, an approach to ending homeless that is also consistent with broader goals of promoting integration and recovery.

The study reported outcomes through just fewer than four years for participants randomly assigned to Housing First, and included a treatment-as-usual control group at the 20-month time point. Such longitudinal outcomes are uncommon among studies of formerly homeless persons. Additionally, two different programs were implementing the Housing First approach, allowing for a comparison of the number of persons engaged and the housing retention rates. Discrepancies between the programs suggest that the agencies were not operating under all of the same procedures for enrollment and discharge, two areas in which the Housing First approach makes critical departures from traditional programs. Enrollment of targeted chronically homeless individuals with serious mental illness on a first-come, first-served basis and providing services through housing loss or housing transitions are fundamental aspects of Housing First. It is essential to ensure that agencies adopting a Housing First approach implement it in the form in which it has demonstrated the greatest effectiveness. The study sample was comprised of individuals who were chronically homeless, had severe mental illness, and did not benefit from traditional services. The high rates of co-occurring disorders among this sample further confirm that Housing First approaches can be successful with persons who experience multiple impairments.

Nevertheless, the study also has considerable weaknesses. First, demographic data were only available for the first cohort of participants who enrolled into the study and so we cannot accurately describe the entire study sample. Second, despite employing random assignment, not all participants were enrolled by their respective Housing First agency. Though the initial groups were roughly equivalent after random assignment, unfortunately, individual-level data were not available to compare those who were actually housed within each study condition. Consequently, it was not possible to determine how comparable the groups of participants were who actually received housing. Further, the absence of such data made it impossible to determine whether there were any significant demographic differences between those who were housed and not housed within each condition. We cannot, therefore, identify the characteristics that are associated with entry into housing for each condition and across the sample as a whole. Further, almost half of the control participants’ whereabouts were unknown at time of follow-up, resulting in substantial amounts of missing data. This weakens our ability to estimate the relative strength of the Housing First approach as compared to the usual sequence of services. Limited resources also did not permit continued follow-up of the control group over the entire 47-month study period. Though unlikely, it is impossible to determine whether the control group caught up with or exceeded the rates of permanent, independent housing reported here for the two Housing First groups after four years.

A final limitation is that the impact of specific agency support and treatment services that were received by participants and their role in maximizing housing retention, were not examined. Because both agencies were funded under the same mechanism, they were very similar in terms of their ACT teams’ organizational structure, staffing, and general practices. Given the disparities in housing retention and selection outcomes across agencies, these structural similarities accentuate the potential impact that overarching agency philosophies and more informal, daily team support services may have on consumer outcomes. For example, the Consortium’s affiliation with a medical center may have resulted in their ACT team having a lower threshold of tolerance for psychiatric symptoms and/or substance use among their clients. Such a service perspective may have led to greater residential instability if consumers were more abruptly removed from housing and their engagement with the team threatened by having their ability to exercise choice more restricted. Unfortunately, the current study was not able to examine these potential philosophical and services discrepancies between agencies in-depth.

Overall, Housing First has proven to be an effective and less costly alternative for housing chronically homeless individuals with psychiatric disabilities. This study demonstrates that the Housing First approach is effective in the long-term in reducing homelessness and can be successfully implemented in suburban areas and with populations of chronically homeless shelter users with multiple disorders. Other Housing First replication sites also report housing retention rates of 80% or better through 12–18 months (D. Dunbeck, personal communication, December 2006). Officials in the county described here have corroborated the positive outcomes. With Housing First as a vital component of the county’s overall approach to ending homelessness, it was recently reported that the county had reduced homelessness by two-thirds over a period of approximately five years and was considering “a top-to-bottom shift to the Housing First model” (Scharfenberg 2006). Considering that certain modifications may need to be made to the Housing First model in order to adapt the program to non-urban locales, it is imperative to understand which adjustments are effective and acceptable, and which changes represent unacceptable departures from the model’s standard operation and philosophy, and result in poorer client outcomes. Essential to shifting services, in this county and elsewhere, will be to clearly articulate the principles and philosophy of Housing First. To guide program development and to ensure that providers are indeed implementing a Housing First model, it is necessary to develop valid measures that will assess fidelity to Housing First, differentiate it from other models, and further identify the essential ingredients that are associated with positive outcomes. Given the research to date, Housing First holds promise for ending and preventing homelessness and promoting community integration and recovery.