Sustaining Housing First After a Successful Research Demonstration Trial: Lessons Learned in a Large Urban Center

As public health units increasingly turn toward evidence-based interventions, interest has grown in investigating how and why particular interventions successfully bridge the gap between research and long-term sustainability. Meta-analyses of such investigations (1,2) suggest that most evidence-based health interventions proceed with some modification to the initial model and that sustainability with high model fidelity remains rare. Common factors associated with sustainability beyond the research demonstration stage include organizational champions committed to sustainability, strong collaboration among stakeholders, and the intervention’s fit within local contexts (1,2). However, evidence regarding the sustainability of complex intersectoral interventions is limited (3,4). To our knowledge, there has been no investigation of the sustainability of Housing First, despite its widespread adoption internationally.

In 2008, the Canadian government announced an unprecedented $110 million investment to support research on approaches to ending homelessness for individuals experiencing serious mental illness (5). Over the following five years, a multisite randomized controlled trial, At Home/Chez Soi (AH/CS), demonstrated the effectiveness of Housing First in five cities with distinct homeless populations and varied service contexts (5–8). Like other Housing First interventions, AH/CS offered rent supplements and either intensive case management (ICM) or assertive community treatment (ACT) to more than 1,100 individuals across Canada. The demonstrated effectiveness of the AH/CS trial prompted the federal government to endorse the Housing First approach in its 2013 Homelessness Partnering Strategy (9). Furthermore, in 2013 Ontario’s Ministry of Health committed $4 million in annual funding to sustain Toronto’s Housing First service teams, supporting more than 250 individuals. Toronto was thus presented with both the opportunity and the enormous challenge of transitioning from a research demonstration project to a full-fledged Housing First program.

Given the paucity of work on the sustainability of Housing First over time, the goal of this research was to identify key factors in securing the long-term sustainability of Toronto’s Housing First services. In this report, we highlight some early local successes and challenges, discuss the integration of Housing First within existing service systems, and provide insights that can assist decision makers as they adopt complex interventions in other jurisdictions.

Methods

The Toronto AH/CS project included three community mental health teams and a housing team, each led by a different organization. The qualitative research study reported here engaged 36 AH/CS stakeholders between September 2014 and January 2015. We conducted key informant interviews with organizational leaders, research representatives, and one Ministry of Health representative (N=13); three focus groups with service team members (N=14); and one focus group with clients (N=9). Interviews and focus groups lasted 45 minutes to one hour and were audio-recorded and transcribed. Key informant interviews largely focused on strategies for securing sustainability and other contextual influences in the transition to a funded program. Focus groups with clients and service team members focused on experiences “on the ground,” with particular emphasis on changes in service provision. Clients were provided with two transit tokens and a $25 honorarium for their participation. The study was approved by the St. Michael’s Hospital Research Ethics Board, and all participants provided written informed consent.

Transcripts were analyzed using NVivo 10 software. Using a thematic analysis approach (10), a set of key concepts or codes was identified through initial readings of each data source. Excerpts from each transcript were assigned to corresponding codes, and these were compared within and between transcripts to ensure consistency. Codes were then aggregated into broader themes, supported by direct quotations from transcripts and interviewer field notes. Investigator triangulation during data analysis and member check-in with interviewees were used to establish trustworthiness of the data.

Results

The interviews and focus groups revealed that the demonstrated effectiveness of Housing First (5–8), early and strategic engagement of funders and policy makers, and collaborative local leadership were instrumental in securing long-term funding for Toronto’s Housing First services. However, study participants also identified emerging challenges to sustainability, including reductions in operational funding and limited oversight after the transition from the research demonstration phase to a provincially funded program.

Organizational leaders described how sustainability was a core consideration as early as 2011. A local advisory committee (LAC) was established in 2010 and included policy makers from relevant provincial and local health and housing authorities. The AH/CS leadership team made the crucial strategic choice to engage stakeholders at the provincial and local level, rather than at the municipal level (11,12). AH/CS leaders perceived that municipal service providers faced resource constraints and thus may have been apprehensive about funding Housing First services for fear of jeopardizing existing funding priorities. Conversely, as the provincial Ministry of Health funded other supportive housing initiatives with rent supplements, Housing First could be situated as complementary to existing and emerging policy initiatives. As a key informant noted, “They [Ministry of Health] knew what we were talking about.”

Through frequent briefings with the AH/CS research team, provincial representatives on the LAC had also become familiar with the rigorous methods of Housing First and the impressive improvements in housing stability among participants. After being closely engaged with the project over two years, LAC members became champions of the project at the provincial level and ultimately helped secure permanent funding. Service team members and clients also noted that the province’s support was likely strengthened by mounting public pressure to maintain housing for individuals who had already received it during the demonstration trial.

This study also found that although the federally funded AH/CS project enjoyed a high level of fidelity to the Pathways Housing First model (13), the shift to provincially funded Housing First programs resulted in a net reduction in operational funds, which led to compromises in service delivery. For example, the AH/CS project included dedicated funding for one-time expenses, such as damages, furniture and moving costs, and apartment insurance. As one key informant pointed out, these “little things . . . really made a difference to whether clients succeeded or not.” One service team member said that without dedicated funds to cover damages and other costs, “Landlords have said, ‘You know what, no clients from this program anymore.’” A client noted that without access to apartment insurance, “Your [housing] choices are cut in half.” Service team members said that without these funds, securing housing in the increasingly expensive private market was a significant challenge.

An additional challenge was the misalignment between funding for clinical support services (administered by the local health authority) and rent supplements (administered by the Ministry of Health), which resulted in unused service team capacity because of the limited number of rent supplements funded. As local authorities encouraged Housing First teams to meet expected caseloads, one key informant asked, “Is it really Housing First if it doesn’t come with rent supplements?” Furthermore, because the program’s funding arrangement tied rent supplements to support services, service teams were reluctant to discharge clients no longer in need of support services because of the absence of other adequate affordable housing options. As one informant observed, “[Clients] shouldn’t be penalized for doing well by taking away their subsidy.” This lack of flexibility, which was posed by the funding agreement, made it challenging to support clients along the recovery continuum. Another informant said, “If they could stay in their apartments, they could have been transitioned out of ICM or ACT.”

Finally, at the conclusion of AH/CS, there was no longer distinct reporting for Housing First initiatives. As a result, the unique benefits of Housing First were no longer monitored. Because Housing First can be more resource intensive than other community mental health interventions, the lack of appropriate accountability mechanisms was concerning. As one informant explained, “The [funder] doesn’t care to look at the At Home program separately. . . . If you’re trying to decrease the cost per unit of service and then you introduce a model that is more costly, then you’ve got this inherent tension.” Furthermore, losing the AH/CS project’s federal accountability framework, coupled with limited oversight by local authorities, made it difficult to identify and respond to emerging challenges in model fidelity.

Organizational leaders spoke about their long-established working relationships and their eagerness to collaborate to address emerging challenges. In the absence of other oversight, the organizations committed to new accountability mechanisms through a joint steering committee. As one informant said, “We are accountable to each other by choice, because we wanted to keep the fidelity of the program in place.” Organizational representatives have continued to meet regularly to establish service protocols, share best practices, and plan ongoing peer-led fidelity assessments (14). They have also pooled their resources to hire a housing coordinator, who is shared among the organizations, and to cover one-time expenses for clients. A service team member noted, “We support each other, all the three teams . . . . Even with the funding [for furniture and household items], it's divided among all of us . . . . We dipped into theirs or they dipped in [to ours].”

Discussion

Our findings align with the existing evidence on sustainability of evidence-based interventions (1–4). Although strong collaboration among stakeholders and the presence of local champions help facilitate sustainability, Housing First poses a challenge in terms of its fit with existing housing and mental health services. Structural and administrative divisions between the mental health and housing sectors, lack of central oversight, and limited understanding of the changing needs of the target population have the potential to undermine the sustainability and effectiveness of Housing First over time and may be of concern for many jurisdictions.

Client choice is a central value of the Housing First model (14), but stakeholders highlighted the significant limitations on service choice posed by rigid local funding and administrative requirements. Overall funding reductions and a limited number of available rent supplements diminished the range of housing options available for clients. Although local organizations successfully aligned Housing First services with existing provincial health initiatives, the lack of coordination between housing and health funding streams limited service teams’ ability to respond to clients’ changing needs and further threatened choice in the degree and type of services offered.

Coupled with oversight and accountability mechanisms that fall short of the rigor of a demonstration trial, these factors may threaten fidelity to the evidence-based Housing First model. Partial sustainability is a common outcome for health interventions during the transition from research to practice (1,2); however, stakeholders at the Toronto site expressed concern that as the model is adopted throughout Toronto and elsewhere, agencies might adopt “watered down” versions of Housing First. Although adaptations to the model may be warranted over time, a recent AH/CS study demonstrated that high model fidelity was associated with improvements in clients’ quality of life, community functioning, and housing stability (15). Despite early and strategic engagement of stakeholders, strong local leadership, and collaboration between service teams, managing these constraints will remain a key future challenge.

Our study had some limitations. The sample was relatively small, and the findings were from a single large urban center with a wide range of mental health services (5), which limits generalizability. Nonetheless, our findings are relevant in many jurisdictions facing similar challenges and may offer guidance to future efforts to disseminate and sustain Housing First.

Conclusions

Housing First has been lauded as an effective model for supporting people with complex housing and mental health needs. Ensuring that Housing First fits within existing service delivery systems and is adequately supported to maintain model fidelity remain key considerations.