COVID-19 isolation and recovery sites evaluation: Lessons learned for systems innovation

Abstract

Objective

The City of Toronto opened COVID-19 Isolation and Recovery Sites (CIRS) in 2020 in response to the need for supported isolation spaces for people experiencing homelessness. As a team of academic researchers and community partners, we assessed how lessons from the CIRS can inform post-pandemic policies for people experiencing homelessness. We focused on identifying models and systems of care for understanding existing services and integration, identifying innovations, and imagining how care can be transformed to be more caring and just.

Methods

We conducted 43 key informant interviews with clinicians, peer and harm reduction workers, executive leaders, operational leaders, and people who worked for funding or decision-making organizations. Data were coded and analyzed using deductive content analysis.

Results

The CIRS presented a new service model that moved away from being overly medicalized towards having a community orientation, which involved greater integration of services for people experiencing homelessness, transparent decision-making, and open communication. Key to this approach were clinical case management, harm reduction services, strong support of peer workers, and adequate and consistent funding. Future models will need to address power imbalances and hierarchies through formal agreements and processes as well as establishing collaborative cultures.

Conclusion

The CIRS offer an important opportunity to rethink services offered within shelter settings. Such change will require a strong vision, sufficient funding, political will, and accountability among all stakeholders. The COVID-19 pandemic offers decision makers an opportunity to consider system-level transformation by looking at largely successful elements of the crisis responses.

Supplementary Information

The online version contains supplementary material available at 10.17269/s41997-025-01031-7.

Introduction

In March 2020, the government of Ontario declared a state of emergency due to the COVID-19 pandemic, which included a directive for people to isolate from others if they tested positive for COVID-19 or were in close contact with someone who tested positive for COVID-19. This requirement for isolation posed distinct challenges for people experiencing homelessness, especially people who were staying in congregate settings such as shelters. The City of Toronto, in partnership with several organizations, opened COVID-19 Isolation and Recovery Sites (CIRS) in the spring of 2020, in response to the need for supported isolation spaces for people experiencing homelessness (Kolla et al., 2020).

The CIRS were similar to models implemented in other cities in the United States and Canada during the early stages of the COVID-19 pandemic to provide safe spaces for people experiencing homelessness to isolate (Fleming et al., 2022). In San Francisco, people who stayed in hotels had fewer emergency department visits, hospitalizations, inpatient hospital days, and psychiatric visits than matched controls without a hotel placement (Fuchs et al., 2021).

Over the course of the pandemic, three CIRS opened in Toronto: one in Scarborough, one Downtown, and one in Etobicoke. CIRS provided an isolation space for people experiencing homelessness who had tested positive for COVID-19, were waiting for their COVID-19 test results, or had close contacts with someone who tested positive. The CIRS provided clients with a private room for 24 h to 14 days as well as supports and services throughout their stay. The CIRS offered integrated health and social services delivered by staff from several partner organizations. Staff at the CIRS included peer workers, harm reduction workers, nurses, physicians, and shelter staff. Each CIRS had a distinct structure, timeline, and funding mechanism. The City of Toronto provided funds to run the Scarborough site when it first opened as a shelter/motel from March to June 2020 and it had fewer supports and services than other CIRS. A funding agreement was reached between the City of Toronto, Inner City Health Associates (ICHA), the Neighbourhood Group Community Services, and the Inner City Family Health Team to fund physician time, nursing, and peer support at the site. The Etobicoke and Downtown sites first operated as formal Hospital facilities in repurposed hotels under the Emergency Management and Civil Protections Act; in July 2021, the sites came under the operation of the City of Toronto. At the Etobicoke and Downtown sites, funding flowed from Ontario Health to the University Health Network to the partner organizations. ICHA had oversight for the budget for doctors and nurses and released funds to The Neighbourhood Group, who had budgetary oversight for peers. Parkdale Queen West Community Health Centre received funding separately from the Ministry of Health and had budgetary oversight for harm reduction workers (Firestone et al., 2021). The Etobicoke and Downtown CIRS included an onsite overdose prevention site (a room where people could use substances, primarily by injection, under the supervision of trained staff).

The City of Toronto reported that 10,460 people experienced homelessness in the 3 months prior to November 2024, defined as having used the shelter system and not moved to permanent housing (City of Toronto, 2024a). The City estimates that 74.4% of people experience chronic homelessness (minimum of 180 overnight stays in the past 365 days or cumulative recurrent overnight stays over the past 3 years of at least 546 nights). On December 18, 2024, about 14% of people experiencing homelessness were in bridging hotels and programs supported by the Canadian Red Cross (City of Toronto, 2024b). The increase in the number of people experiencing homelessness across Canada is due to a number of factors, including the rising cost of living, underfunding of supportive housing and long-term housing options, increased evictions, and a shelter system that is over capacity (StatsCAN Plus, 2023).

We—a team of academic researchers and community partners—conducted a developmental evaluation of the CIRS, which considered the changing context and populations in complex environments and the emergence of innovative approaches (Fagen et al., 2011; Patton et al., 2016). The evaluation was guided by a logic model that we developed with community partners (see Appendix in Supplementary Material), which aimed to build program theories about how CIRS operated, including how the CIRS developed, the structure of the programs and services, and how these influenced decision-making at the sites (Funnell & Rogers, 2011). The program evaluation has been previously published on our website (Firestone et al., 2021).

Within the context of the evaluation, participants provided rich information about how lessons from the CIRS can inform post-pandemic policies for people experiencing homelessness. Herein, we explore these emergent themes using a qualitative approach. Our focus is not on evaluation of the CIRS as specific programs, but rather on identifying models and systems of care that provide a framework for better understanding existing services for people experiencing homelessness in congregate settings such as shelters, identifying innovations, and imagining how care can be transformed to be more caring and just.

Methods

We recruited nurses, physicians, peer and harm reduction workers, executive leaders, operational leaders, and people who worked for funding or decision-making organizations who worked in Toronto CIRS to participate in interviews. We used purposeful sampling, drawing upon community partner collaborations. We circulated information through email, posters, and word of mouth about the study to the city organizations and institutions who were funding and operating the CIRS. Interviews were conducted by several members of the research team, including researchers with lived experience. We developed an interview guide informed by our logic model and research questions. Interview questions were designed around the strategies identified in the logic model: partnerships, funding, staffing and management, and direct service delivery and activities. The interview guide also asked participants about how and why the CIRS were meeting the needs of clients and under which circumstances. Interviews were transcribed using an automated transcription service, checked for accuracy, and analyzed using Dedoose, a web-based application for mixed methods analysis (Salmona et al., 2019).

We used a deductive content analysis approach, in which we first developed a set of predetermined, theory-driven codes and subsequently tested and iteratively refined these theories using hypothesis testing in the data and adding emergent, data-driven codes (DeCuir-Gunby et al., 2011; Young et al., 2020). The codebook was finalized through an iterative process of coding, refinement of existing codes, and addition of emergent, data-driven codes. We developed 24 unique codes organized under three categories: (1) Model Building or Conceptualization, (2) Site Functioning or Operation, and (3) Site Outcomes. Six members of the CIRS evaluation team participated in coding. Each transcript was coded by at least two members of the evaluation team. Discrepancies with coding were resolved through team discussions and having additional team members revisit relevant transcript excerpts. We then implemented thematic analysis, focusing on salient concepts both within and across codes, and identified which hypotheses were supported in our data, which were challenged, and why (Braun & Clarke, 2014; Guest et al., 2012). Our team met regularly to discuss and interpret the identified themes, including considerations of our positionality, as some team members had leadership roles in community agencies. Our study was approved by the University of Toronto and Unity Health Toronto Research Ethics Boards.

Results

Models of care

The logic model program theory identified two models of providing services for people experiencing homelessness in congregate settings (Table 1). The first, which we labelled a “medical model”, emphasized physical health needs related to COVID-19 infection, as well as other health conditions common among people experiencing homelessness, including substance use–related concerns. The second, which we labelled a “community model”, considered health and wellbeing within the context of socioeconomic determinants such as poverty and systemic racism. Our definition of model refers to the approach to caring for clients and the underlying ideas about how to address health and social issues. This definition is distinct from the organizational structure or considerations of which organization had authority for operating or funding the site. Characteristics of both models were observed across all three CIRS from the perspectives of clinicians, peer and harm reduction workers, executive leaders, operational leaders, and people who worked for funding or decision-making organizations.

Table 1.

Medial and community models of care for people experiencing homelessness

Category	Model
	Medical	Community
Decision-making	Hierarchical	Recognition of power imbalances and opportunities for dialogue and sharing ideas
Composition of executive table	People with power, particularly those with medical expertise	Individuals from all organizations, including people with lived experience
Funding decisions	Prioritize medical/clinical outcomes and impacts	Include allocation of resources for social supports
Distinctive team roles	Nursing team and primary care team provided general medical care and monitored COVID-19–related symptoms and facilitated access to medical care on-site	Workers with lived experience provided harm reduction–focused services for people who use drugs
Care for people who use substances	Substance use physicians with specialized experience provided prescription medications (opioid agonist therapy and safer opioid supply)	Harm reduction supports were integrated into housing care. Peer and community workers advocated for clients and provided harm reduction and social support
Post-discharge care	Handed off to others; if no care givers are available for hand-off, clients are lost to follow-up	Attempt to provide continuity of care across institutions and sectors through case management
Primary role for peer workers	No defined role for peer workers	Peer workers connected with new clients and conducted wellness checks

Participants

We conducted 43 semi-structured qualitative interviews, between January and June 2021, with clinicians (nurses and physicians), peer and harm reduction workers, executive leaders, operational leaders, and people who worked for funding or decision-making organizations (Table 2). For this analysis, we grouped peer and harm reduction workers together into a single category and we defined operational lead as including team leads, managers, and site leads and executive lead as including members of the executive leadership team. Finally, while several nurses and physicians had multiple roles (for example, as both an executive lead and a clinician), we categorized them by their clinical role in this report.

Table 2.

Study participants

Category	Count (N = 43)	%
Cliniciana	10	23.3
Executive leader	8	18.6
Funder/decision maker	3	7.0
Operational leader	12	27.9
Peer or harm reduction worker	10	23.3

^a5 clinicians also had executive leadership roles and 1 had an operational leader role

Results are presented around five key themes.

Theme 1: Integrated services and partnerships

Participants described how CIRS operated within broader social and health systems, including the shelter and housing system for people experiencing homelessness and the health care system, including hospitals. Historically, these systems largely operated in silos. In contrast, the CIRS offered important lessons for how services for people experiencing homelessness can integrate the work of partners across multiple sectors and organizations.

“I hope to God we don’t all go back to our corners after this is said and done, I think we’ve seen more - I’ve seen more integration across our sector throughout COVID. This is a great example, but I’ve seen it in many areas. I’ve seen us actually start to function like a system. That brings me great hope. I hope we can see our funding models continue to support that.” [Executive Lead]

Many participants reported that new partnerships had been formed and strengthened during the implementation of the CIRS and they were committed to ongoing collaboration.

Some organizations had already implemented programs based on lessons learned from CIRS, such as moving towards blended medical-community models of care with greater integration of peer workers in hospital emergency departments and expanding health care teams for people experiencing homelessness to be more interdisciplinary.

“I hope that, like in the future, particularly thinking about as we build up the harm reduction services … and as we try to have more social and peer engagement … ideally, I think if we can learn from the site … the importance of doing that in a in a way that is more collaborative and less hierarchical, that would be a really big benefit.” [Physician]

“On the harm reduction side, we’re rolling out more supervised consumption space within the shelter environment and partnering very closely with the harm reduction community in order to operate those services. We’re bringing more clinical supports into our program that offer a whole range of primary and allied health care to people who experience homelessness. So, I think there’s been great learning and things are already changing the landscape of what the broader homelessness service system looks like.” [Executive Lead]

The CIRS were successful in integrating many aspects of community care, such as harm reduction education, distribution of harm reduction equipment, telephone or in-person check-ins with residents, and referrals within their structure and services. However, they were constrained in their ability to maintain this level of care for clients who were discharged. It was particularly challenging to continue innovative harm reduction interventions, including prescribing of safer supply opioids following discharge, and to arrange ongoing follow-up care for clients with chronic physical and mental health conditions. One participant identified continuity of care as a characteristic of community care, in contrast to “handing-off” care, which was a feature of a medical model:

“Planning for discharges is probably one of the rockiest practices that we have … So looking at like the case management needs of any particular client, especially higher needs clients who may need referrals to safer supply programs out in the community as they get discharged, because, again, it’s kind of that like short term, like clinical lens or it’s like once you’re discharged, you’re not our problem anymore. Versus like … holistic care would be ensuring that people are connected before they’re discharged and that it is a problem if somebody gets disconnected from care or support because what happens to them in the community is something that we’re responsible for.” [Operational Lead]

Several participants spoke about their frustrations with discharging clients back into congregate shelter settings where they faced high risks of ill health and persistent lack of permanent housing options.

“We were discharging people back to sites where they came from, where they were concerned. They said, “Look, I got COVID because of that site”. And we sent them back to that same site literally after hearing them say, “I’m terrified to go back to that site”, we went and sent them back there. Like, it just doesn’t make sense ethically by any means. But we had to send them there because they had nowhere else to go.” [Operational Lead]

“Not having a shelter space or housing that really is a dignified space is the biggest barrier by far. And because of our limited mandate, we can’t hold people forever until a good space comes up.” [Physician]

Theme 2: How to partner and build a robust collaborative model

The models of care at the CIRS evolved rapidly over the course of the pandemic. Both the medical and the community models co-existed, to some extent, at all times within each of the CIRS. At the executive level, the different models were characterized by differences in decision-making, control over budgets, and relationships between partners, each of which evolved over time to be more collaborative and inclusive. These shifts occurred both because of formal mechanisms—including an interim program evaluation, establishing a memorandum of agreement, inviting an additional community-based partner organization to the executive table, and greater budget transparency—and informally, as the representatives for partner organizations learned from and listened to each other.

“What I saw was a real need for leaders to really have to have people’s trust. And that that it was it was really kind of a unique situation where you … had to establish and earn that trust quickly in a situation where, you know, you couldn’t have counted on people saying, oh, I know you, from like we’ve worked together for many years. We have an ongoing relationship. It kind of required that trust building.” [Executive Lead]

There was a strong consensus among participants that careful planning and transparency were required to build a partnered, integrated approach to case management and health care for people experiencing homelessness. Participants stressed that the ideal approach would avoid perpetuating hierarchical power structures and lack of accountability within existing institutions and would prioritize the needs of racialized, Indigenous, queer, and trans communities to avoid operating within, as one participant described, “institutional, oppressive and… white supremacist lens without even knowing it.”

“The medical infrastructure is a totally colonial driven experience. And so it doesn’t consider the integration of needs of racialized communities, Indigenous communities … queer and trans communities. [It] historically has rejected the needs of those groups. And so this just replicates itself without planning. And so I think we absolutely, before doing anything, need planning and consultation with all groups of people to then design and implement a new system and a new structure that’s accommodating to … a variety of people’s specific needs.” [Operational Lead]

Theme 3: Towards a new model for supportive housing

Several participants noted that the CIRS offered an alternative model not just for shelters but also for permanent housing for people experiencing homelessness. Some participants noted that CIRS helped to identify the kinds of supports that would allow people experiencing homelessness with complex needs to live in supportive housing under the “most robust model possible.” Key to this approach, which is characteristic of the community model of care, were clinical case management, accessible harm reduction services, strong peer support, and adequate and consistent funding. As one participant noted, housing models can do a “huge amount more … to support people.” Another participant noted that the CIRS lessons should prompt decision makers to consider how shelter and housing support can provide people experiencing homelessness with pathways out of homelessness.

“While we provide[d] isolation for COVID through this model, we’ve also learned what kinds of support allow people to be maintained in residential sites with high complex needs supportively and what supportive housing might look like if it actually had the most robust model possible. And so I think there were some real learnings beyond COVID isolation toward supportive housing about what kinds of support need to come together and how they can do that in a way that is mutually respectful, supportive and most impactful for clients.” [Executive Lead]

Participants also noted that this model for supportive housing could incorporate elements of both the medical and community models and include people with lived experience (peer workers) as staff.

“We started to improve people’s wellbeing through … social aspects. And I think blending those models like the medical and social kind of world … it worked really well. It was quite beautiful to see. And I think … the opportunity for peer workers as well to be involved in medical environments would be really great. Like peers could play a really great role in emergency departments and detox centres, for example. And so I think that was very helpful.” [Operational Lead]

Theme 4: Client-centred care

The CIRS prioritized client independence and dignity. Staff at the sites provided clients with non-judgemental, compassionate care in addition to basic health and shelter support. Client-centred services and support included both medical and social care, including wellness checks, access to outdoor space and social opportunities, harm reduction services, and linking clients to health, and social and financial programs in the community. The CIRS hired people with lived experience as peer and harm reduction workers; one of their main responsibilities was to conduct wellness checks at the sites. Through these regular check-ins with clients, peer and harm reduction workers established trust with clients, which facilitated reassurance for clients during their time in isolation. One harm reduction worker described the emotional supports they provided:

“Because sometimes it’s like you’re offering counselling too. So I would say if it’s changing, it just depends on because if you knock doing a wellness check and they start talking to you sometimes they will start crying. I mean the most you can do is give a listening ear so you can encourage them. Just be there. Because in this time, it’s really hard for everyone, even for the workers … So we just be there for each other in ways that we can. This is the most common way to just listen to something, listen to their concerns, listen to what they have to say.” [Peer/Harm Reduction Worker]

CIRS peer workers described their approach to client-centred care and how they adapted to clients’ needs and built relationships. Consistent with the community model, the CIRS provided a low-barrier approach to substance use, in which clients were supported in their choices and not required or expected to be abstinent from drugs. Peer workers helped clients to understand this harm reduction approach and navigate available supports.

“Whatever they needed to stay comfortable in that isolation site … was more often provided here, except medication and alcohol. We weren’t allowed to do that. But we can if they needed it and they hadn’t yet, they would say, hey, can you find out what’s going on with my medication or can you find out what’s going on with my meth so that we follow up, we ask harm reduction. Can you let me know what’s going on with this person? And they haven’t gone through medication yet? Or we bring them out for cigarettes. We … built a rapport with them while they’re downstairs enjoying their cigarette, like we stand outside with them trying to build a relationship. So it’s very, very supportive role what the peer played over there.” [Peer/Harm Reduction Worker]

Participants noted that clients felt more comfortable sharing information and requesting items such as harm reduction kits from peer and harm reduction workers than from other staff.

“For peers, it was more on an emotional level, like a more like heart to heart, where some clients really feel as open telling nursing or doctors about, you know, something that they’re struggling with. Might it be like they want a kit and they don’t feel comfortable letting the doctors or nurses know that, hey, I need a kit for their substance. They’ll usually come to peers or harm reduction and just say, hey, can I grab a kit? We’ll bring it up. We’re more of I guess they feel no judgement, kind of, because majority of them know peer support. You know, we’ve all been there. We all kind of have to understand we have lived life experience, so they can they feel more inclined to open up to us about what they’re going through or like if they need any support, they know that we will probably have an answer as opposed to nursing. And doctors might not.” [Peer/Harm Reduction Worker]

Theme 5: Barriers to building an effective model for supportive housing

Many participants identified inadequate funding as a key barrier to implementing changes towards a new housing model for people experiencing homelessness.

“It does come with a need for real funding, both for the support and for the housing and in general. While there is a now a national housing strategy that will fund a substantial amount of housing, the supportive housing models historically have not included robust clinical case management and harm reduction and peer support. They tend to be lightly supportive for the most part outside of the developed disability sector. And what we’ve learned is that you can do a huge amount more than that to support people.” [Executive Lead]

Participants noted that there was no pre-pandemic emergency plan for funding CIRS. Instead, the municipal government spent relatively freely to implement the three CIRS sites at an estimated incremental cost, at the time of interview, of $200 to $250 million. While it was politically expedient to obtain this funding during a public health emergency, decision makers were unclear whether and how future funding would support long-term solutions. Furthermore, given that health care is a provincial mandate but housing and shelters are a municipal mandate in Ontario, it was unclear which level of government would assume responsibility for funding a new model for housing.

Respondents also noted that advancing a new integrated housing model will require political leadership, collaboration across organizations and disciplines, and a commitment to a mixed staffing model. Some participants expressed concerns that implementing a system, in which providers compete for funds to implement new models, will result in powerful winners keeping most resources with no system-level change. As well, there was a need for ongoing evaluation and accountability.

“So you can’t just take the dollars and then not be accountable. That’s the other piece. And sometimes, and I think we need to set very clear expectations. Accountabilities need to align with funding. And we need to ensure that people are delivering.” [Executive Lead]

Despite the shared vision that the CIRS inspired, many participants were skeptical that actual change could happen beyond the pandemic, given the scale of the obstacles to implementation.

Discussion

The CIRS, which were an essential component of the response to the COVID-19 pandemic in Toronto for people experiencing homelessness, can serve as a case study for how to improve integration of health care and social services for people experiencing homelessness. Integration and intersectoral action has been championed by health ministries internationally to improve efficiency, enhance patient or client experiences, and address inequities (World Health Organization, & Public Health Agency of Canada, 2008; Public Health Agency of Canada, 2024; Smith & Weinstock, 2019). While much has been written about the importance of integrating social determinants of health into health and social care delivery (Raphael, 2016; WHO Commission on Social Determinants of Health, & World Health Organization, 2008), the CIRS offered a practical example of how institutions can implement such approaches in ways that directly helped clients. Our results indicate that successful integration is possible but requires both a diverse team of service providers—including health care providers and harm reduction and peer workers—and careful attention to power imbalances and hierarchies between partners, including historical imbalances in which the health sector received greater funding and stature than others. Transparent funding and decision-making processes were critical for all the success of CIRS and need to be addressed in future initiatives.

As demonstrated by the CIRS, peer workers or people with lived experience provided critically important services at the CIRS that had direct benefits for clients. Peer workers shared their uncertainty around the sustainability of their roles and the limitations around their ability to voice concerns about operational issues. Our findings are consistent with evidence generated in similar settings where expansion of formalized and resourced peer worker involvement in safe consumption programming can have an impact on overdose prevention (Kennedy et al., 2019). Despite the recognition of lived experience as being invaluable to harm reduction work, barriers around access to training, inconsistency of equitable workplace conditions, and a lack of monitoring and accountability around the abuse of power by other staff and leadership continue within the shelter system (Firestone et al., 2023). Future models will need to include people with lived experience at all levels of decision-making, including the most senior, and will require stable and allocated funding and infrastructure.

There was a strong consensus among participants that the CIRS offered a new model for thinking about the ideal provision of a broad range of supports for people experiencing homelessness who have complex needs within sheltering settings and supportive housing, that transcended the binary distinction between medical and community models. The development and evaluation of this new model will need to be co-led by community groups, health groups, and governments and must incorporate a full set of coordinated services for clients and include mechanisms for regular reporting on outcomes. The long-term implementation will require broad consultation, a strong vision, sufficient funding, political will, and accountability among all stakeholders. The alternative, however, is to return to a system widely recognized to be siloed, inefficient, and not working well for clients. As one executive lead said, “we run the risk of just going back to our own sandboxes, and that would be tragic.”

Since the pandemic, health outcomes have worsened and rates of comorbidities and mortality have increased for people experiencing homelessness in Toronto. Ontario Health has funded the Homelessness Initiative, which aims to comprehensively address sector gaps by offering integrated primary care, intensive mental health case management, harm reduction, and peer worker services. The Homelessness Initiative services are offered across shelters, drop-ins, and encampments and continue to implement some of the integrated and collaborative approaches developed within the CIRS.

While we conducted a large number of interviews from a variety of perspectives, the environment at the CIRS was dynamic and differed across the sites and participants’ experiences may have been specific to a certain time and place. Given the specific focus of the evaluation in which our study was nested, we focused solely on staff and leadership perspectives and did not collect data from clients. While this is a significant limitation of the current study, our findings resonate strongly with other components of our study that included the perspectives of people who use substances in one CIRS and of people who stayed in encampments during the COVID-19 pandemic (Boucher et al., 2022; Kolla et al., 2021).

While the CIRS experiences were specific to Toronto, many of the most important lessons and recommendations are consistent with findings from other cities (Lew et al., 2022; Brothers et al., 2022; Scallan et al., 2022; Kimmel et al., 2020). For example, partnerships between shelter operators, harm reduction groups, and health care providers in Hamilton, Ontario, were critical to the successful implementation of an integrated safer supply harm reduction program in shelters during COVID-19 (Lew et al., 2022). In Boston, COVID-19 respite and recuperation facilities that included harm reduction workers were able to provide training and support to other staff around de-escalation of critical incidences and overdose prevention as well as to facilitate essential community referrals at the time of discharge (Kimmel et al., 2020).

Conclusion

The CIRS need to be understood in the context of social oppression and the ongoing displacement of people experiencing homelessness, both of which were exacerbated during the COVID-19 pandemic, as were poverty, racism, and other forms of marginalization. Participants strongly emphasized the need for reform of the larger systems of care—shelter, housing, income, and health care, in particular—that limit the ability of any single setting, such as a CIRS, to have a meaningful impact on clients’ trajectories of health and housing. The COVID-19 pandemic offers decision makers an opportunity to consider system-level transformation by looking at largely successful elements of the crisis responses, such as the care delivered at the CIRS.

Contributions to knowledge

What does this study add to existing knowledge?

• Our study describes COVID-19 Isolation and Recovery Sites, as implemented in Toronto, a novel service model for people experiencing homelessness.

• This model was viewed by many involved parties as a successful element of the response to the pandemic, with strong implications for post-pandemic care.

• Key aspects of the model include clinical case management and integrated harm reduction services.

• The inclusion of people with lived and living experiences of homelessness and substance use as peer workers was critical to the success of the model.

What are the key implications for public health interventions, practice, or policy?

• COVID-19 Isolation and Recovery Sites are a model for thinking about how to restructure health and social services for people experiencing homelessness.

• The findings underscore the limitations of current approaches to delivering health and social care in aggregate settings for people experiencing homelessness (like shelters), including fragmented care, lack of system integration, and overly medicalized models of care.

• System change will require attention to power imbalances, strong political will, built-in systems of accountability, and dedicated funding.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Firestone, Khoee, and Bayoumi. The first draft of the manuscript was written by Firestone and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The Marginalization and COVID-19 (MARCO) project gratefully acknowledges funding from The Temerty Foundation and the University of Toronto through the Toronto COVID-19 Action Initiative/Connaught Community Partnership Research Program; The University of Toronto’s Faculty of Medicine Equity, Diversity, and Inclusion fund; and St. Michael’s Hospital Foundation.

Data availability

Source materials are not available due to privacy considerations.

Code availability

Not applicable.

Declarations

Ethics approval

This study was approved by the University of Toronto (Protocol# 39506) and Unity Health Toronto (REB# 21–012) Research Ethics Boards. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committees and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent for publication

Not applicable.

Conflict of interest

Bayoumi has received grant support as principal investigator for other projects from the Canadian Institutes of Health Research, the Ontario Ministry of Health and Long-Term Care, Ontario Health, and MITACS. He has received consulting fees from the Canadian Centre on Substance Use and Addiction. Bayoumi was supported by the Baxter and Alma Ricard Chair in Inner City Health, a joint Hospital-University Professorship between the University of Toronto, St. Michael’s Hospital, and the St. Michael’s Hospital Foundation. Kolla was supported by a Banting postdoctoral fellowship from the Canadian Institutes of Health Research. O’Campo was supported by a Tier 1 Canada Research Chair in Population Health Interventions. All authors declare no competing interests for this work.