Nowhere to go: exploring the social and economic influences on discharging people experiencing homelessness to appropriate destinations in Toronto, Canada

Abstract

Objectives

A main component of discharging patients from hospital is identifying an appropriate destination to meet their post-hospitalization needs. In Canada, meeting this goal is challenged when discharging people experiencing homelessness, who are frequently discharged to the streets or shelters. This study aimed to understand why and how the ability of hospital workers to find appropriate discharge destinations for homeless patients is influenced by dynamic social and economic contexts.

Methods

Guided by critical realism, we conducted semi-structured, in-depth interviews with 33 participants: hospital workers on general medicine wards at three urban hospitals; shelter workers; and researchers, policy advisors, and advocates working at the intersection of homelessness and healthcare.

Results

Historical and contemporary social and economic contexts (e.g., shrinking financial resources) have triggered the adoption of efficiency and accountability measures in hospitals, and exclusion criteria and rules in shelters, both conceptualized as mechanisms in this article. Hospitals are pressured to move patients out as soon as they are medically stable, but they struggle to discharge patients to shelters: to prevent inappropriate discharges, shelters have adopted exclusion and eligibility rules and criteria. These mechanisms contribute to an explanation of why identifying an appropriate discharge destination for people experiencing homelessness is challenging.

Conclusion

Our results point to a systems gap in this discharge pathway where there is nowhere for people experiencing homelessness to go who no longer need acute care, but whose needs are too complex for shelters. Systemic changes are needed to better support hospital and shelter frontline workers to improve discharge processes.

Introduction

The growing number of homeless individuals worldwide is a pressing public health concern. Around 150 million people globally are homeless, accounting for 2% of the total population (Chamie, 2017). Housing is a key social determinant of health (SDoH), whereby those who are homeless or vulnerably housed experience extremely poor health. Extensive research shows that homeless populations suffer from higher rates of multiple morbidities (Hwang et al., 2013) and mortality compared with the general population (Frankish et al., 2005). These health inequities have been attributed to experiences of social exclusion, defined broadly as the “processes driven by unequal power relationships that interact across economic, political, social, and cultural dimensions” (Popay et al., 2008, p. 7). Such processes include social and economic policies that lead to the inequitable distribution of SDoH, such as income, food, education, employment, and housing (Ruckert & Labonte, 2014). For instance, divestment from areas such as housing supports can create or exacerbate social and health inequities (Ruckert & Labonte, 2014), thus contributing to a growing burden of homelessness and associated poor health outcomes (Stuckler et al., 2017).

In Canada, where roughly 235,000 people experience homelessness each year and 35,000 on any given night (Gaetz et al., 2016), the federal and provincial governments have made cuts to health budgets and public health, and social services such as social assistance and housing, intermittently since the 1980s (Ruckert & Labonte, 2014). Although these policy decisions have been implemented in response to economic crises and budget deficits, they are nonetheless troubling. Research indicates that combined expenditures on health and social services have positive impacts on health outcomes (Ruckert & Labonte, 2014), and austerity-driven policy decisions during economic crises worsen health, particularly for the most vulnerable such as homeless individuals (Stuckler et al., 2017).

As people experiencing homelessness have high rates of morbidity, they have a concomitant high need for healthcare services; however, they face multiple barriers to accessing primary and preventive healthcare and frequent hospitals at higher rates than the general population (Hwang et al., 2013). Homeless patients often have complex health and social needs, requiring complicated discharge planning that involves service coordination across multiple sectors and service agencies. Discharge plans that include housing, physical and mental health services, and substance use services can improve recovery post-discharge, shorten lengths of stay in hospitals, and decrease hospital readmissions (Backer et al., 2007). Yet, such supports are not readily accessible for homeless patients who are frequently discharged from hospitals to emergency shelters or the streets (and less often to other destinations such as long-term care or medical respite) (Forchuk et al., 2013). These discharge destinations lack the social and health-sustaining resources needed for recovery, exacerbating health inequities faced by this population while highlighting a key challenge in the discharge process: identifying and discharging homeless patients to an appropriate discharge destination.

This paper responds to key gaps in the literature. The discharge process for people experiencing homelessness is not well characterized in the Canadian context. Research primarily examines hospital length of stay and discharge outcomes such as hospital readmission (Hwang et al., 2011; Hwang et al., 2013; Saab et al., 2016), or discharge from long-term and acute mental healthcare (Forchuk et al., 2013). Less research focuses on discharge after hospitalization for physical health conditions, despite frequent admissions and long lengths of stay for homeless patients admitted to medical and surgical services (Hwang et al., 2011). A small number of studies have described barriers or challenges to discharging homeless patients from acute care (e.g., Buccieri et al., 2018), but are not explicit on why and how these challenges exist. We seek to add to current knowledge by providing further explanations for how and why discharging homeless patients from general medicine to an appropriate discharge destination remains challenging.

Methods

Philosophical framework

We undertook an exploratory critical qualitative project, employing a critical realist framework to characterize the hospital discharge process for people experiencing homelessness. Critical realism is a philosophy positing that the world is made up of social structures and their mechanisms, which we cannot see but come to know through the phenomena they produce (Lawson, 1997). Mechanisms are the causal powers embedded within structures (physical objects or social processes). They can be triggered within certain contexts to generate social phenomena (Lawson, 1997). We understand context as “the local mix of conditions and events, social agents, objects and interactions which characterize open systems” (Poland et al., 2008, p. 309). Within critical realist research, mechanisms can be identified through examining the relationship between social structures (e.g., norms, resources, policy, institutional practices) and agency (e.g., individual action and sense of identity) (Poland et al., 2008). Agents’ actions are constrained by structures; over time, their actions, intentionally or otherwise, can also either reinforce or change structures. This offers researchers ways to understand social reproduction and change in society. Methodologies derived from critical realism see that “institutions in the health, educational, prison, and other social domains are micro-scale versions of society” (De Souza, 2013, p. 142), supporting a critical realist approach to examining a process within healthcare and social institutions.

Study settings and participants

This study took place in Toronto, Canada’s largest city with roughly 8715 people experiencing homelessness (City of Toronto, November 28, 2018). We employed purposive and snowball sampling to recruit 33 participants: 11 key informants working on issues of homelessness in advocacy, research, middle management, or policy positions; 16 hospital workers across three urban hospitals; and 6 shelter workers (see Table 1 for participant demographics). All three hospitals regularly serve homeless patients and are in different geographic regions of the City of Toronto. Shelter participants did not always disclose where they worked. The disclosed locations included different types of shelters (e.g., men’s only and women’s only shelters) in diverse geographic locations across the City of Toronto. These details hold implications for how this research might be understood in other contexts.

Table 1.

Participant demographics

	Hospital workers (n=16)	Shelter workers (n=6)	Key informants (n=11)	Total (n=33)
Age (years)
24–29	2	3	-	5
30–39	7	1	2	10
40–49	4	2	3	9
50+	3	-	6	9
Gender participant identifies as*
Female	11	5	6	22
Male	5	1	5	11
Education
Undergraduate	4	2	3	9
Graduate	12	4	8	24
Occupation
Physician	3	-	-	3
Nurse	2	2	-	4
Social worker	5	2	-	7
Other (e.g., outreach counsellor, supervisor/manager, occupational therapist, etc.)	6	2	11	19
Time in current position (years)
0–5	9	4	6	19
6–9	1	1	2	4
10–14	2	1	-	3
15+	6	-	3	9
Time in occupation (years)
0–5	2	2	-	4
6–9	2	2	1	5
10–14	6	-	-	6
15+	6	2	10	18
Prior to current position worked in a hospital? (yes/no)	Y = 7 N = 9	Y = 4 N = 2	Y = 5 N = 6	Y = 16 N = 17
Prior to current position worked in the community? (yes/no)	Y = 11 N = 5	Y = 5 N = 1	Y = 7 N = 4	Y = 23 N = 10
Prior to current position work(ed) with homeless? (yes/no)	Y = 15 N = 1	Y = 6 N = 0	Y = 9 N = 2	Y = 30 N = 3

*Gender was asked as an open-ended question (“What gender do you identify as?”). All participants answered either male or female

Data generation

Semi-structured, in-depth interviews were conducted by the first author between November 2018 and April 2019. Interviews with hospital and shelter workers focused on the practice of discharging or receiving people experiencing homelessness, respectively; their understandings of a successful discharge; and what facilitated or impeded this type of discharge. Interviews with key informants situated the discharge process within broader social and economic contexts.

Qualitative data analysis

Transcripts were uploaded into Dedoose Version 8.2.14 online coding software. Analysis was iterative and involved movement back and forth between identifying demi-regularities, abduction, and retroduction (Fletcher, 2017). Demi-regularities are tendencies that can be identified in the empirical data through qualitative coding, abduction is the process of moving from individual experiences to a more general social phenomenon, and retroduction moves from describing events to identifying possible causes and conditions of the social phenomenon. We used a flexible deductive approach to coding that pre-determined codes of interest, but allowed space for in vivo codes derived from the data that better explained the discharge process. Deductive codes were predetermined based on the hospital discharge literature (i.e., safe discharge, effective discharge, high-barrier), CR concepts (i.e., structure, agency), and broad domains of context from a CIHR/NIHR guidance document on context in intervention research (i.e., social and economic, organizational) (Craig et al., 2018). As analysis proceeded, these codes were added to, changed, eliminated, or exchanged with codes that were drawn from the data (i.e., bed flow pressure, exclusion criteria). From this final list of codes, dominant codes were identified as demi-regularities in the data.

Dominant codes were then redescribed using existing theories and concepts, a process of abduction. It involved “trying on” multiple explanations to see if existing theory explained our findings or could be built on using our data, a form of theoretical triangulation. We drew on concepts and theories from the critical social sciences (i.e., neo-liberalism, stigma, moral distress, etc.), interpreting “single events as expressions of more general phenomena” (Blom & Moren, 2011, p. 69). Those that best explained our data have been used to frame our results and are woven into the results and discussion sections.

Finally, we employed retroduction. The final codes were grouped into context (C), mechanism (M), or outcome (O) categories, and then arranged into linked CMO configurations, an analytic tool used to illustrate causal relationships (see Table 2). This was done following an approach used by Sayer (1992, as cited in De Souza, 2013) to move “from observable actions ➔ to reasons individuals give for engaging in action or non-action ➔ to finding formal and informal rules existing in the context of social action within which the reasons given make sense ➔ to identifying the existing social structures (or the aspects) that generate these rules, influencing the perceptions and actions of agents” (p. 145). Descriptive and analytic field notes were often the starting point for generating analytic ideas and have been integrated into the general analysis.

Table 2.

Results—CMO configuration

Context	Mechanism	Outcome
Social  • Aging population, adding pressure to healthcare system  • Increasing homeless population Economic  • Cuts to healthcare  • Cancelling Canada’s national housing plan (a new National Housing Strategy was implemented in 2017)  • Cuts to social services, including shelters	• Policies and accountability measures (efficiency metrics) established in hospitals • Shelters develop exclusion and eligibility criteria, and occupancy rules for clients, to prevent inappropriate patient discharges from hospitals to shelters	• Hospital workers pressured to discharge patients quickly • Limited time to address social determinants of health, such as housing • Shelters reject discharge referral requests • Shelters receive patients with needs inappropriate for their services/resources • General outcome: Challenges finding an appropriate discharge destination for people experiencing homelessness

Ethics

This study was granted ethics approval from the Toronto Academic Health Science Network (TAHSN) Board of Records REB, and administrative approval through the University of Toronto.

Results

Our results suggest identifying an appropriate discharge destination for people experiencing homelessness is complicated by hospital policies and accountability measures that prioritize resource and time efficiency, and eligibility and exclusion criteria, and occupancy rules imposed by shelters that prevent many patients with complex needs from being discharged to their care (see Table 2). Key informants framed these mechanisms in what we have characterized as broad economic and social contexts. Results were consistent across workers at all three hospitals and with shelter workers, pointing to a common challenge across healthcare and social service sectors. In presenting the results, we first discuss the social and economic contexts that triggered the two main mechanisms we identified, which follow: resource and time efficiency, and eligibility and exclusion criteria, and rules.

Social and economic contexts

Hospital and shelter workers pointed to the broader pressures they each faced within healthcare and social service systems, respectively. While hospitals had limited resources to meet increasing health and social needs and were frequently overcrowded, shelter workers were poorly supported to meet the health needs of homeless patients, and their facilities were often operating at or over occupancy capacity. Within hospitals, resources were inadequate for the increasing needs of the population. While “admissions have gone up about 50% [over the last 7 years] ... the number of beds that we have been allocated hasn’t gone up 50% ... it’s gone up minimally” (P_10, Hospital).

Hospital workers discussed hospital resource constraints primarily in terms of limited bed space. When treating people experiencing homelessness, who can require longer stays in hospital to address their complex health needs, hospital workers were challenged to justify extending stays based on their social needs once patients were medically stable. When contending with general patient surges, delaying discharge for homeless patients for non-acute or non-medical reasons (i.e., chronic conditions or housing) meant fewer beds were available for other patients. Shelter workers were equally pressed to meet the demand for services and the complex needs of individuals discharged to their facilities:

Emergency shelters are being asked to be a shelter, a transition house, a counseling center, all these things. I think a lot of it is the capacity of the shelter, what they’re able to do, what qualifications staff have to be able to do that sort of stuff. (P_S8, Shelter)

Both shelter and hospital workers pointed to the inadequate supply of affordable housing as a major challenge for discharging homeless patients to appropriate destinations. While many referred to Housing First (https://www.canada.ca/en/employment-social-development/programs/homelessness/resources/housing-first.html) as a promising model for this patient population, they also noted that it was near impossible for them to implement. As a key informant working in the shelter system explained, “There’s a gap between that philosophy [Housing First] and the access to housing that would help us follow through on that … even when a caseworker is involved and has filled in every form, the chances of getting anyone transitioned into housing is rare” (KI_1b, Shelter System). Without viable and accessible housing options, hospital workers frequently discharged homeless patients to shelters. Although hospital workers worried about this discharge disposition, they were unable to keep patients in hospitals for housing reasons.

Our data point to additional ways the social and economic contexts complicate finding an appropriate discharge destination for people experiencing homelessness. Within healthcare, budgetary cuts have led to a “paradigmatic shift towards auditing” (Whiteford & Cornes, 2019), where hospital transfer payments are informed by accountability measures such as time and resource efficiency metrics: “You get paid if you meet your metrics, you don’t get paid if you don’t” (KI_9, Hospital). Such metrics include reduced in-patient length of stay and reduced hospital readmissions within a certain time period (see below). Within the shelter system, to protect against receiving discharged patients they cannot care for, and in response to limited operating resources, some shelters have developed exclusion and eligibility criteria (i.e., individuals must be self-ambulatory) and occupancy rules (i.e., no smoking or substance use). Such measures can create high-barrier spaces that are not suitable for many patients with complex needs.

Resource and time efficiency

Hospital workers addressed the pressure hospitals are under to meet performance metrics, such as lowered readmissions and shorter lengths of stay. This pressure was often downloaded onto them as front-line workers. They referred to resource and time efficiency, bed flow pressures, the need to keep the system moving, etc., as challenges they faced.

I put pressure on my discharge planners because we gotta get them out, because I am accountable to those numbers, I’m accountable to that turnover for the length of stay… we’re pretty actively trying to get people out as quickly as possible, we really do pride ourselves on that, but sometimes it’s just the system has nowhere for them to go. (P_3, Hospital)

The same participant explained “they’re [hospitals] cutting our positions … they’re getting rid of the social workers, they’re getting rid of discharge planners at other hospitals” (P_3, Hospital). Less staff on the front-line, combined with efficiency and bed flow pressures, can further exacerbate the situation, especially for a high-needs population. Such pressures also meant patients were to be discharged when they are medically optimized, which frequently left little time to address their social needs such as starting housing applications (“The context of how quickly our turnover happens, we can’t find housing for people ... we don’t have the luxury of time” (P_12, Hospital)), or even develop a holistic discharge plan (“They’re so pressured to get them out that they often can’t even do a thorough plan with them because they need to get the next patient in” (P_S19)). The push for quick turnaround times in hospitals stood in stark contrast to shelter workers who suggested that having more time to participate in discharge planning would improve their ability to implement the patient’s discharge plan, which might include coordinating follow-up care, helping patients get to follow-up appointments, accessing prescription medication, etc. 1

Failed discharges had different meanings among our participants, but among hospital participants this commonly referred to when a patient was readmitted to the hospital for the same or connected health issues within a certain time frame (this differed between hospitals). Some hospital workers experienced pressure to quickly discharge these patients without a specialized plan:

Some of our patients who come in frequently [pause] there isn’t the same patience once they’ve been in a few times … I had someone the other week who management encouraged me not to see because so many other things had been tried in the past, so once he was stable, they told me that I had an hour before he would be leaving the hospital, which with housing [laughs] you could spend a week on something and still not have somewhere for someone to go. (P_1b, Hospital)

This description reinforces the struggle to house patients but also implies an underlying assumption on the part of hospital management that failed discharges result from patient characteristics. Individuals are often held responsible for their homelessness or their inability to follow discharge, treatment, etc. directives (Lyon-Callo, 2000). However, shelter workers were adamant that discharge plans misaligned with the resources homeless patients have access to were doomed to fail. For example, patients discharged to shelters experienced structural barriers to receiving follow-up care: “Even when the follow-up includes … in-home care, whether it’s wound change, dressing changes, that’s one of the places where it kind of falls apart because the organizations that provide that care don’t view Shelter as an address” (P_S19). Without the necessary follow-up care, an individual’s health may worsen, leading to hospital readmission.

At times, hospital workers met system pressures with resistance, pushing back against managerial pressures and keeping admitted patients longer in hospital than medically necessary.

When they [discharge planners] call shelters, 50% of the time they’re full. So, guess what, we’re going to keep them at Hospital, just for one more day and check again tomorrow. That happens a lot as well…We keep them here because there’s no place that can take them. They are full. (P_16, Hospital)

Some of our participants spoke of value-conflicts they experienced when systems pressures intersected with patients with complex needs. In the face of these pressures, hospital staff experienced moral distress, where they identify “the ethically appropriate action but that action cannot be taken” (Epstein & Delgado, 2010, as cited in Georgiadis et al., 2017). In addition to facing pressures to discharge patients when they were medically optimized, hospital workers faced great challenges finding somewhere to discharge patients. Potential discharge destinations could include rehabilitation hospitals/facilities, medical respite, long-term care facilities, or shelters; however, hospital workers explained that discharging to any of these locations could be complex given their limited number of beds and extensive rules and regulations.

There’re not enough long-term care homes, and … [they] won’t manage people who have smoking or “behavioural” needs... So, in the hallway here there’s a whole bunch of guys hanging out that have been rejected because once upon a time … they were seizing and they hit someone. And then it’s months of telling them [long-term care] “no, they’re fine”. And then 10 homes rejecting them before reapplying again and then finally getting maybe 1-2 homes accepting them. (P_3, Hospital)

Without a final fixed discharge destination, most patients are not eligible for programs such as rehabilitation or medical respite. Participants addressed the need for alternative and long-term options to avoid discharge to inappropriate destinations. As one hospital worker noted, they felt that there was little they could currently do to improve hospital discharge: “If we had a different level of care, if we had something that was somewhere between a hospital bed and a hostel bed, we might be able to do something” (P_10). With few other options, shelters become the main discharge destination for homeless patients.

Exclusion and eligibility criteria, and rules

Changes to the healthcare system and cuts to social services have also impacted the shelter system. According to our key informants, shelters have become a “catch-all”, receiving hospital patients who have nowhere else to go:

We become the last stop and there’s no other option and we wouldn’t always be in the best position to adequately service the needs of people who present, but with no other option that is what happens and that is a challenge (KI_2b).

Shelters are often the go-to discharge destination as interim care facilities can present challenging barriers, noted above. Hospital workers characterized finding an appropriate discharge destination as a web of rules and restrictions:

Recently I had a homeless patient who jumped off a fence and broke both of his heels. [He] needs total care … most places that provide that total care for the 6-8 weeks … require a discharge disposition [a guaranteed place he will go after his stay at this interim facility]… [W]e can’t send him there without knowing that he has housing afterwards; this patient is also not able to go to a regular homeless shelter because he can’t physically care for himself and he has to be in a wheelchair for the next 6 weeks, even though he’s 25 years old. (P_2, Hospital)

Except for two participants who were nurses, shelter worker participants were not trained medical professionals, yet they still received patients who can have complex health issues. Shelter workers described situations where they received patients with infectious diseases (e.g., Strep A), mobility issues, or other needs that require a Personal Support Worker (PSW). Shelter workers suggested that hospitals would knowingly discharge patients to inappropriate spaces that could not match their needs, such as shelters or drop-ins, to get them out of the hospital:

He was in a wheelchair. He showed up here, and a worker from Hospital was with him and basically was, “Okay. He’s being discharged from the hospital. We’re trying to find him somewhere to go, but we haven’t been able to yet.” He could not go to the bathroom by himself, he couldn’t get in and out of his chair by himself, and you’re sleeping on the floor here. They had arranged for CCAC to come twice a day. They would come in the morning … bring him to the bathroom… and then leave him sitting in his chair for the day. [T]hen they would come back at night and help him go to the bathroom and get down on his mat to go to sleep ... Anyway. The guy just disappeared. He wasn’t here for very long... In that case, the hospital knew. (P_S5, Shelter)

Participants reported that to prevent hospitals from discharging patients with complex needs to shelters, some shelters had reacted by developing and implementing exclusion or eligibility criteria, and many have operating rules. Operating rules included curfews and prohibition of substances, which can deter some patients from accepting their discharge to certain discharge destinations and can act as high-barrier spaces. Key informants pointed to the condescension imbued in such rules, identifying the paternalistic experience of, for example, having a curfew as an adult. Exclusion and eligibility criteria imposed by shelters were another challenge to patient discharge. Criteria included expectations that patients are independent, can walk stairs (if necessary), do not require oxygen tanks, and so on.

We actually had to end up putting some exclusion criteria because we were getting a lot of inappropriate discharges, like people were still on oxygen and we, of course, have clients that sometimes smoke in the building and so we couldn’t have people on oxygen or people with PICC lines or people that were not self-transfers and they were discharged to shelter. (P_S19, Shelter)

Participants suggested criteria were aimed at preventing inappropriate discharges, and that shelters were simply not designed nor mandated to address complex health needs. Criteria were described as ways to protect both individuals currently residing in the shelter and those discharged to their care by ensuring staff are not overburdened. These criteria were also characterized as push-back mechanisms, an attempt to reinforce that emergency shelters are not a discharge solution for this patient population with complex needs.

Shelter workers expressed frustration that even when trying to enforce these exclusion criteria, hospitals might discharge patients to them anyway. For example, one participant describes a situation where the hospital ignored their referral refusal:

We had one particular client who was in hospital for quite some time, had really serious health issues, and our management team made a decision that she wasn’t supposed to come back here because her health needs were too high for our space ... Her health deteriorated in our space as well, so it was not good for anybody. And they discharged her back to us twice, and we had kept letters ... Like, formal letters from management saying not to do so, and they sent her back. It’s really unfair to the client. (PS_7, Shelter)

This quotation demonstrates the power hierarchy between hospitals and shelters. Hospitals might ignore shelter directives prohibiting a discharge to them. In this paper, we have characterized such actions as responses to systems pressures created by the need to meet efficiency metrics.

Discussion and conclusion

Coordinated and comprehensive discharge planning for people who are experiencing homelessness can help improve their well-being and prevent hospital readmissions. A key part of the discharge process is referring patients to accommodations, programs, or services to meet their needs; however, in this paper, we highlight how social and economic contexts, over time, created challenging conditions for achieving this in Toronto, Canada.

The adoption of neo-liberal policies since the late 1970s, combined with austerity-driven policy responses to three economic recessions beginning in 1981, 1990, and 2007/2008, have contributed to an increase in the homeless population and concomitant health inequities (Gaetz, 2010), and have caused resource strains in the healthcare and social service sectors (Ruckert & Labonte, 2014). Such federal and provincial policies include but are not limited to cuts to affordable housing programs and the end of Canada’s national housing strategy (a new strategy was announced in 2017 (https://www.placetocallhome.ca/)), cuts to healthcare spending and social assistance (Ruckert & Labonte, 2014), and over-reliance on and under-funding of homeless programs and services, such as shelters, by municipal governments (Gaetz, 2010). Added to this is a large aging general population that requires higher levels of healthcare and a housing crisis nearly eliminating affordable housing options, leaving many individuals who experience homelessness to rely on the emergency shelter system.

Our results point to a systems gap where there is nowhere for people experiencing homelessness to go who no longer need acute care, but whose needs are too complex for shelters. Hospitals have limited resources to meet increasing needs and are frequently overcrowded (Premier’s Council on Improving Healthcare and Ending Hallway Medicine, January, 2019), and shelters, which are best suited for clients who are temporary and self-sufficient, are operating at capacity (City of Toronto, N.D.). Permanent and supportive housing is an ideal discharge location, but for participants in our study, the inadequate supply of affordable housing was a major challenge for discharging homeless patients to appropriate destinations. Other Canadian studies report that “the lack of affordable housing was believed to undermine efforts at discharge planning”, without which patients will continue to be discharged to shelters (Buccieri et al., 2018, p. 10). These patients can have complex needs, adding strain to an already stretched shelter system. Research indicates that individuals with complex needs represent a fraction of the shelter population, but use a disproportionate amount of the services (Aubry et al., 2013). Other discharge destinations, such as medical respites, are associated with lower odds of hospital readmission and are better suited for individuals with complex health needs, but many programs will not admit patients who use substances or whose behaviour is considered challenging (e.g., those with mental illness who may lash out physically, or those who are unable to adhere to curfews, such as sex workers) (Doran et al., 2013). While discharge planning aims to help decrease readmissions, Cornes et al. (2018) note that discharge destination is the strongest correlate of hospital readmission for this population and suggest that “lack of system capacity in the community, geared to the realities of persons’ needs, equates to acute care being a fallback option” (p. e353). We add that shelters become the fallback option for hospital discharge destinations. With a lack of systems planning to address the multifaceted needs of people experiencing homelessness, the discharge process contributes to a scenario where the health of people experiencing homelessness is likely to get worse.

Within public health sciences, the contribution of critical realist research is to “generate richer conceptualizations and deeper understandings of complexity for the development of more sophisticated explanations and more effective solutions” (Angus & Clark, 2012, p. 1). We intend that this research ultimately contributes to systemic improvements in the hospital discharge process for people experiencing homelessness. One avenue for exploration is re-imagining the role of hospitals in addressing the SDoH of homeless patients, while restructuring funding streams that can create barriers to this. Health systems must “reach beyond their walls” (Kangovi et al., 2018) and integrate social services into their processes to address the SDoH of people experiencing homelessness. We might take guidance from the United Kingdom, where the Homelessness Reduction Act 2017 mandates that hospitals refer homeless patients on to other services and not discharge them to the streets or emergency shelters (see http://www.legislation.gov.uk/ukpga/2017/13/section/6/enacted). In Canada, recent clinical guidelines for improving social and health outcomes for homeless and vulnerably housed patients recommend immediate linkages to permanent supportive housing as the highest priority for homeless healthcare (Pottie et al., 2020). How this will be implemented during hospital discharges, particularly with current systems pressures and resource challenges outlined in our study, is yet to be determined. Policies that address these structural factors would take the pressure off hospital staff to make decisions between discharging someone to inappropriate destinations or pushing back against managerial pressures.

Our study has limitations. We experienced challenges recruiting shelter workers and we may be missing important voices and experiences. Further research that includes voices of people experiencing homelessness who go through the discharge process would contribute to a more robust explanation of difficulties in the hospital discharge process. Our study also has notable strengths. To our knowledge, this is one of the first empirical studies in Canada to characterize the hospital discharge process for people experiencing homelessness when leaving general medicine. Furthermore, including both hospital and shelter workers provided a more rounded perspective of the complexity of this discharge pathway. Our use of critical realism and CMO configurations as an analytic tool allowed us to expand on and complicate current explanations of challenges in hospital discharge for homeless patients by identifying additional explanatory mechanisms. The discharge process is complex, messy and emergent, similar to the open-system characteristic of the social world. Critical realism offers a philosophical standpoint from which to embrace and explore complexity without requiring us to artificially simplify it. In critical realism, explanations are fallible and partial; our study presents one small piece of explanation. More research is needed to identify other possible explanations, informing the development and implementation of future interventions.

Contributions to knowledge

What does this study add to existing knowledge?

• This study provides a contextualized explanation for why many homeless patients are discharged to emergency shelters from general medicine. In doing so, it nuances current scholarship that can fall short of explaining why or how these issues persist.

• There is very little written on the discharge process for homeless patients after hospitalization for physical health conditions in Canada. This study identifies some of the drivers of current challenges in discharging homeless patients, helping policy makers, hospitals, and shelter workers identify clear areas for intervention.

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

• A human rights approach to housing must be employed to navigate improvements in the hospital discharge process. Without this, short-term solutions that do not address housing need may fail to trigger long-term changes.

• Hospitals should develop specific discharge protocols for people experiencing homelessness to guide and improve the discharge process.

• Multi-governmental investments are needed into increasing the availability of intermediate care facilities where homeless patients can recover post-hospitalization, and to improve post-discharge supports for the diverse needs of homeless patients (e.g., older adults experiencing homelessness).

• Hospitals and shelters should develop organizational-level relationships to support integrated and effective hospital discharges for homeless patients.

Author contributions

Jesse Jenkinson: Conceptualization, methodology, investigation, resources, data curation, formal analysis, writing (original draft), and visualization. Stephen W. Hwang: Conceptualization, methodology, and writing (review and editing). Carol Strike: Conceptualization, methodology, and writing (review and editing). Erica Di Ruggiero: Supervision, conceptualization, methodology, and writing (review and editing).

Funding

This work was supported in part by the Social Sciences and Humanities Research Council of Canada [752-2017-1416], and the University of Toronto Open Fellowship. The funders had no role in the design of the study and collection, analysis, and interpretation of data, or in writing the manuscript.

Data availability

Not available due to participant privacy.

Declarations

Ethics approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the University of Toronto, 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

All authors have read and approved the final version of the manuscript.

Conflict of interest

The authors declare no competing interests.