Exploring the Impact of a Housing Support Service on Hospital Discharge: A Mixed-Methods Process Evaluation in Two UK Hospital Trusts

.


Introduction
Delayed hospital discharge is a global health care problem and policy concern [1,2]. Delayed discharge is defned as a patient remaining in hospital even when they are medically ft due to issues such as a lack of social care, needing adaptions to their home, awaiting residential home availability, or being homeless [3]. Several countries including Spain, France, the US, and Canada have identifed issues with delayed discharge (e.g., [4][5][6][7]). For example, there were 155,782 hospital bed days attributed to delayed discharge in February 2020 within England, UK [8]. Poor information exchange and coordination of support between care providers as well as inadequate transfers of care are key causes of delayed discharges [9,10]. Delayed discharge has many negative efects on patient outcomes and the healthcare system. It is problematic because it incurs signifcant healthcare costs, is detrimental to patient safety, and increases the likelihood of hospital readmissions [3,[10][11][12][13].
A key reason for delayed discharge is patients not having suitable housing to be discharged to that meets their needs [14]. Tis includes older people with mobility issues who cannot live independently, people with enduring mental health illness who need supported living, and those experiencing homelessness [14][15][16]. Te latter cohort is signifcant due to increasing numbers and complexity of needs [15]. In one UK mental health trust, homelessness was independently associated with a 45% increase in length of stay, and the need for housing was the most common cause for delayed discharge from hospital for homeless individuals [17]. Leaving hospital is often a traumatic experience for people without a fxed address, and questionnaires suggest that 30%-70% of homeless inpatients are "discharged to the street" (i.e., sleeping rough immediately after discharge) [18,19] greatly increasing the likelihood of hospital readmission [20]. Compared to housed patients, those experiencing homelessness have high rates of emergency readmissions for nonmedical reasons. Tis shows the importance of adequate discharge arrangements which take into consideration wider social and economic needs (such as housing status) [20]. Recognising the impact of homelessness on hospital systems and patient outcomes in the UK, an additional pledge of further investment in homeless prevention (e.g., such as pathway teams based in hospitals), was included in the National Homelessness Reduction Act in 2017 [21].
Many initiatives developed to address housing-related barriers to hospital discharge have involved the integration of health, social care, and housing services to identify innovative solutions [22][23][24], such as funding housing support-related roles and developing hospital transfer services [25]. Examples of initiatives in the UK include housing staf providing a triage service to support people who need extra support to return home [3] and the introduction of a housing to home support worker to provide a single point of contact for housing-related discharge issues [26]. An independent evaluation of the triage service demonstrated clear benefts to hospital discharge and healthcare staf [3], whilst a small audit revealed patients in the housing to home service to be positive about the initiative, although there is little evidence on the impact of the intervention on hospital discharge processes [26]. Indeed, whilst there is now a plethora of international and UK evidence on the impact of hospital discharge, there remains a dearth of evidence into the efectiveness of nonmedical interventions such as housing support to improve the discharge process. Te aim of this research was to bridge this gap in evidence through an evaluation of a newly integrated housing and health service which aims to facilitate hospital discharge in Wakefeld, UK.
To facilitate more timely discharge from hospital, housing support coordinators (HSCs) were introduced in two hospitals in the UK to provide specialist support to hospital inpatients who were identifed as having housing issues. Te scheme was a joint venture between Wakefeld District Housing (WDH) (social housing provider [27]) and two NHS hospital trusts: a mental health hospital (Southwest Yorkshire Partnership Trust (SWYPT)) and an acute hospital (Mid Yorkshire NHS Hospital Trust). Te HSCs were experienced housing ofcers previously working in a variety of tenancy support roles prior to taking this role but did not require specifc professional qualifcations. Similar to social workers, HSCs would provide support for the wider determinants of tenancy sustainability such as health and fnances through signposting to onward provision, as well as providing one point of contact for specialist housing support. WDH is one of the UK's largest social housing providers, managing 32,000 homes across Wakefeld and the North of England. As well as managing homes, WDH provides a range of in house support including housing, tenancy, and wellbeing services to ensure wrap-around support for individuals and increase tenancy sustainability. Clinicians would refer hospital inpatients with identifed housing needs to the HSC. People did not have to be WDH tenants to access the service. Te HSC would hold several meetings with the service user to understand their housing issues and identify solutions. Much of the support took place outside of the formal meetings, with HSCs undertaking case work including locating appropriate housing, making referrals to other services, completing assessment forms on behalf of the service users, and attending multidisciplinary discharge planning meetings.
To inform service development and commissioning decisions, the University of Shefeld (UoS) undertook an evaluation to understand whether and through what mechanisms, and to what extent, the HSC service has an impact on the hospital discharge pathway and service user's health and housing outcomes.

Methods
Here we report fndings from a mixed-methods process evaluation to explore the delivery and impact of the service using both routinely collected data and a qualitative interview study. A mixed-methods study design was conducted to provide a rich insight into the delivery of the intervention and capture its complexity [28]. An economic evaluation was also conducted but encountered challenges due to a lack of available data to quantify hospital costs and therefore has not been included within this paper (see separate economics report: [29]). In addition, a questionnaire element was included in the evaluation, but the sample size was too small to undertake meaningful analysis (see paper by Foster et al. [30] and limitations section for further refection). Further details on the methods and the wider report fndings are included in the full project report [31]. In this paper, we focus on the evaluation fndings which are relevant to wider policy and practice.

Routinely Collected Data Analysis.
We undertook secondary data analysis of service user data collected by the housing association between April 2018 and June 2021 (n = 488). For each service user supported, the HSCs recorded demographics, reasons for referral, the nature of support provided, and outcome of support on the housing association's data management system. Te data were anonymised before being shared with the research team. Researchers cleaned the data and transferred it into SPSS Version 28 for analysis. Researchers undertook descriptive analysis to explore aspects of service delivery such as demographics and outcomes [32]. Further descriptive analysis was not undertaken because the focus of the process evaluation was to understand who was accessing the service, the nature of delivery, and the associated outcomes. Tese issues were explored through the descriptive analysis.

Qualitative Data Collection.
Due to onward funding of the intervention not being confrmed in the acute hospital at the time of the qualitative data collection, interviews with service users and health care professionals (HCPs) were conducted within the mental health trust only. In addition, we interviewed the HSCs in both hospital sites, alongside key housing staf in WDH with knowledge of the intervention.

Sample and Recruitment
(1) Service Users. Service users were recruited via the housing association. A tick box for people who were willing to be contacted for the interview aspect of the study was included on the housing referral form. If a service user ticked that box during the recruitment period (September 2020-August 2021) (N = 103), a member of housing staf approached them to seek permission to pass on contact details to the UoS research team to discuss the interview aspect of the study. If they agreed, housing staf telephoned the research team to pass on the contact details. Te researchers then contacted the service user to provide information about the study and if agreeable, to arrange an interview. A total of 103 service users ticked the box during the recruitment period of which 32 had incorrect phone numbers. 23 agreed for WDH to pass on contact details to the UoS research team. Te UoS research team contacted all 23 for which for 16 there was no response or they did not want to take part. Seven people agreed to an interview in principle; however, one person did not respond when trying to confrm a date for the interview, and another did not pick up the phone on the day of the interview. Service users received a £10 voucher for participating in an interview.
(2) Healthcare and Housing Professionals. To understand how the introduction of the HSC role has had an impact on the wider health system, we invited healthcare professionals (HCPs) who had experience or knowledge of the intervention and relevant staf from the housing association to take part in the interview. Te study was introduced to HCPs through an email or discussion with the HSC, and those who were interested in taking part were referred onto the researchers to make the arrangements for the interview. Similarly, key housing staf were invited to take part in an interview via email by the research team.

Data Collection
Methods. Semistructured interviews were undertaken between October 2020 and March 2021 by EH and EL using separate topic guides for housing staf, HCPs, and service users developed around the aims of the evaluation. All interviews took place over the phone or via video call using Microsoft Teams software due to COVID-19 restrictions and lasted between 27 and 136 minutes (mean: 52 minutes). With participants' consent, interviews were digitally recorded, transcribed verbatim, and loaded into NVivo 11 software for data management and coding. Verbal consent was taken at the start of each interview.

Analysis.
Framework analysis was undertaken which involved (1) familiarization with the data, (2) identifying a framework, (3) indexing, (4) charting, and (5) mapping and interpretation [33]. An initial thematic framework exploring issues relating to context, implementation and delivery, and the outcomes and impact of the intervention was derived from discussions with our housing and healthcare partners. Initial codes from these discussions were combined with codes derived from the in-depth reading of a small number of transcripts before being modifed to refect the emerging themes. Te development of the coding framework was largely inductive, but initial scafolding was provided through discussions with housing partners. Te research team met regularly to validate the coding framework and discuss emerging fndings [34].
(1) Integration of Findings. We undertook a "following the thread" method to integrate the fndings [35]. Tis involved regular meetings between the researchers and partners to discuss the arising fndings and develop overall themes. Tis process was iterative, for example, when we identifed within the routine data that the support delivered was relatively short term, fndings from the qualitative interviews demonstrated how this was often due to people not being referred in a timely manner.

Ethics.
Ethical approval for this project was granted by the North of Scotland Research Ethics Committee and the Health Research Authority (REC reference: 20/NS/0050) (4th May 2020).

Participant Characteristics of the Qualitative Interviews
(1) Healthcare and Housing Professionals. We conducted 11 interviews with housing, health, and social care professionals. People held a variety of roles including managerial responsibilities and directly delivering care (Table 1).
(2) Service Users. Five service user interviews were conducted. We interviewed fewer service users than planned due to COVID-19 restrictions and the vulnerable nature of the population (see limitations section). Service users varied in their age (range of 24-66 years, mean: 28 years) and living situation prior to hospital admission ( Table 2). All participants identifed as White British ethnicity and were either single or divorced with no dependent children. Te demographics collected through the routinely collected data are described in the next section alongside who accessed the service.

Who Accessed the Service and Reasons for Access.
Te service supported 488 service users between April 2018 and June 2021: 238 from the mental health trust (service started in April 2018) and 250 from the general hospital trust (service began in September 2018) (demographics described in Table 3). Most service users were male (n � 325, 66.6%). However, the age profles of the two services varied, with the HSC in the general trust supporting an older population (over half of service users were over 65, n � 132, 52.8%), whereas in the mental health trust, half of service users were under 45 years old (n � 142, 59.6%). Te majority of service users were White British (n � 451, 92.5%), refecting the local geographical area.
It was apparent that service users led complex lives and often experienced several social issues which interplayed with their housing needs. Tis included substance misuse, debt, and safeguarding issues such as abuse at home. Furthermore, many mental health service users had experienced multiple hospital admissions, which destabilised their housing.
Tere were diferences in the housing issues experienced between the mental health and general hospital service users (Table 4). Over a third of service users in the mental health trust experienced homelessness (n = 90, 38.6%), and there were several people unable to return home due to issues like violent behaviour or damaging previous property (n = 39, 16.7%). In contrast, service users from the acute hospital often faced physical barriers to their property because of their medical conditions such as having strokes, being at risk of falls, and amputations. For example, some service users no longer felt their previous accommodation was accessible (n = 33, 13.3%) or people needed their living accommodation to be on one foor (n = 28, 11.3%). Approximately a ffth of service users were experiencing homelessness (n = 46, 18.5%) (half the proportion of the mental health hospital).
Te variability of the issues that necessitated HSC support highlights the specialist and wide-ranging knowledge that HSCs require, in addition to demonstrating how this may difer between settings. HSCs working in an acute trust will need knowledge of organising adaptions and supporting people to access care homes, whereas in mental health trusts, they need skills in supporting people with homelessness and working with people with histories of substance misuse.

Support Provided.
HSCs provided a variety of support including assistance with bidding on properties (n � 133/ 351, 37.9%), home search applications (n � 199/351, 56.7%), and applications for Health and Medical Rehousing (n � 158/351, 45%). Other support included help with addressing rent arrears, organising a house clean, arranging an occupational therapist visit, and supporting people to stay within their current accommodation. Te range and complexity of support demonstrates how the HSCs were skilled in providing personalised care tailored to the individual needs of each service user.   4 Health & Social Care in the Community Te intervention was generally short term: over half of service users were supported for less than a month (n � 296/ 479, 61.9%). Only 10% of service users received support from HSCs for more than three months (n � 48/479, 10%). HSCs found it benefcial if people were referred to the service early on in their admission, to give the HSCs sufcient time to address issues. Tis was a particular issue for HSC3 who often received referrals late on in the discharge process partly due to supporting a more elderly population who may need support in respite care before returning home. Timely referrals were particularly important given the short-term nature of some hospital stays.
"I think I've had a couple of amputee's probably being a delayed discharge, and again I think I would put that down to the fact that they've not referred them soon enough to me sometimes. . .so I might get a referral when they're ready for discharge and suddenly you've got to sort something out" (HSC3) 3.4. Impact of the Service 3.4.1. Impact on Service Users. Generally, service users felt receiving the support was benefcial in terms of improving their housing outcome and reducing housing-related stress. People discussed receiving support to organise repairs, move accommodation, access fnancial support for housing, and referral to additional support services. For example, SU5 was put on higher priority for properties following the HSC intervention: Service users valued the support provided by the HSCs to help them address fnancial and wellbeing issues that were contributing to their housing problems, for example, supporting people with completing applications for welfare benefts.
"I wouldn't have been able to do it myself at that time. So, she just basically said right, we'll fgure out your fnances and do your fnancial forms, we'll fgure out, cos they needed that, to fgure out the housing side and put me in where you need to be because I was on a low priority, then they, they'd put me on a higher one. So, she did that for me as well and then flled all the forms out erm, to do it anyway and now I can do it online, so I wouldn't have been able to fgure that out myself." (SU5)

Impact of the Service on the Hospital Discharge
Pathway. Many HCPs discussed how the introduction of the role appeared to have improved the overall hospital discharge system, reducing unnecessarily delayed discharges and average length of stay. In particular, the service had assisted in discharging service users who had received effective treatment and were feeling stable mentally, who otherwise would have had to remain in hospital for a longer period due to housing-related issues. It was acknowledged that having a particular person with specialist knowledge of housing within the multidisciplinary team was essential to ensure service users wider needs were met.   Te service was seen as particularly benefcial for service users who were experiencing homelessness. Before the HSC was in place, people experiencing homelessness were left to access homelessness services independently, which was a barrier to engagement (see Section 3.5 for further information).

Impact on Healthcare Services and Individual Job Roles.
HCPs at all levels commended the service for alleviating pressure on overstretched healthcare services and reducing staf workloads, freeing up staf time to focus on front line clinical work. Having a dedicated person with specialist knowledge of housing-related issues was considered "vital" (HCP4) for improving overall hospital services, in addition to reducing pressure and stress on individual clinical staf, and the hospital social work team. Before the job role was in place, the responsibility for housing-related issues would fall to front line clinical staf who did not have the knowledge of housing systems, or the time to appropriately deal with the issue.
"[before the service was in place] we would hope that the care workers were involved in it, but it would just be us ringing to talk to somebody and then they didn't know the answers so we'd get passed to somebody else and then we'd still not have the answer at the end of that day so we might have to follow it up. So it's lightened the workload for the nursing staf dramatically." (HCP7) 3.5. Strengths of the Service. Key mechanisms which appeared to contribute to the success of the service were identifed during the evaluation. Te HSCs' previous experience and knowledge of the housing service were at the forefront of successful implementation of the role. Both HSCs had come from a previous customer facing housing role and were skilled at developing rapport with service users and external organisations. Te wide-ranging support provided by the HSCs shows how they delivered personcentred care. People valued that the HSCs proactively liaised with both service users and members of the care team. In this sense, each HSC acted as a coordinator facilitating communication between diferent partners-service users, health and social work professionals, and family members. Tis reduced tension between diferent professional groups and ensured that service user's housing situation was addressed at the earliest opportunity.
"What they do is have a bit better knowledge than I do of housing processes and you know how to apply for things. And they are quite good for the service users to contact directly because then they feel that they're getting a sort of expert answer and they can sort of keep track of what's happening. .

.they're quite good at contacting me and letting me know what I need to do and say "you need to do this form" and what have you." (HCP2)
Coordinating communication between diferent agencies and delivering patient-centred care were particularly pertinent for supporting vulnerable service users such as those experiencing homelessness. Te HSC acted as an important liaison between the service user, the council homelessness team, and the housing association, remaining engaged with the service user as they moved through the process. Before the HSC was in place, the hospital or community teams would make a referral to homeless services with no further follow-up, increasing the likelihood of ongoing homelessness or hospital readmission. Having a dedicated housing specialist involved with the homeless individual from the start meant that they provided consistency of support through the discharge pathway, and this continued with referral to more specialist support if required.
Supporting people to access other appropriate services was a key component of the service. Over three quarters of service users were referred for further support (n � 376, 77.7%). Te referrals were to a range of organisations including social care, debt advice, and befriending support. Te housing association delivered many of these services in-house which may be benefcial for service users because it prevents them from being passed onto multiple agencies.
3.6. Challenges to Service Delivery. We identifed several challenges to service delivery including integration of the HSC with hospital teams, the availability of suitable housing, the complexity of service users' circumstances, and the need for follow-on support.

Integration of the Role into Hospital Discharge Teams.
A key challenge was how each HSC was integrated diferently into the existing hospital systems. In one hospital, the HSC was based with the hospital discharge team with access to the hospital electronic patient notes system and regularly attended multidisciplinary team meetings and ward rounds with clinical staf. HSC2 refected that this helped in raising awareness of the service and making them feel part of the clinical team, a view which was also shared by healthcare professionals: "the [HSC] calls into our meetings, well she's part of our team, the patient fow team so she attends sort of daily meetings with us, so we discuss any issues that we've got. . ." (HCP5) In contrast, the other HSC was based in the social work team and did not have the same access to hospital systems, nor did they feel as well incorporated in the clinical team. Te HSC felt this lack of integration resulted in less awareness of the role across the trust:

"the health professionals I probably don't think the health professionals even know I'm involved sometimes though, I think sometimes you know, doctors wouldn't really know much about my role I don't believe I think it's the discharge team 'cause, you know the discharge coordinator sits on every ward they are sort of ultimately responsible for the discharge" (HSC3)
It was suggested that placement within the discharge team with access to electronic patient management systems and line management from healthcare staf would support better integration with the ward teams.
"I think the model feld is slightly diferent to [

Availability of Suitable
Housing. Te ability to meet people's housing needs and speed up hospital discharge was constrained by the lack of suitable housing, especially social, adapted, or supported housing for people with mental health issues or physical support needs. Often, whilst HSCs could help facilitate a more efective discharge and support in the community, they could not necessarily speed up how quickly someone was discharged because of constraints in the wider system, e.g., delays from home adaptations. HSCs, and service users themselves, sometimes felt that the lack of housing provision meant people were discharged into unsuitable accommodation, which may increase the likelihood of readmission. For example, a service user discharged to shared housing subsequently declared themselves homeless: "I had a place down [name of area] that I thought was going to be a private place. But it didn't turn out. It turned out to be shared living so I ended up coming back and I ended up homeless" (SU3) 3.6.3. Complexity of Peoples' Lives. Whilst the HSCs were highly skilled and experienced in supporting service users with complex needs, they encountered challenges with supporting people due to problems within the wider system. One key issue was that people's situations or prior experiences could prevent them from accessing housing options. For example, service users who had previous rent arrears or issues with crime were disqualifed from accessing some social housing. Tose who were homeless often lacked the correct ID and documentation to register and apply for properties. HSCs believed it was crucial to intervene as early as possible for individuals experiencing homelessness, allowing them ample time to ofer the necessary support and avoid discharge into temporary housing options such as hostels. In turn, some follow-up should continue beyond discharge to avoid potential readmissions: "I think homelessness is one of the main barriers for someone being discharged, so I think of the pressures for that cohort is actually securing them some permanent accommodation but it is not always possible at the point of discharge though. I think again it's around just ensuring that that support in some way continues when they leave hospital. And making sure that support's there and it doesn't kind of break down quite quickly. So you know I think the main challenges I would say was from both sides really that for certain clients the challenge has been the supply of accommodation. Getting the right accommodation available for them." (HSC4)

Some Service Users Needing Support after Discharge.
Te HSC service generally stopped when people were discharged from hospital. However, some service users expressed a desire for HSC support after discharge, particularly those experiencing homelessness or more complex mental health needs. HSCs and housing management suggested that having ongoing support may prevent readmissions, especially as there was not an equivalent service for people in the community.
"I think maybe after you get discharged I think people should follow-up whereas they normally discharge you after you've been discharged it's weird and I think you need more support while you're in the community you know why would you need support while you're in [name of hospital

Discussion
Te aim of this study was to understand the impact of the HSC service. Several key fndings emerged relating to the strengths, challenges, and impact of the HSC service which have wide relevance to other housing and health organisations wishing to implement similar integrated services.
A key strength of the service is that it is delivered by experienced housing ofcers who have a liaison role, which links together diferent services. Additionally, being employed by a housing association meant that the ofcers were able to refer people into follow-on services. Tis refects the current policy preference for integrating health, social care, and voluntary sectors in delivering patient-centred care [36,37]. Tere is strong existing evidence that integrating roles which span the boundaries of health and housing such as the HSC reduces demand on overstretched healthcare services and improves population health, particularly for underserved populations [38]. Tis is important because a lack of coordination between services and an inadequate transfer of care have been found to be key causes of delayed discharge [10].
However, our study identifed several challenges relating to the integration of the role within hospital settings. Te HSCs did not always have access to the necessary IT systems nor felt part of the service users' multidisciplinary discharge team. Previous research on the challenges of integrated housing/health interventions has also demonstrated the difculty in accessing IT systems and information sharing between sectors [39]. Our fndings also highlight the importance of managers within the hospital setting taking responsibility for integrating the new HSC role within existing clinical teams and IT systems.
A notable beneft of the service was that it signifcantly reduced stress and the time spent on housing problems, a crucial fnding given current policy concerns over the burden faced by healthcare staf [40]. Providing more specialist services such as the HSC role may reduce some of this stress by freeing up time clinical staf spend on social issues which they are often not equipped to address. Such fndings echo previous research regarding the introduction of social prescribing link workers into primary care [41], which demonstrated that the role reduced stress on GP workloads, in addition to improving outcomes for patients.
Another key beneft of the service was how the service supported complex cases such as those service users experiencing homelessness. As the introduction to this article describes, issues relating to hospital discharge and homelessness have a great impact on hospitals. Te complexity of homeless cases in this service required specialist knowledge of housing systems and a fexible, patient-centred approach tailored to the circumstances of each individual, which overburdened healthcare staf were unable to provide. In turn, it provided consistency of support throughout the discharge process, which has been shown to be a key component of safe and efective hospital discharge [16].
Commissioners funded the HSC service with the expectation that it would facilitate the speed of hospital discharge, but this does not consider constraints within the wider system such as a lack of appropriate housing. Although our research demonstrates positive impacts on hospital discharge from the perspectives of HCPs, it remains the case that even with HSC support, service users cannot be discharged quicker from hospital if there is no suitable housing or provision of care available to meet their needs. A shortage of afordable and appropriate housing options is a longstanding issue that afects other Western countries such as Canada [42]. As well as increasing the likelihood of homelessness [43], a diminishing housing stock increases pressure on housing associations and the level of support they can ofer [44]. Given this, commissioners need to consider funding the HSC service alongside increased investment in wider housing options and other services across the healthcare system.
In addition to managing the expectations of commissioners, an additional challenge is managing service user expectations. Tere were times when service users could not be provided with the optimum housing solution because of a lack of suitable housing stock, such as the need for more step-down supported housing options for mental health service users.

Recommendations for Similar Services.
Our fndings point to some recommendations for policy and practice and for those who are interested in commissioning and implementing similar services. Some recommendations for the specifc service, such as the need to introduce follow-on support, are detailed in our full project report and have been taken forward by the service [30]. Here we outline recommendations relevant to wider policy and practice.
As this study has demonstrated, any service intending to develop a HSC role needs to be aware of the challenges associated with successful implementation of the role and give consideration to where the role would be situated (e.g., health or social care team), the range of skills required, how the HSC role may need to be adapted depending on the patient group or trust in which it is situated, and how the complex background of some service users may require an extended period of support and impact on anticipated outcomes.
Our study has shown the vital importance of integrating the HSC within appropriate existing hospital team such as those with responsibilities for discharge planning, with access to healthcare IT systems. HSCs need appropriate managerial support, from healthcare staf within a patient fow or discharge team who understand the role, and can help promote it within the organisation to strengthen referral pathways and promote buy in.
Te background and experience of the HSCs were instrumental in the successful implementation of the role which demanded a range of skills including having experience of providing housing-related support, the ability to deliver person-centred care to those with complex needs such as homeless patients, a high level of profciency in establishing relationships, and the ability to undertake service development without the support of a wider housing team within the hospital. Recognising the level of skill and experience required for the role, those looking to commission such services need to provide a pay scale which is appropriate to attract senior-level housing ofcers with the necessary skillset.

Knowledge
Mobilisation and Impact. Given current issues around hospital discharge, the results we present here are timely and signifcant.
Our fndings are likely to be relevant to other hospitals in diferent geographical locations facing similar challenges. Globally, there is a lack of specifc discharge policies or processes tailored to the homeless population [45]. In Canada, the level of homelessness has been described as a national disaster, yet there is a lack of discharge planning and policies that cater to homeless individuals [43]. Echoing our study, Weldrick et al. [46] found the need for individualised support, rapport building, and partnership working across agencies to be instrumental in supporting homeless patients. Tis indicates that the introduction of a specifc role to provide specialist housing support and provide a liaison function between sectors would be benefcial in other Western settings.
A recent scoping review [2] identifed several interventions aimed at improving hospital discharge, yet only a small number of the included interventions were nonmedical or housing-related specifc support roles (for example, [26]). Two UK and US studies discussed the impact of introducing discharge criteria to nursing roles and found a notable reduction in delayed discharges [47,48]. Although the evidence on housing specifc roles is scarce, fndings echo the current study in that consistency of support has a positive impact on discharge planning. However, our study found that front line hospital staf are already overstretched and found incorporating housing-related issues into their remit stressful. Rather than adding further burden on nursing staf, our fndings indicate that introducing a specifc role to provide specialist housing support would be benefcial for vulnerable patients such as the homeless. It is clear that learning from this evaluation, such as the need for skilled housing ofcers providing specifc patientcentred, housing support, can be incorporated into other national and international settings experiencing similar challenges.

Strengths and Weaknesses.
A key strength of the study is that it produced learning that has informed development of the service and will be of use to people developing other health-based housing support services. Te fndings are timely given the current impetus to integrate housing and health services to reduce pressure across the healthcare system. As a result, there has been great interest in the study results by practitioners at several conferences (e.g., [49]).
However, Te study encountered several limitations, predominantly because of COVID-19 restrictions. A key challenge was recruitment of service users to both aspects of the study. Researchers planned to undertake a pre-and postservice questionnaire, but due to delays relating to COVID-19 such as increased staf sickness and reduced capacity of the research teams to conduct the study, as well as COVID-19-related research (such as vaccine trials) taking priority within the trusts, the sample was too small to undertake meaningful analysis [30]. It was also difcult to recruit service users to interview due to COVID-19 which restricted access for research staf visiting the wards to conduct consent procedures and the vulnerable, transient nature of the population. Tis resulted in only fve service user interviews in the mental health trust; therefore, data saturation is unlikely to have been achieved, although similar themes and issues were emerging across the transcripts. Still further, given the small numbers and specifc population, the fndings may not be wholly transferable to other hospital settings. Given the vulnerable nature of the population, it is likely that those with more capacity agreed to be interviewed. In turn, the perception that the HSC role saved clinical staf time by removing housing support from their workload was a key fnding of the qualitative data; however, it was difcult to quantify. Terefore, future studies may want to consider implementing ways of quantifying impact on hospital staf and their workloads through a formal economic evaluation.
Future studies using quantitative methods such as randomised controlled trials are needed to further quantify the impacts of interventions in terms of length of hospital stay, rehospitalisation, and mental health. In turn, future studies need to be aware of the challenges of recruiting service users and explore how to capture relevant system outcomes. Tere is a need to improve the recording of issues such as additional bed days because of delayed discharge alongside developing ways to access this information to enable better quantifcation of costs.

Conclusion
Our study identifed that delivering specialist housing support to hospital inpatients, both those with acute medical and mental health service needs, has multilevel benefts for individual service users and clinical staf. Key to the success of the intervention was experienced housing staf with links to both hospital and housing systems delivering patientcentered support. However, the impact of the service was reliant on the availability of suitable housing stock and other support services. Given the positive benefts of the service, our fndings indicate that it would be benefcial to implement the HSC intervention in other hospitals facing delays in discharge due to housing-related issues.

Data Availability
Te data used to support the fndings of this study are available on reasonable request from the corresponding author. Te data are not publicly available due to privacy and ethical reasons.

Additional Points
What Is Known about the Topic? (i) Timely discharge from hospital improves health outcomes and reduces hospital costs. (ii) Housing-related issues contribute to a large proportion of delayed hospital discharges. (iii) Several housing-related interventions have been developed to improve hospital discharge, but little is known about their impact. What Does Tis Paper Add? (i) Healthcare professionals highly valued having a specialist housing support coordinator supporting inpatients because it reduced the time they spent on housing-related issues, enabling them to focus on clinical tasks. (ii) Key to success was having housing support coordinators with highly developed skills in providing patient-centred support, developing efective relationships with healthcare staf, and being placed within a multidisciplinary team with oversight and management from the health service. (iii) Key challenges were service users needing long-term support and a lack of suitable housing options such as supported housing for mental health service users.

Disclosure
Te views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.

Conflicts of Interest
Te authors declare that they have no conficts of interest.