Work package 1
We adapted an existing quality assessment tool for longer-term mental health facilities (the QuIRC) and an existing patient-reported outcome measure (the CAT) for use in supported accommodation services. This process comprised (1) review of the content by focus groups of staff and service users of the three types of supported accommodation services in England and three expert panels; (2) agreement of suggested amendments by the PMG; (3) assessment of the psychometric properties (item response variance, internal consistency and inter-rater reliability) of the adapted QuIRC; and (4) assessment of the psychometric properties (internal consistency and convergent validity) of the adapted CAT.
Amendments to the wording of 28 QuIRC items were made, 20 items were deleted and 10 new items were added. Only five adapted QuIRC items were found to be unreliable but further amendments to four of these were possible and only one item was dropped completely. The psychometric properties of the adapted QuIRC were acceptable. An online application has been developed to provide web-based access for supported accommodation service managers to use the tool, the ‘QuIRC-SA’, to assess the quality of their services directly.
Minor amendments to the wording of all seven CAT items were made. The adapted version, the CAT-SA, demonstrated good internal consistency and satisfactory convergent validity.
Both measures can be recommended for routine use in mental health supported accommodation services.
Work package 2i
A sample of 87 supported accommodation services were recruited from 14 nationally representative areas of England (residential care, n = 22; supported housing, n = 35; floating outreach, n = 30). Interviews with each service manager and 619 randomly selected service users and their support staff were carried out to collect descriptive and quantitative data. Around two-thirds of service users across all three types of accommodation were male and the majority were single and unemployed. Around one-third of floating outreach service users had a primary diagnosis of depression or anxiety whereas the majority of those in residential care and supported housing had a primary diagnosis of psychosis. This suggests a slightly more varied diagnostic profile than was identified by Priebe et al.’s11 telephone survey. We also identified greater variability in support provided between service types than Priebe et al.11
Staff rated those in residential care as having the highest needs and a greater requirement for assistance with personal care than the other two types of supported accommodation; across all three types of service, most needs were met. The three types of service had similar referral and assessment processes and most service users were subject to the Care Programme Approach and, therefore, received additional input from a local statutory community mental health service. Around one-fifth of participants had a recent history of significant self-harm, but this was much higher among those in supported housing and floating outreach than among those in residential care. Around one-third were vulnerable to exploitation and over half were at risk of self-neglect, with the highest proportion among those in residential care (almost three-quarters). Users of supported housing and floating outreach services were more likely to have been a victim of crime in the last year than users of residential care.
Supported housing and floating outreach services worked to the same time frame, with an expectation that service users would move on to less supported accommodation or manage with less support within 2 years of being taken on. In regression analyses that accounted for differences in service user characteristics, service users’ quality of life did not differ between supported housing and residential care but was lower for those in floating outreach than for those in residential care. Supported housing was associated with higher autonomy than residential care but floating outreach was not.
Floating outreach services provided more places per service than the other two types of supported accommodation and costs of care were lowest (and highest for residential care). However, with regard to cost-effectiveness, there was a reasonable probability that supported housing provided better value for money with regard to quality of life than the other two types of service. Supported housing also produced better outcomes than residential care with regard to autonomy and was less expensive.
Work package 2ii
We followed 586 out of the 619 participants recruited in WP2i over 30 months, representing a 95% follow-up rate, to assess the proportion who successfully moved on to more independence and identify service and service user factors associated with this. We also planned to collect more comprehensive follow-up data on a subcohort of those who had been in their supported accommodation for < 9 months at recruitment, to compare quality of life, autonomy and social function between service types. Unfortunately we were only able unable to recruit half (73/147) of this group, which precluded statistical comparisons.
Two-thirds of those in floating outreach achieved the primary outcome of managing with less support at the 30-month follow-up. One-third of those in supported housing and 1 in 10 of those in residential care successfully moved on to less supported accommodation. Successful move-on was positively associated with service quality, specifically the degree to which the service promoted human rights and recovery-based practice (as assessed by the QuIRC-SA). Service users with more unmet needs, more risk and longer length of stay in the service (all of which are markers of greater morbidity) were associated with lower odds of successful move-on. We also identified a paradoxical finding that higher scores on the social interface QuIRC-SA domain at recruitment were associated with lower odds of successful move-on. This domain reflects the degree to which family members are involved in service users’ care and the degree to which the service engages service users with local community resources. It is possible that services that achieve greater family engagement may find greater resistance from family members for service users to move to more independent accommodation, an issue identified in previous studies.50 In addition, it may be that service users who were more integrated into local community activities may have been reluctant to move to alternative accommodation in a different locality.
Work package 3
In our qualitative investigation, we found considerable agreement between staff and service user views on the goals and purposes of supported accommodation services. Participants’ narratives consistently returned to themes of building individuals’ skills and confidence for more independence incrementally and avoiding inadvertently creating dependency on services. Staff understanding of the facilitators of effective practice included many elements that can be conceptualised within rehabilitative and recovery-orientated practice. However, the time-limited nature of services provoked anxiety among service users and staff about how the person would manage when they moved to a less intensive service. Staff and service user participants agreed that a trial comparing the effectiveness of two models of supported accommodation, supported housing and floating outreach, would be valuable.
Work package 4
We conducted a feasibility trial to assess the viability, required sample size and potential outcomes for a randomised evaluation of two supported accommodation models: supported housing and floating outreach. This component of the programme was conducted in three sites: North London, East London and Gloucestershire. We aimed to recruit at least 60 service users referred to supported accommodation services and randomise them to receive either supported housing or floating outreach.
We collaborated with local supported accommodation referral co-ordinators, inpatient ward managers and care co-ordinators to identify eligible service users. We used the findings from WP3 in our recruitment approach. When we identified a potential participant, we contacted the referring clinician to explain the rationale for the study and enquire whether or not it was clinically appropriate for the service user to be considered for recruitment (i.e. whether or not they could potentially manage in either supported housing or an independent tenancy with floating outreach support). If the clinician agreed, we approached the service user to explain the purpose of the study and gain their informed consent for participation. Participants were randomised on an equal basis to move to supported housing or an independent tenancy with floating outreach support and the outcome of randomisation was communicated to the local referrals co-ordinator to process accordingly. If the individual declined to be randomised, they were given the opportunity to participate in the naturalistic arm of the study, whereby we followed them over the course of 12 months but had no influence on their supported accommodation allocation. For all participants, randomised or naturalistic, we conducted baseline interviews and 6-month and 12-month follow-up interviews with both the individual and a member of staff involved in their support.
Recruitment was challenging. We screened 1432 potential participants, but only consented 17 to participate: eight service users were randomised and nine participated in the naturalistic arm. The majority of potential participants were not approached as they were deemed to be clinically inappropriate for the study by their clinical team (n = 851).
We also conducted qualitative interviews with staff and service users involved to gain a fuller understanding of their views regarding the feasibility of a larger trial. We interviewed 11 service users (six who were randomised and five from the naturalistic arm) and 10 staff members (six who had referred service users to the trial and four who had refused) and analysed our data using thematic content analysis.
Four main themes emerged from the service user and staff interviews: preference for a certain type of supported accommodation, rejection of randomisation, complexity of randomisation and the value of a trial. Service users voiced concern about their housing situation being decided at random, mainly because they had a preference for either supported housing or floating outreach, or because they perceived differences between the two models. In spite of this, they were in support of a larger trial, citing the value of further research and the potential benefits for others. Staff often held a strong belief in the need for the existing ‘step-down’ model, revealing a lack of perceived equipoise between supported housing and floating outreach services. They also cited a lack of availability of supported accommodation and housing as a potential recruitment barrier. Some clinicians felt that it was unethical to use randomisation in relation to a person’s accommodation but most felt that a larger trial would be valuable.
Successes and challenges
The QuEST project achieved all four of its original objectives, on time and within budget. We delivered the first specialist quality assessment tool for supported accommodation services, which is now also available as a free online self-assessment tool. We also delivered the first specialist ‘patient-reported outcome measure’ for these services, which is also freely available for use. These tools have the potential to help drive up the quality of existing services through benchmarking and monitoring of quality improvement initiatives at the local or national level. They can also be used in future research studies in this field.
We completed the first, detailed national survey of the provision, quality and effectiveness of mental health supported accommodation services in England. Our results provide important data to assist our understanding of the similarities and differences between supported accommodation service types and the users of these services. Our cohort study was characterised by low attrition and identified service characteristics that contribute to successful move-on for service users that can help to focus future skills training for the staff working in these settings. It has also helped to clarify the characteristics of service users who require longer-term support.
However, a number of challenges were encountered. Most striking was the low recruitment rate in WP4. Despite adapting our recruitment strategy, we failed to recruit an adequate sample to justify investment in a full-scale trial. Nevertheless, this is an important and useful finding that helps to inform the direction of future research in this area. Similarly, our attempts to follow-up in person a subcohort of participants in WP2ii who had been in their supported accommodation service for < 9 months at recruitment proved very difficult, and we were unable to carry out the data analyses that we had hoped might inform the comparative effectiveness of different service types because of the small sample size.
Summary of findings from the whole QuEST programme
The QuEST research programme allowed us to adapt a standardised measure to assess the quality of care provided and a patient-reported outcome measure for specialist mental health supported accommodation services. These were used, along with other tools, to survey a nationally representative sample of the three main types of supported accommodation in England (i.e. residential care, supported housing and floating outreach). This represents the first, detailed description of these services and their users in England.
Although our findings indicate that most people residing in mental health supported accommodation are single men with a diagnosis of a psychotic illness, one-third of those receiving floating outreach have depression or anxiety. Although people residing in residential care have the highest needs and require most support, the majority of users of all three types of service were subject to the Care Programme Approach (the statutory framework for mental health services in England) and, therefore, received additional clinical input from a local statutory community mental health service. This is appropriate given our finding that around two-thirds of participants had a significant risk history, with the most prevalent risks being self-neglect and vulnerability to exploitation. Users of supported housing and floating outreach services were more likely to have been a victim of crime in the last year than users of residential care.
A number of similarities were found between supported housing and floating outreach services and both worked to the same time frame, with an expectation that service users would move on to less supported accommodation or manage with less support within 2 years of being taken on. Service user ratings of satisfaction with care were similar across services but, in our regression models that adjusted for differences in clinical characteristics, we found that quality of life for those in residential care and supported housing was similar and higher than for those receiving floating outreach, and autonomy was higher for those in supported housing. This has important implications for commissioners and service planners, because, although floating outreach services are the cheapest form of supported accommodation, there was a reasonable probability that supported housing provided better value for money with regard to quality of life and autonomy.
In our cohort study, we found that 41% of people progressed to a lower level of supported accommodation over 30 months but there were large differences in this outcome between service types. After adjusting for clinical differences, individuals receiving floating outreach were almost eight times more likely to be able to manage with less support than those in residential care and almost three times more likely than those in supported housing. Similarly, those in supported housing were almost three times more likely to move on to less supported accommodation than those in residential care. Successful progression to less supported accommodation was associated with the degree to which the service promoted human rights (including access to advocacy, privacy and a complaints process) and adopted a recovery orientation (which includes involving and empowering service users as much as possible in decisions about their own care and holding hope for their progress). Service users with greater morbidity were less likely to achieve successful move-on. Unsurprisingly, successful progression to less supported accommodation was associated with lower costs of care.
Our qualitative investigation of staff and service user experiences of supported accommodation suggested that staff were adopting a rehabilitative and recovery-orientated approach in their work with service users. However, the time-limited nature of services provoked anxiety among service users and staff about how the person would manage when they moved to a less intensive service.
Staff and service users were supportive of our plans for a feasibility trial to compare the effectiveness of supported housing and floating outreach services. Concerns about the process of randomisation were addressed in our approach to recruitment to ensure that referring clinicians’ and service users’ preferences were taken into account; participants were only randomised to one of the services where they had capacity to give informed consent and where there was clinical equipoise about the most appropriate service. We also included individuals who did not consent to be randomised but who agreed to complete research interviews. We included outcomes suggested by staff and service users in WP3 and a qualitative component to ensure that we captured participants’ experiences and views about the feasibility trial. Nevertheless, the feasibility trial struggled to recruit. Although we identified > 1400 potentially eligible participants, we randomised only eight. Clinical teams often declined for us to approach service users as they felt clear about which supported accommodation the individual required. Our qualitative investigation of the impediments to recruitment identified that the current ‘step-down’ pathway is ingrained in the culture of mental health services and clinicians perceived a lack of equipoise between the two models of supported accommodation we hoped to compare. Ultimately, the feasibility trial was successful in clarifying that it is not feasible to assess different models of supported accommodation in England through means of a RCT.
Implications for practice
Our results highlight the need for a range of supported accommodation services to be provided at the local level that can respond to the differing levels of individuals’ needs and we identified the pros and cons of different models. This will, we hope, facilitate a more thoughtful discussion about future service investment; during the course of the programme we were aware of ongoing cuts to the more intensive, higher-supported settings in favour of cheaper, floating outreach services. Our results suggest that all three types of supported accommodation are of value and that service planning should certainly not only respond to financial drivers. In particular, our health economic findings suggested that supported housing provided good value for money when outcomes such as quality of life and autonomy are considered. Nevertheless, we found through our cohort study that floating outreach services were able to support the highest proportion of individuals to manage with less support over time. Facilitating successful move-on is the key aim of supported accommodation services and this is, therefore, an important finding. We note that in our survey both supported housing and floating outreach service managers reported that they expected to work with individuals for around 2 years. This rather arbitrary time frame has been adopted by commissioners across the country in response to non-evidence-based parameters set, historically, by the Supporting People programme. Our results show that most people in supported accommodation do not move on within this time frame. The concrete interpretation of this 2-year expectation can pose significant risks to individuals who require longer-term supported accommodation, putting them and staff under unnecessary pressure to move on to more independent accommodation prematurely. Indeed, our findings highlight the tension for supported accommodation services between promoting service user autonomy and providing care; more supported environments were associated with higher levels of quality of life, and the promotion of greater independence seemed to place individuals at greater risk of being a victim of crime. Our qualitative findings suggested that staff and service users experienced appropriate anxiety about how the person/ they would manage with less support after move-on. It was extremely encouraging that, despite their non-professional training, staff in supported accommodation appear to have adopted many aspects of a rehabilitative and recovery orientation in their work. As we found that this approach, along with the promotion of human rights, facilitated successful progression to more independent accommodation, this should be encouraged through training and supervision.
Conclusions
The QuEST programme delivered two standardised outcome measures to assess the quality and user experiences of supported accommodation services. Our findings suggested that the current ‘step-down’ pathway, whereby individuals move from higher- to lower-supported accommodation over time, is deeply ingrained in the system without clear evidence for this approach. The two more intensive models (residential care and supported housing) were associated with better quality of life than floating outreach, and supported housing was associated with greater autonomy. Overall, supported housing represented good value for money. Services that promoted human rights and recovery were more successful in supporting service users to move on (or, for floating outreach, to manage with less support). This outcome was more likely for users of floating outreach services than supported housing and least likely for users of residential care. Users with the greatest morbidity were less likely to move on successfully.
Our feasibility trial showed that the effectiveness of different models of supported accommodation cannot be assessed through randomised trials in this country.
Our findings cannot provide clear guidance on the most effective model(s) of mental health supported accommodation. It seems likely that a range of options will continue to be required to provide appropriate support to assist the recovery of individuals with differing needs. Future investment in this area should be guided by assessment of the mental health needs of the local population and the pros and cons of the different services that our study identified, rather than being based on purely financial drivers.
Recommendations for future research
Research in the field of mental health supported accommodation is underdeveloped and there is huge scope for further studies to inform and build on the evidence base about what works and for whom. Given our finding that trials comparing the effectiveness of existing models of supported accommodation appear unfeasible, future research will need to adopt alternative methods. International surveys of supported accommodation are needed to map and compare the different models in operation. However, these need to be detailed surveys that take account of contextual factors (e.g. the morbidity of the local population and the availability of other components of the mental health system). Service mapping technology has been developing over recent years, allowing the assessment of ‘relative technical efficiency’ of specific components of health systems using Monte Carlo Data Envelopement Analysis.51 This has been applied to mental health systems and is beginning to incorporate social care system components.52 It allows information about local service provision to be understood in the context of the particular needs of the locality, providing useful intelligence for service planners. This field is growing and may prove particularly useful in enhancing understanding of the relative merits of different types of mental health supported accommodation.
Our programme included a national cohort study comparing outcomes in the three main types of mental health supported accommodation in England. Future prospective cohort studies could compare different subtypes of the three main models to inform their relative effectiveness. However, cohort studies are time consuming and expensive to conduct and it may be that developments in ‘big data’ technology will ultimately prove fruitful in harnessing comparable outcome data on a national scale. This area is currently underdeveloped.
In the meantime, more qualitative studies are needed to gain a deeper understanding of the specific components of care delivered by different subtypes of supported accommodation services that are most useful to service users. Although we concluded that it is not feasible to compare existing models of supported accommodation through randomised trials, it may well be feasible to develop and test through randomised trial interventions that aim to improve delivery of specific aspects of care, for example those interventions that we found to be associated with better outcomes in our cohort study, such as promotion of human rights and recovery-orientated practice. Qualitative methods would also help in the development of these kinds of interventions, in the process evaluation of trials evaluating them and in studies designed to inform their successful implementation.
Finally, although there is a growing international evidence base pertaining to supported accommodation, interpretation of the literature53 is hampered by a lack of comparable description of different supported accommodation services. This could be addressed by encouraging researchers and journals to adopt a consistent taxonomy, such as the one that one of our research team members (PM) developed alongside the QuEST programme,54 so that results can be compared and collated.
Research priorities
- Assessment of the sensitivity to change and test–retest reliability of the QuIRC-SA.
- International surveys of mental health supported accommodation services and service user characteristics to provide comparative data on the different models in operation in different settings and their target service user group(s).
- Cohort studies to identify the aspects of care and support associated with better outcomes to assist our understanding of how best to focus staff training and resources.
- Qualitative research to complement the findings from surveys and cohort studies to further enhance our understanding of the specific approaches and components of care that constitute ‘best practice’ in supported accommodation from staff and service user perspectives.
- Results from these studies will inform the development of complex interventions to enhance the quality of care delivered in mental health supported accommodation services. These should then be evaluated in terms of their clinical effectiveness and cost-effectiveness through RCTs.