Service provision for older homeless people with memory problems: a mixed-methods study

Dementia

Two major policy documents have shaped dementia services and provide the background to the emerging interests in dementia recognition at patient and society levels in England: the National Dementia Strategy,1 superseded by the Prime Minister’s Challenge on Dementia. 2 Both emphasise the need for earlier or more timely recognition of dementia. Commenting on the progress made under the National Dementia Strategy and the Prime Minister’s Challenge, Knapp et al. 3 reported that the provision and use of memory assessment services (sometimes referred to as memory clinics) has increased, although it is hard to detect what changes there have been and where they have occurred. The development of memory service accreditation has been slow, meaning that commissioners cannot rely on strong evidence of how best to deliver memory services. Knapp et al. 3 added:

Local services will and should respond to local circumstances and be more person-centred, but the wide variations currently seen in staff-mix, methods of clinical assessment and provision of post-diagnostic care are unacceptable if the aim is to equalise access to diagnosis and post-diagnostic support.

P. 2.

Reproduced with permission from the Policy Innovation Research Unit

Although the National Dementia Strategy1 addressed issues of equalities, generally there has been a lack of focus on people who are homeless who have memory problems or cognitive problems that are indicative of a dementia syndrome. The Ministerial Advisory Group on Dementia Research4 recommended that public and professional awareness and understanding of dementia should be improved. However, whereas it made reference to caring for people with dementia at home, in hospital and in care homes, it contained no mention of those who are homeless or socially excluded from settled accommodation. It also recommended that health and social care staff working in care homes with residents who have dementia should receive appropriate training and continuous professional development, and that community mental health teams (CMHTs) working with older people should provide specialist services to care homes; however, practitioners working in homelessness services were not mentioned. The Ministerial Advisory Group on Dementia Research advised that everyone with dementia should have access to a pathway of care services, including specialist assessment, treatment, care and support, and that this pathway should be responsive to individual need and preferences. It made no reference to possible difficulties in providing care and support to older people with dementia who are homeless, or to the staff groups working with them.

Policy has also promoted broad public health and community development approaches to responses to dementia. Knapp et al. 3 noted the considerable emphasis on improving awareness about and attitudes towards dementia through campaigns such as Dementia Friends and the development of dementia-friendly communities. Such initiatives have generally focused on people living in their own homes. 5

Equalities

Part of the justification for a National Dementia Strategy1 was that dementia has long been overlooked as a health problem, despite its often devastating impact in human terms and its impact on public expenditure. There have been arguments that this lack of priority reflects dementia’s associations with old age and the impact of ageist beliefs and stereotypes; that the condition is stigmatised; and that there is therapeutic nihilism – or a feeling that nothing can be done. 6 There is increasing interest in seeing dementia as needing a rights-based dimension, building on the ideas of the disability movement. 7,8 A report from the Mental Health Foundation suggested that this would bring several benefits, such as greater legal protection and improved service entitlement, thereby promoting cultural shifts to how dementia is perceived that will have benefits to society as well as people with dementia. 9

Thomas and Milligan10 have argued that adapting the social model of disability to people with dementia also encourages us to think beyond the social and institutional disadvantages that may be experienced by people with dementia, and instead notice the impact this can have on the places they live. As people with dementia may lose their sense of orientation of time and space, formerly familiar places may start to seem unfamiliar or frightening to them. 10 The premise that people with dementia may lose their memories of settled and familiar locations overlooks people without settled accommodation or who are unfamiliar with their present neighbourhood.

The premise that people with dementia may lose their memories of settled and familiar locations overlooks people without settled accommodation or who are unfamiliar with their present neighbourhood.

More recently, there has been greater interest in the heterogeneity of the ‘dementia population’ and the possible impact of social inequalities being sustained when a person has dementia or of such inequalities being reproduced or enhanced. Research, for example, has investigated differences in service responses, such as prescriptions of ‘anti-dementia’ medication, by socioeconomic group,11 variations in symptom recognition between ethnic groups,12 fear of prejudice about revealing sexual orientation13 and gender disparities affecting women throughout the dementia trajectory. 14,15 Concluding their literature review, Ludwin and Parker15 observed:

As we have seen throughout the material we reviewed, individuals with dementia are marked by the social locations and identities they occupied prior to diagnosis; these axes of identity may form important parts of individual sense of self. Failing to recognise them in interactions and provision of support may thus contribute to an erosion of personhood and perpetuate marginalisation.

P. 2015

Reproduced with permission from the University of York

The importance of considering equalities in respect of people who are homeless is not generally recognised, despite the very poor health and QoL of this population. 16 According to the Equality and Human Rights Commission,17 this population is among the most disadvantaged groups in UK society. Although the Equality Act 201018 provides a legal imperative to ensure that ‘protected characteristics’ are addressed in assessments of service impact (many of which are relevant to people with dementia2), homelessness is not readily translated into the ‘characteristics’ for which there is legal protection. Some advocates from the homelessness sector have suggested that the NHS Constitution instead may be an effective lever to ensure the rights of homeless people to health services. 19

Housing people with dementia

There has been growing interest in the living accommodation and housing choices of people with dementia. Much has focused on how to design new accommodation and how to modify existing accommodation so that people with dementia can stay in their own homes or move to better accommodation offering choices over patterns of care and tenure. 20 Several organisations (e.g. the National Federation of Housing21) have argued that dementia-friendly housing must be an integral part of dementia care. Research has also begun to explore particular housing needs of some groups of people with dementia; for example, Lipman and Manthorpe22 investigated how social housing providers (housing associations) were beginning to respond to tenants with dementia from black and minority ethnic groups.

As Knapp et al. 3 reported, local authorities have been encouraged to develop, in collaboration with the Homes and Communities Agency and the Greater London Authority, partnerships with housing associations (social housing providers that may offer housing and care services) and private builders to meet the needs of older people with long-term conditions, including dementia. However, the focus of such developments has been on people with dementia who already have homes, some of whom have considerable assets from home ownership and who are potentially attractive new purchasers to ‘housing with care’ providers.

Methods

Process of the review

The process of the review was that two reviewers independently screened all titles and abstracts that had been retrieved for relevance; disagreements were resolved by consensus. If relevant, the full-text article or publication was obtained. The full text of obtained articles was judged to be potentially relevant by the two reviewers. A list of key questions was generated, and a pro forma (data extraction tool) was developed to address these from the included papers. These were piloted on a small number of selected papers. Owing to the few papers identified, unlike Burra et al. 61 and Ennis et al. 62 we did not adopt a critical appraisal approach to rating the quality of studies, thereby following procedures generally categorised as scoping reviews. 63 One key reason for this is that such an approach risks overlooking studies with small sample sizes and, thus, might exclude individual case study or whole-hostel research if the numbers of older residents were not sizeable. Ennis et al. 62 considered a sample size of over 100 to be ‘good’, but, as discussed below, we found few studies reporting such numbers.

We identified 20 articles for inclusion in the review, after screening 260 articles retrieved from our database searches (the searches were run in May 2016). We searched English-language databases only (Web of Science, Scopus, Social Care Online, PubMed and The Cochrane Collaboration) for material published from 1980 to 2016, including ‘grey’ literature [Figure 1 shows the flow chart – a modified PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram – for the process of literature review].

FIGURE 1.

Literature review A flow chart (a modified PRISMA diagram).

We cross-checked our findings with a recent (2014) literature review from Australia conducted by a researcher at La Trobe University,64 in which the period 2004–14 had been searched for items from English-language literature, using the terms cognitive impairment, dementia, care needs and illness of homeless people. Chenco undertook one search for ‘dementia’ and ‘homeless’ and another for ‘illness’ and ‘homeless’, and then combined the findings into a single larger list. 64 This resulted in the location of 344 references, once Chenco had removed duplicates. Of these, 35 were deemed directly relevant for the purposes of Chenco’s Australian review, which, as noted previously, took a much broader scope than our study by including the term ’illness’. 64 In this search of four main databases, Chenco located the following numbers of items:

• Social Services Abstracts (via ProQuest) (n = 241)

• AgeLine (via EBSCOhost) (n = 24)

• Australian Family & Society Abstracts Database (via Informit) (n = 84)

• Health & Society Database (via Informit) (n = 73).

In addition to comparing our searches with Chenco’s findings,64 we took particular account of the findings of other scoping and systematic reviews of similar subjects. In the review by Spence et al. ,65 for example, the following terms were searched for in the titles, abstracts and any subject heading fields: ‘homeless*’ or ‘roofless*’ or ‘fixed abode*’ or ‘hostel*’ or ‘bed and breakfast*’ or ‘night shelter*’ or ‘hotel*’ or ‘housing benefit*’ or ‘street dwell*’ or ‘tramp’ or ‘tramps’ or ‘vagrant*’ or ‘vagabond*’ AND ‘cognit*’ or ‘executive function’. We did not search for the terms ‘tramp’ or ‘vagabond’, but it is interesting to note that such terms were evidently in use just over a decade ago in the UK. There are wide variations between services that provide hostel accommodation and night shelters, both internationally and in the UK and in terms of their residents or user profiles; these need to be taken into account in any comparison and are mentioned in Findings.

Findings

Our searches found a small number of studies covering homelessness and memory problems/deficit/dementia and Appendix 1 contains a full table of the studies included in the second literature review. Of these studies, one of the earliest to examine these subjects in terms of prevalence was undertaken in Australia by Teeson and Buhrich. 66 Over 20 years ago, these researchers used the Mini Mental State Examination (MMSE),67 which is still commonly used in UK dementia services, and estimated that severe memory problems affected 25% of one hostel’s 65 residents, and another 40% showed signs of mild memory problems.

To establish if the high prevalence of severe memory problems was more commonly borne out in other hostels, and because no women had been included in their 1993 research,66 Buhrich et al. 68 later conducted a broader study of homeless people, again in inner-city Sydney, NSW, Australia. Using the MMSE67 with 204 participants (155 men and 49 women) in seven hostels, they found that 10% (20 people) had some form of memory problems. Of the 14 men with memory problems, five had severe memory problems. The six women had mild memory problems:

The prevalence of cognitive impairment among the homeless subjects in this sample was 10 percent . . . the prevalence of cognitive impairment in the general adult population is 1.7 percent . . .

Buhrich et al. 68

In this study, MMSE scores of 0–17 were interpreted as indicative of severe impairment and scores of 18–23 were interpreted as mild impairment (although six people were not included in these figures as they were unable to complete the MMSE, were deaf, or could not speak English). The researchers found that people with memory problems were likely to stay in the hostel accommodation for longer than those without, postulating that this might be because ‘move on’ accommodation was much harder for them to find, or because their motivations to move on were compromised. 68

Services

One point implicitly raised by Rota-Bartelink25 was whether or not memory problems matter in the context of providing services to homeless people. She suggested that they do indeed matter because they affect people’s ability to keep or get stable housing and to benefit from support. The person may need support to make decisions and judgements, regulate their behaviour, and build and keep social relationships. A person’s ability to benefit from interventions has to take into account their cognitive abilities and disabilities. Services should beware of devising or agreeing to outcomes that do not reflect these challenges. The thesis that memory problems bring extra vulnerability was echoed in Grenier et al. ’s55 review, which pointed to the likely needs of older homeless people as being for safety and, by extension, the heightened vulnerabilities if a homeless person has dementia.

Two further reviews of the service needs of older homeless people were found. Stergiopoulos and Herrmann77 noted that some researchers had concluded that older homeless people are likely to be missed by studies that recruit from large services because they are ‘hidden’ or do not use such services and sometimes do not access other related services, such as primary health care. The service needs in their review were mainly indicated by reports of surveys and interview-based studies that pointed to the range of problems commonly found among older people who are homeless or the ‘chronically mentally ill elderly’. 77 They noted that better understanding of the events that precipitate homelessness might inform prevention measures or the pathways to homelessness. However, they also suggested that there exist ‘age-specific unmet physical and mental health needs among homeless seniors, particularly among those with mobility and memory problems’. 77 They proposed that it would be helpful to think about services for all homeless people as well as those for older homeless people; the latter experience particular barriers to accessing some services, such as physical inaccessibility, and often have a legacy of dissatisfaction with services and lack of health card identification (proof of entitlement). One concrete suggestion was that local consideration should be given to day centre programmes to engage older homeless people and to provide more acceptable and accessible assistance.

Stergiopoulos and Herrmann’s study77 was one of the few found that collected data from multiple sources, obtaining data from local health and community services in Toronto, ON, Canada, and surveying 11 hostel directors (eight responded) (in addition to completing a literature review). The directors reported that although older (aged > 65 years) male hostel residents were a very small minority of hostel users, and likely to have been homeless long term, the problems that gave rise to most concern among hostel staff were older residents’ memory difficulties, verbal aggression and alcohol misuse. Fewer in number, older female hostel residents were more likely, in the directors’ experiences, to have memory problems, symptoms of paranoia and depression. The directors expressed concern that overall they were not meeting older residents’ memory and mobility needs.

Burra et al. 61 systematically reviewed studies of cognitive deficits in homeless adults, and discussed their implications for services. They found that, among the studies that had administered the MMSE, between 4% and 7% of homeless people were reported as having global cognitive deficits. The consequent implications for services were thought to lie in problems with clients’ adherence to treatment regimes, their difficulties following rehabilitation programmes and their lack of participation in ‘training’.

Likewise, Brown et al. ,73 whose study included older homeless people living in a variety of environments, concluded that there were few differences in the mental and physical health status of older homeless people by location (temporary or rooflessness), many of whom had disproportionately high levels of long-term conditions compared with the general older population, and even when compared with older people living in poverty. These authors suggested that current service efforts in areas such as rehabilitation, environmental adaptation and addressing polypharmacy may not have much traction with older homeless people, whose key need seems to be for permanent supportive housing.

Some researchers concluded that homeless people not only need stable housing but also need support. For example, Lippmann78 argued that older homeless people need to be able to access more of mainstream older people’s (aged care in Australia) resources and funding. Van Wormer79 similarly proposed that older homeless people need assistance with gaining and retaining housing. A review of the literature on housing options for older people80 reported the wide spectrum of housing services, noting the importance of distinguishing between supported and supportive housing, for example the different models of shelters, and the growing evidence base for the ‘housing first’ approach that sees housing as a necessary preliminary for rebuilding people’s lives.

For older people, it is reported to be even more difficult to provide care and resources if they are living on the streets. 81–83 O’Connell et al. 84 noted that diagnosing dementia is difficult ‘on the streets’ as assessment necessitates access to tests and the ability to take details of a person’s history. In this unique longitudinal study of older homeless people in Boston, MA, USA, where the authors followed up 30 people aged ≥ 60 years for 4 years, multidisciplinary teams of health professionals undertook outreach work regularly.

There have been few evaluations of services for older people with dementia or similar conditions who have been rehoused in care facilities. One of the few is of the Wicking Project in Melbourne, VIC, Australia, a model of residential care for older people with dementia and challenging behaviour arising from alcohol-related brain injury. This started in 2006, as described by Rota-Bartlink and Lipmann60 (Lipmann is the chief executive of the provider organisation, Wintringham, a not-for-profit homeless service for people aged ≥ 50 years). Using a household model with an emphasis on harm minimisation and on meeting the needs of people who ‘shuffle transiently between organisations that cannot provide long-term care management solutions,’60 this facility offers individual care plans and specialist case management.

Findings from interviews with 50 stakeholders in Sydney, NSW, Australia, about possible best practice for older homeless people were summed up by Conroy et al. 72 as suggesting the need for persistent engagement with individuals, taking a harm minimisation approach, targeting people for assessment of possible mild cognitive impairment, and improving links and referrals systems between agencies. 72 Backer and Howard85 had earlier noted that, although prevalence estimates of cognitive problems varied, the effects of such impairments were that older homeless people could find it difficult to access help and support or to sustain engagement with services when their housing was at risk.

Barrett et al. ,83 reporting on older homeless service provision in Florida, USA, claimed that there was an increasing need to understand and work on the different ways in which older homeless people gain access to services and the pathways designed. Unsurprisingly, perhaps, they suggested that better collaboration between services and agencies was needed.

In one of the few discussions about distressing or challenging symptoms of dementia that may be experienced in service settings supporting previously homeless older people, the homelessness provider mentioned previously, Wintringham,78 welcomed proposals for a funding supplement to cover the extra costs needed to provide such support. It commended an approach that was ‘behaviour based’, although it queried whether or not the proposed assessment tool would be appropriate for people who had been homeless. Furthermore, it suggested that a wider range of staff be approved to undertake such assessments, noting that, in homelessness services, members of staff were not generally clinically qualified. It further speculated that some people’s behaviour might be so challenging that a ‘top-up’ or an upload to the funding supplement should be considered. Such detailed considerations of cost-effectiveness and service models are rarely encountered in the literature.

Staff training and skills have been considered in some studies, although there does not appear to have been any evaluation of skills development or of training packages or workforce models. Concurring with the conclusions of other studies,81,86,87 Rogoz et al. 70 suggested that improved access and delivery of services to homeless older people with memory problems should be undertaken by specialist, multidisciplinary teams of health professionals. The skills reportedly needed were summarised as spanning the areas of outreach, assessment and multimorbidity management. In addition to advocating for more research on older homeless people following her review, Chenco64 argued the importance of better understanding of the educational, training and resource needs of health professionals who work with homeless older people.

More specifically related to the content of clinical services, Andersen et al. 88 recommended screening for traumatic brain injury after finding that this seemed to explain the poor performance in cognitive testing among a sample of 34 homeless people in Canada, whose average age was 58.8 years. Further related to health services, Abdul-Hamid et al. 89 had earlier raised questions about whether clinical services for older homeless people should be under the responsibility of geriatricians or psychogeriatricians.

However, as reported by other commentators, some older homeless people refuse to be assessed or evaluated, and questions of competency and decision-making autonomy may forestall service engagement. 84 The legal position of having a proxy or court-appointed decision-maker is not generally addressed in most studies, although the first report of the Wicking Initiative60 mentioned briefly that some residents had an administrator appointed to make decisions (in England and Wales this would perhaps be an appointee for benefits because other provisions of the Mental Capacity Act 200590 would apply only if decision capacity was substantially impaired).

Backer et al. ,85 among others, have pointed to the need for services to prevent homelessness by addressing its causes and risk factors. The skills to follow that path are not delineated, nor is how to change service commissioning. The Wicking Initiative, discussed above, remains one of the few service examples of tertiary provision described in any detail. We do not know much about what happens over time to older homeless people in other settings who are recognised as having memory problems or a dementia. As one example of the importance of this, of the 15 older homeless people with dementia contacted by an outreach initiative91 in the city of Tel Aviv, Israel, 11 later moved to a nursing home, with unreported outcomes. More generally, this study recommended intensive outreach and case management to effectively assist older homeless people.

Phase 1 design

The first phase of the study aimed to recruit and conduct short ‘screening’ interviews with up to 200 older homeless people aged ≥ 50 years living in hostel accommodation in England, with a maximum of 50 people in each of four diverse study sites. These were the Midlands (five hostels), Central London and South London (three hostels) and the North East of England. As noted in Chapter 1, we used the former UK Coalition on Older Homelessness definition of ‘older’ homeless people as those aged ≥ 50 years.

At the time the study was being designed, managers of several hostels in the sample areas were contacted to find out if they had any residents aged ≥ 50 years and to ascertain their interest in participating in the study. There was an enthusiastic response, with the managers welcoming the need for the research as they perceived the needs of older homeless people to be growing. In the North East of England, however, the numbers of older homeless people were thought to be small in the areas approached, and, consequently, four further hostels were identified in the North East of England that potentially could supplement numbers.

The recruitment and screening interviews of older homeless people aimed to obtain a profile of each participant and an overall assessment of their problems and needs. These also aimed to enable the research team to consider whether or not a participant had any memory problems and, if so, to determine whether or not depression was an influencing factor. This enabled participants to be selected for the case study phase (phase 2). These interviews were designed using a semistructured interview schedule to collect the following information: (1) demographic factors (age, gender, ethnicity); (2) brief histories of homelessness (i.e. length of time homeless, duration of stay in present hostel); (3) previous and current employment, education, income and welfare benefits; (4) use of time and family and social contacts; (5) support received with personal care and household tasks, such as obtaining meals and personal hygiene; (6) past and current physical and mental health problems (including administration of a validated depression tool), substance misuse problems, histories of head injuries and other accidents or traumas, other diagnoses including chronic or long-term conditions, past hospitalisations, and medication and treatment; (7) concerns about memory and support or treatment received, and administration of the MMSE; (8) the use of different health services; and (9) QoL, and hopes and plans for the future in terms of housing and support.

We were conscious that older people with memory problems and in marginal living conditions could be a vulnerable group in terms of research as well as in much of their lives. In our application for ethics approvals, we addressed the risks that they might feel obliged to co-operate in the study, might find the questions intrusive or upsetting and might have expectations of more help. Two senior researchers (MC and LJ) had considerable experience of research and practice in the homelessness sector and they supported the wider research team in fieldwork. One ethical challenge was how to acknowledge participants’ time in study participation. We decided that, at the end of each interview, older homeless participants would be offered £5 to thank them for their time, either in cash or as a shopping voucher: the type of incentive was negotiated with each hostel manager.

Once participant baseline interviews were completed in each hostel, each person’s key worker was contacted and invited to take part in the study. An information sheet about the study was designed for hostel workers, and workers were shown the consent form signed by the hostel resident that gave consent for the worker to be interviewed about them. Hostel workers were also required to provide written consent to be interviewed. This interview included questions on (1) the characteristics of the hostel resident; (2) how the worker assessed and planned support for the resident; (3) their perspective on any physical, mental health, substance misuse and memory problems and the support and treatment provided; (4) what assistance, prompting or supervision was provided on a day-to-day basis by hostel staff or other services; and (5) future plans for the hostel resident. If a worker suspected that the participant had memory problems, they were then asked, based on the informant interview questions in the General Practitioner assessment of Cognition (GPCOG) tool,92 the reasons for this, such as whether the person (1) had problems remembering things that had happened recently, (2) could not recall conversations a few days later or (3) had difficulty finding the right word or tended to use the wrong words.

To gain an understanding of the hostel staff’s training and experience, a short structured questionnaire was also designed to be administered at the first interview with each practitioner. Topics were (1) current role, length of time working in the hostel and hours of work; (2) previous experience of working with homeless people; (3) other work experience; (4) training related to working with homeless people and training on supporting someone with mental health or memory problems; and (5) training that was/would be helpful in working with older hostel residents.

To collect background information about the hostel and the services it provided, an interview was also planned with the manager or senior practitioner in each hostel. An interview schedule was designed to collect information about (1) the hostel management, facilities and funding; (2) client groups and access policies; (3) tenure policies; (4) staffing levels and training available for staff; (5) help and support provided to hostel residents; and (6) characteristics of residents with memory difficulties, support provided and services available.

After the baseline interviews were completed with hostel residents and hostel staff, data were summarised so that each participant could be discussed by the research team psychiatrists to enable participants to be allocated to observation or control groups. It was initially intended that decisions regarding allocation to the observation or control group would be made with the 30 hostel residents suspected of having memory problems and a random selection of eight residents not suspected of having memory problems being invited to take part in an interview with the research psychiatrists using the Addenbrooke’s Cognitive Examination tool version R (ACE-R),93 which would be used to confirm the group allocation. This tool was later replaced by the Addenbrooke’s Cognitive Examination III (ACE-III) following further discussion. The ACE-III is one of the most popular and commonly used cognitive tests in dementia clinics and in the assessment of other neurological disorders. It provides a sensitive, reliable, secure and easy-to-administer clinical tool to assess cognition as part of the process of assessing for dementia. The ACE-III has five subdomains, which provide a cognitive score out of a maximum of 100: (1) attention, 18 scores; (2) memory, 26 scores; (3) fluency, 14 scores; (4) language, 26 scores; and (5) visuospatial, 16 scores. The assessment also involves asking participants if they have problems with face recognition, remembering names of people or objects, and problems with their short- or long-term memory. As with other hostel residents who participated in interviews, £5 was given for participation.

Phase 2 case study design

The 6-month longitudinal case study phase aimed to conduct case studies of 60 older homeless people – 30 with memory problems and 30 without memory problems – divided between the study sites with the samples matched on age and duration of their stay in the hostel. These 60 participants would be selected from the 200 participants recruited in phase 1. A case study methodology was chosen as it would enable an in-depth exploration of the predominant factors and challenges that staff in homelessness services face in a real-time context,94 which can be especially useful in researching practice settings. 95

Recruitment of study sites and participants

Ethics approval was obtained from the London and South East Research Ethics Committee in April 2014 (14/LO/09373) and fieldwork began in May 2014. Once fieldwork for phase 1 commenced, however, the recruitment plan for the study sites changed substantially in some areas. This was due to the closures of some hostels that had initially agreed to participate, and changes to the resident profiles. The team contacted many hostels, considered other areas to try to supplement numbers and attempted to identify replacement sites. A summary of the changes to the sites, the nomenclature used and the numbers of participants recruited is shown in Table 1.

Recruiting, obtaining consent and interviewing older hostel residents

Once the initial sites had agreed to participate, three members of the research team (MC, LJ and KS) met with the site managers to further explain the study and provide information sheets for residents and staff. A few days before the team visited each hostel, hostel staff spoke to residents aged ≥ 50 years, provided them with an invitation letter and study information sheet, and sought their permission to tell the interviewers if they were interested in taking part. When the interviewers visited the hostel, residents who had expressed an interest were contacted. The researcher explained what was involved, specifically that (1) participation was voluntary and they could withdraw at any time or refuse to answer any question; (2) that they would be seen three or four times by a researcher over the following 6 months, one of whom would be the research team psychiatrist for some questions related to memory and concentration; (3) that their consent was needed to collect information from hostel staff, their hostel records and their medical records; and (4) if they said something that indicated that they or someone else was at risk of harm, the interviewer would be obliged to take appropriate action. Permission was also sought for the researchers to make enquiries about participants’ whereabouts from hostel staff, and nominated services, family or friends if they could not be contacted for future interviews.

Written consent was obtained before the first interview, with verbal consent obtained before subsequent interviews. Only hostel residents who gave informed and written consent were recruited. Participants’ capacity to give informed consent was determined before each subsequent interview. It was planned that, should a participant lose capacity during the following 6 months, they would be kept in the study and, in accordance with the Mental Capacity Act 2005, a nominated consultee would be involved. This, however, was not necessary. For participants who were heavy daily drinkers of alcohol, interviews were arranged in the morning before they became intoxicated or fell asleep.

There were several challenges in recruiting residents to the study. Despite numerous visits to hostels, recruitment took much longer than planned and a greater number of older people than anticipated declined to take part. One problem was that the title of the study led to misunderstandings. Some hostels persistently referred to the study as the ‘dementia study’ and at times staff perceived that only people with memory problems were needed. Consequently, some residents were reluctant to take part as they felt that they did not have memory problems or dementia. Repeated efforts were made by the research team to challenge these misconceptions but it is likely that this had a negative effect on participation.

Being reliant on staff to initially speak to residents and explain the study brought other challenges. Some members of staff were very interested in the study, discussed it with residents, encouraged them to take part, and readily liaised with the research team. In other hostels it was more difficult. This was largely because of work pressures but insufficient interest in and unfamiliarity of being involved in research may have also influenced the way in which the study was explained and the extent to which residents were encouraged to take part. In one hostel in North England, only 2 out of the 16 eligible residents agreed to participate.

Practical aspects also affected when interviews could take place as the research team had to work around the shifts of hostel staff, staff handover meetings and other activities, such as resident meetings and groups run by visiting alcohol/addictions or health workers. In two hostels, members of staff were only on site until 4 p.m. so, for safety reasons, researchers could not visit after this time. Another hostel only permitted researchers to visit once per week.

Arranging and conducting interviews were also complex tasks. As outlined previously, many hostel residents were heavy consumers of alcohol and illicit drugs. On two occasions, interviews had to be completed at a second visit: one person fell asleep in their first interview and another had to attend an unexpected case review. Some participants had been homeless for many years and were vague about past details. Therefore, capturing the required information was often difficult. Some questions had to be rephrased and asked in different ways, and details needed to be checked at different points during the interview if contradictory information was provided.

Nonetheless, a total of 62 older homeless participants aged ≥ 50 years were recruited to the study. Fifteen people (24%) were lost to follow-up and 47 were included in the case study analysis. Several reasons were offered as to why people either dropped out of the study or could not be followed up. These are shown in Table 2.

Undertaking cognitive assessments

Changes were made to the study design in relation to the various cognitive assessments undertaken in both phases of the study. In phase 1, the MMSE was replaced by the Six-Item Cognitive Impairment Test (6-CIT). This was because it was felt that the 6-CIT was more achievable to conduct as an initial screening tool, as it was shorter than the MMSE and did not involve a requirement to draw.

Although the study was designed so that only some participants would receive further cognitive assessment by the research team psychiatrists, by the time the data of the first group of residents were discussed by the research team it was apparent that recruiting up to 50 participants per site was not feasible and that it would be a challenge to obtain 60 participants for the case study phase. It was therefore decided that (1) all participants would be taken into the case study phase and (2) to ensure consistency in allocation to observation or control group, all participants would be invited for cognitive assessment. Appointments were consequently arranged with each participating hostel resident for a psychiatrist or a psychologist (RN, Tanya Walton and Elspeth Dustagheer) to conduct a short (20-minute) assessment of cognition with them where they lived. Although it was planned that the ACE-R would be used, this was replaced by the Addenbrooke’s Cognitive Examination tool III (ACE-III) as copyright permissions were needed for the ACE-R.

On reviewing the first set of residents’ data, the research team noted that some participants struggled to answer the 6-CIT questions about memory and concentration and therefore might struggle with the ACE-III. It was therefore proposed that the shorter Montreal Cognitive Assessment (MoCA),96 which captures executive function of memory more succinctly, should be used with participants for whom the ACE-III could be too burdensome. This would theoretically enable all participants to undertake a cognitive assessment and reduce the number of missing data. The ACE-III was considered to still be useful for any subanalysis but the MoCA would give a good indication of whether or not an individual had memory problems. An application for a substantial amendment to ethics permissions was made and accepted at the end of July 2014. Nonetheless, 48 ACE-III assessments were conducted and the MoCA was attempted with only one participant.

Interviewing hostel staff

There were challenges in completing interviews with hostel staff. They were often very busy and found it hard to make time and, in a few cases, long-term illness meant that some could not be interviewed as planned. Consequently, on some occasions baseline questions were combined with 3-month interviews. In one hostel, members of staff were interviewed only once at 6 months. When possible, the same member of staff was interviewed each time; however, the high turnover of agency/locum staff in one hostel meant that different staff were acting as key workers for residents at baseline and at 3 and 6 months. Even when appointments with workers had been arranged, at times interviews were cancelled if there was a crisis in the hostel. Nonetheless, interviews were held with 44 hostel workers. All were recorded with consent, transcribed and analysed thematically, and linked to any relevant resident for whom the staff member acted as a key worker (see Data sources).

Data sources

Extensive attempts were made to obtain a ‘complete’ data set that included participants’ self-reports, hostel worker reports and information from hostel and medical records. However, the challenges in conducting the fieldwork explained previously meant that this was not possible in many cases. Although the majority of hostels provided full access to electronic and paper records, ensuring that all relevant information could be collected, in a few cases staff either completed the extraction themselves or acted as gatekeepers to data, making it difficult to be certain whether or not all of the relevant information had been provided.

Health centres were also difficult to contact, and, although medical records were obtained for 30 participants, nine participants’ GPs did not respond to data requests, one hostel resident did not have a GP and another’s GP was contacted only to find that the resident had not used the practice during this time; four practices reported that they were unable to provide the data as the resident was no longer registered with them. In one case, the hostel resident changed his GP in the last month of the case study period. Efforts were made to contact the second practice, but it did not respond. Not being able to pay general practices for the provision of data was a stated barrier to their willingness to provide data, but this did not appear to be the only factor. The different data sources and the number of participants for whom data were collected are shown in Table 3.

Data analysis

Sites

We outline in this chapter the salient characteristics of the hostels participating in this study. We have grouped together some identifying features, such as the exact numbers of beds, and some other details that might be easily recognisable and so compromise anonymity. We have divided the hostels into three groups for ease of reading – the Central London group (two hostels), the South London group (two hostels) and the North England group (three hostels). This does not mean that the hostels were totally representative of these geographical areas. Data from the interviews with managers are used to further set out the context of the participating hostels.

Duration of resident stay

In this section, we report on the salient characteristics of the participating hostels. As outlined in Chapter 1, there is an increasing trend for hostels to require residents to move on ‘quickly’, and hence the average length of stay or maximum period of stay was of interest, in particular in noting whether or not the participating hostels made exceptions for certain residents or groups. We found considerable variation in how long residents were able to stay in each hostel. The reasons for longer stays were multifactorial, and appeared to include lack of housing options, increasing care needs, and whether or not the hostel owners/managers considered themselves as offering a ‘home’. For example, in one of the London hostels (Central London hostel 1), although the maximum stay was 2 years, we were told that if a resident is not able to cope independently with their alcohol problem or is ‘elderly’, or ‘does not fit into categories [for move on], or can’t look after themselves’ – or they are in ‘limbo’, for example not being ‘ill enough for a care home’ (not meeting local authority eligibility criteria), then they ‘may be able to stay’. Central London hostel 2 declared a maximum 18 months’ stay, whereas South London hostel 2 cited a maximum of 3–4 months but then acknowledged that this might be ‘flexible’. Although still aspiring to a 2-year maximum stay, in South London hostel 3 about 15 residents had been living in the hostel for some years, and the staff reported that suitable housing with care to move them on had not been found for half of them. According to the manager:

Those on older contracts predating 3 years ago, this is generally when no other service can meet their need adequately. And, in some cases, it is when social services has come in, assessed the person, and despite very obvious high needs have deemed them as not needing another place of care. Social services know we provide a lot of support to clients here so they are happy to leave us to it, but it goes against our contract and it increases our workload when social services avoid their duty of care.

Manager, South London hostel 3

Even more variations were evident in the hostels outside London. In North England hostel 1, although there was a limit of 140 days’ (4–5 months’) stay, there were two long-term residents, one of whom had been living in the hostel for nearly 20 years and was said to have a mild learning disability and no close family; ‘this is his home’, declared the manager, who added ‘when he first moved in, there would have been no pressure to move’. The other long-term resident had been living in the hostel for over a decade. In North England hostel 2, the official stay was a maximum of 24 months and we found no exceptions to this. In North England hostel 3, the maximum stay was stated to be 3 years but the manager declared being under no pressure to move residents as there was ‘nowhere for them to go’. By contrast, in North England hostel 4, stays were limited to 6–12 months, with a very high turnover of residents; staff reported that they helped ‘move on’ 8–10 residents each month.

Housing options

Similar variations were found in the housing options open to hostel residents. The metaphor of ‘fighting’ with other agencies was used in a couple of London hostels to emphasise that interagency working was not always smooth-running. In Central London hostel 1, the manager reported that, of their older residents:

A few go into care [homes], especially if doubly incontinent. Few continue to remain in hostel. One is awaiting supported housing; have to fight for it. We have to fight to get them into registered care homes. It can take 1–2 years to get them in. In the meantime we get carers [care workers] in and then we see the person gradually go down. It is cheaper for them to remain here, but then sometimes if they end up in hospital we have to then say we won’t take them back as their needs are too high and we have to fight not to take them back.

Manager, Central London hostel 1

South London hostel 1 reported similar tensions, here described more as a ‘constant battle’, in respect of the local authority adult services department (social services). Its manager explained:

This has been a big bone of contention with social services, as I explained before. They tend to view us as a care home and leave us to look after clients who get ever more frail and needy. We try and fight for nursing home placements for the most vulnerable but it doesn’t always work. We had social services come in and assess our most vulnerable clients as not needing another placement when he had a carer [care worker] come in once a day, needed his food cut up, needed help feeding, was having difficulties swallowing, zero mobility, very, very vulnerable, and obviously needed nursing care that we are not trained to provide. Since I started, I have asked our commissioners to step in three times – they have spoken to the relevant head of social services and clients have been reassessed and rehoused. The one whom I mentioned before went from ‘not needing any more care’ to requiring the highest level of care. So it’s a constant battle that we have to fight.

Manager, South London hostel 1

In Central London hostel 2, moves for older residents were more often to sheltered accommodation, but sometimes to other hostels. The manager here commented that the presence of memory problems among residents constituted a ‘big stumbling block to finding appropriate housing’.

South London hostel 2, the hostel site with the highest turnover and most restricted length of stay, declared itself not suitable for those with memory problems; its manager considered that such residents would need a smaller, semi-independent service with greater support.

Central London hostel 1 reported more day-to-day working with the local authority adult services department than with the local authority housing department. The manager commented that, for its older residents:

Referrals often come via social services and alternative housing for people with memory problems is most often found through social services.

Manager, Central London hostel 1

However, in North England hostel 3, the manager considered that local cost pressures meant that rehousing for older residents was limited:

Care home costs are too high compared to hostels (£550 per week) so the local authority is reluctant to move them.

Manager, North England hostel 3

He added that some residents did not want to move to locations such as care homes because ‘drinking [of alcohol] will not be allowed’. This led to an unsatisfactory situation for residents who, in his view, should not be living in hostels:

There’s nothing, it’s us or nothing. We try our best, not really equipped to deal with it, [we] try and help as best we can.

Manager, North England hostel 3

Experience in North England hostel 4 was that of fairly rapid turnover and most residents moved to ‘own tenancy or try social housing with support, e.g. sheltered or tenancy support and any other possible support’. However, the amount of support for individuals following moves was thought to be shrinking; the manager here reported:

Due to lack of funding, [we’re] looking at church-based community groups for support to plug the gap.

Manager, North England hostel 4

Support for residents with memory problems

Despite their differences in locations and policies, the accounts of hostel managers in describing the extent of support provided by some members of staff for residents with memory problems were similar, and at times such support was considerable. The manager of Central London hostel 1 reported:

We have to escort them everywhere, sometimes we have to do missing persons reports. They [residents] get upset when they realise what’s going on and they get frustrated and if you speak to them about it they can get very angry with you as they don’t believe you . . . If it gets very bad and we can’t cope we get social services in for assessment for a care package. Some are reluctant to get help and say they are OK. We constantly have to update the risk assessment. They also have unrealistic expectations at action plan meetings. For example, one man thinks he will be able to go back to work. How they think they are is very different to what we see. Their relationship with alcohol doesn’t help.

Manager, Central London hostel 1

In Central London hostel 2, substantial support could be provided inside the hostel, as the manager reported:

If they have appointments you have to keep knocking for them. They say – give me 5 minutes and then you have to go back again and again. They forget the real time. You would think your body would tell you. Not being able to do simple tasks.

Manager, Central London hostel 2

In this hostel, staff obtained appointeeship (authority to manage welfare benefits and pensions) for some residents. This indicates the high level of need among some residents in activities of daily living and also illustrates the extent of staff engagement with budgeting and money management. South London hostel 1 also reported having:

. . . a harm minimisation programme, which I think is unique to us. So what we do, with the voluntary agreement of clients, we hold their money, we hold their bank card and we give them a daily allowance. We also buy them and give them alcohol – this is always less than 6% (alcohol) beer or cider and give them in stages – so three times a day and two cans only. We are essentially working with them to avoid them spending all their money in one go and bingeing on potent alcohol and harming themselves. Many feel this is controversial as we are then buying them alcohol, but we feel it reduces the risk of overuse and bingeing.

Manager, South London hostel 1

The manager of South London hostel 1 listed some of the manifestations of memory problems among its residents as confusion, forgetfulness, aggressions, falling, wandering and getting lost, and again spoke of the complexities when such behaviours were hard to differentiate from the impact of drinking; they do ‘virtually everything, ensuring safety, minimising risk, try and arrange memory assessments’. A couple of residents who were ‘wandering and getting lost’ were supplied with wrist bands (at a cost of £20 each) by the hostel containing their identification details. Concerns about safety and adult safeguarding (protection from abuse and exploitation) were also noted in South London hostel 2, where managers reported that staff needed to be alert because certain residents were ‘forgetting they are vulnerable’ and that their lack of judgement meant that others took advantage of them. All of this was complicated by general difficulties in engaging with certain residents, particularly if their behaviour was challenging to staff or others or distressing, or if they were difficult to motivate or slowing down in their progress to resettlement. In North England hostel 2, the manager reported that residents with memory problems required ‘more intense support, reminding appointments, time spent on routine tasks, going to appointments with them’.

Other forms of surveillance or monitoring were evident; for example, the manager of North England hostel 1 reported that staff:

[H]ad to keep a closer eye on and make sure one person had meals – but then he’d forget he had already eaten. We can’t keep any medication for people we don’t have any form of care provision. We do help them to get a dosette box (for pills) and remind them about meals.

Manager, North England hostel 1

However, this was exceptional as the practice in this hostel was to move such people on:

If we do get someone with memory problems, we get them moved on as quickly as we can.

Manager, North England hostel 1

North England hostel 3 similarly reported encouraging residents to eat and complete routine tasks but also efforts to maintain hygiene; as someone’s poor self-care worsened, this could be accompanied by lack of motivation and energy, or, on the other hand, a resident could become non-compliant and aggressive. Several managers commented on the complications of substance misuse, which sometimes excluded people from receiving support from other services but could also mask the extent of memory problems and confusion.

Referral routes for residents with memory problems

As noted above, there were some accounts of accessing specific help for residents with memory problems, who, in addition to needing more time for support than other residents, had higher support requirements. Some needed prompting and frequent reminders, particularly in terms of managing appointments, and generally tailoring the approach to their support. The process for referral to support with memory problems was generally to advise the resident to contact their GP or to do this on their behalf:

We refer them to the GP for a referral to the memory clinic.

Manager, Central London hostel 2

However, Central London hostel 1’s manager had one experience following making a referral of a resident to the local memory service, when the resident ‘turned up drunk’ for his appointment. In his view:

They didn’t cater for homeless people. Now we refer to alcohol services and are happy about it. We also have to refer people to the continence service as well.

Manager, Central London hostel 1

Similarly, in Central London hostel 2 referrals were made to the resident’s GP or alcohol services, who decided whether or not a referral for dementia assessment was needed. The difficulties of accessing help for residents who consumed large amounts of alcohol regularly were again voiced by South London hostel 1’s manager, who knew that, theoretically, residents with memory problems or similar could be directly referred to the local CMHT for assessments, but for this to happen the resident needed to be sober, which the manager observed was not always possible to arrange. Likewise, North England hostel 3’s manager summed up the lack of referral pathway as being again related to ‘alcohol acceptability’, stating more generally that there were:

No services, nothing for them. Can’t be assessed in memory clinic until they’re dry.

Manager, North England hostel 3

Only one out of the four North England hostels had a referral route for residents with memory problems, indicating their generally weaker links with secondary NHS services.

Primary care input

In this section, we report on the variations of primary care services’ input to the hostel as summarised by the hostel managers. North England hostel 3 was an outlier, in that no input from primary care was reported by its staff or observed on our visits or in the records. Other hostels had different types and levels of engagement that varied from proactive to reactive. In the London hostels, overall, there was more regular and planned input from primary care teams to the hostels, although in Central London hostel 1 a GP surgery and a walk-in centre were also located nearby. In this hostel, a nurse from the local drug misuse service also visited weekly to undertake health assessments and to see new residents. A different model of primary health care operated in Central London hostel 2, as staff from a specialist primary care intervention for homeless people visited the hostel weekly.

A similarly separated and specific approach was evident in South London hostel 1, where there was a weekly GP visit, a weekly visit from a nurse from an integrated care pathway service, and, in addition, weekly visits from staff from a Homeless Intermediate Care Project (described by its staff as providing ‘wrap around’ care for the most vulnerable residents, including taking them to appointments and helping hostel staff support them). This was similar to North England hostel 2, where a weekly visit to the hostel was made by a local health outreach team; a nurse visited weekly (for 3 hours) and a nurse prescriber visited weekly for a full day. Although not part of any specialist service or new NHS configuration, the health-care practice local to South London hostel 2 held a weekly clinic (full day) in the hostel; a nurse visited weekly (for a full day) and a further specialist nurse prescriber made a weekly visit (for a full day). A similar pattern operated in North England hostel 4, where the local GP made a weekly visit, and a community nurse was described as coming in when needed. Similarly, in North England hostel 1, a local general practice held a clinic every week (lasting 3 hours) in the hostel; a nurse from a well-being team visited weekly (3 hours) to undertake hepatitis injections and apply dressings, although this seemed to be a new specialist initiative as she was not attached to a general practice.

Mental health team input

Hostels could be busy places once primary care input and other services were viewed together. In Central London hostel 1, a member of the CMHT visited weekly, and an alcohol worker visited once or sometimes twice per week. This hostel was unique in our sample in employing its own part-time drug and alcohol worker. In sharp contrast, Central London hostel 2 reported that its local CMHT was reluctant to work with the hostel because of residents’ drug and alcohol problems. In its stead, a local mental health charity had been visiting regularly and had accepted self-referrals, and another voluntary sector service had formerly provided drop-in mental health support. Both had recently ceased at the time of our study. North England hostel 1 reported that it could not make direct referrals to the CMHT; these had to be made by a GP or by the local authority adult services department (social services). North England hostel 4 reported no contact with mental health services, whereas North England hostel 3 reported that a CMHT worker visited weekly or monthly, depending on the worker; its residents had often been referred to the CMHT following their hospital A&E visits. In this hostel, a drugs or alcohol worker used to make weekly visits to the hostel but this had recently ceased following funding cuts.

Further different patterns of engagement with mental health services were evident in South London hostel 1, where some residents were ‘live’ cases with the local CMHT, although sobriety was required before the CMHT would engage with a resident; this was described, optimistically, as offering opportunities for residents to give up or reduce alcohol consumption. More unusually, a psychologist was available for consultations every weekday in South London hostel 2, where the local intermediate care (rehabilitation) project staff were also available daily. In this hostel, an alcohol worker ran a weekly alcohol support group, and a drugs worker visited the hostel weekly to carry out one-to-one work with some residents.

Background demographics

Forty-three hostel key workers were interviewed for this study (some did not supply full demographic details). Table 6 presents background information about them. All but two were employed full-time and worked shifts, each shift typically lasting 6 hours. Shifts changed weekly and the hours of work were not always consistent. Their job titles ranged from project worker and support worker (SW) to lead case worker, all of which incorporated many elements of the same job descriptions. Six were assistant managers, a role that SWs undertook alongside their resident caseload as key workers. In some cases, the SW was identified as having a specialist role, such as ‘substance misuse worker’ or ‘lead worker in medication and health and support’, the duties of which were in addition to their caseload.

The length of time participants had worked in their current hostel ranged from 1 month to 26 years. Twelve participants had worked in their current hostel for < 12 months, seven had worked there between 1 and 2 years and 12 had worked there for > 5 years.

Current roles and experiences

Descriptions of their roles essentially covered the same functions. Most described having a caseload of six or seven clients, for whom they provided a range of support, including supporting them with day-to-day tasks, such as keeping themselves clean, tidying rooms and laundry assistance; managing money and budgeting; managing their health in terms of appointments, medication and attending to chronic problems; identifying and helping them achieve manageable goals; drawing up support plans; undertaking risk assessments; and checking on each client every day or as frequently as necessary. Such a role was often described as being the resident’s key worker.

Previous roles

Eighteen SWs had not previously worked with homeless people before their current job. Their prior work experience varied; one had been a restaurant manager, another a hotel manager, one had managed laboratories, one had been a personal assistant in a technology company, one had worked in retail, two trained in nursing and one had previously worked as a designer/photographer. Three participants had started as apprentices in the homeless sector, and one as a housing officer in a local authority. The remaining participants had started work as project workers and SWs in hostels, shelters and day centres in the homelessness sector.

Training undertaken and training needs

Hostel staff described undertaking a wide range of training. Some training was described as mandatory, but several voiced enthusiasm for signing up to training as a way of developing their skills. Most appeared to be in-house training, being organised either by individual hostels themselves or by an overarching hostel provider in the not-for-profit sector. This was the case in Central London hostel 1, which ran a comprehensive mandatory training programme for all staff. This covered drug and alcohol use, key working and ‘emotional intelligence’. For any staff commencing apprenticeships, there was an opportunity to take a qualification [National Vocational Qualification (NVQ) level 2] in health and social care. There was also one-off training available on ‘care of the elderly’ and dementia, and ‘end-of-life’ training. In contrast, another hostel put great reliance on being able to access local authority training covering safeguarding, managing dual diagnoses and managing complex needs.

Across the hostels, staff reported recently accessing training in adult safeguarding, working with vulnerable people, first aid, fire training, equality and diversity, risk assessment, mental health, alcohol and substance use, emotional care, health and safety, rehabilitation, debt and benefits advice, dealing with challenging behaviour (their phrase), dealing with difficult families, working with complex clients, administering medication and caseload management. Training specific to this study’s focus included working with older adults, mental health, dementia training, hoarding, managing ‘challenging behaviour’, capacity and decision-making, and the Mental Capacity Act 2005. 90

Although most participants felt that the training they required was available and accessible, when probed several identified specific areas in which more training might be helpful. Two talked about the value of more dementia-specific training, especially identifying early symptoms and supporting people with ‘everyday’ memory problems. One described needing more refresher courses rather than only day-long training courses. A senior member of staff in South London hostel 1 commented that training was not always available to meet the changing needs of their hostel resident profiles:

Dementia training is very hard to access and very expensive. We can’t always get everyone on the course, nor is it easy or cheap to get someone to come here and deliver training. So we send someone with the understanding that they will come back and disseminate the information to the rest of us. Which is not the best way to do it! I would really like us all to have some training in managing behaviours in dementia. There are so many challenging behaviours we come across, that we can never know whether it’s dementia, or their alcohol or their personality and how to deal with it. I saw this lovely clip about someone with dementia presenting with aggressive behaviour and how staff dealt with it, and it was fascinating! Wish we had access to more training like that.

Senior member of staff, South London hostel 1

There was some indication of training being carried out online as e-learning. In one North England hostel, staff accessed some courses online. In this hostel, there was particular emphasis on addictions and mental health; its managers undertook diploma courses, and other staff undertook a 2-day mental health ‘first aid’ course covering suicide intervention skills training, and some had completed a foundation unit in men’s health to help them identify early mental health problems and self-harm and how to guide someone to the right services. However, no specialist training was accessed here that covered memory problems.

Across the hostels, 16 staff participants had obtained NVQs levels 2 and 3. The majority of these were in health and social care (nine participants), followed by mental health (three), vulnerable adults (two), promoting independence (one) and counselling (one). These had been achieved either during the course of their employment at the hostel or prior to their current role. Overall, there seemed low expectations of regular training except in the hostels that had their own in-house provision. High staff turnover clearly affected the extent to which more experienced staff could support new colleagues; this affected one North England hostel in particular, which relied more than others on what it called ‘on-the-job learning’.

Age, sex and ethnicity

Among the 62 older homeless people, 24 were staying in hostels in Central London when they were first interviewed, 19 were staying in hostels in South London and 19 were staying in hostels in North England. The majority (n = 54) were men; just eight were women. One-third (34%) were aged 50–54 years, 29% were aged 55–59 years, 21% were aged 60–64 years, 8% were aged 65–69 years and 8% were aged ≥ 70 years. Their ages ranged from 50 to 82 years, and the median age was 56.5 years. Hence, the sample was relatively young compared with older people generally, but was comparable with the age distribution of ‘older’ homeless people in Western countries. According to statistics from England, Canada and the USA, the majority of older homeless people are aged in their fifties; just 5–10% of homeless adults are in their sixties or older. 16 There was no significant difference in age according to the cluster site of their hostel.

Most residents were white British (76%) or white Irish (8%), and 16% were from other ethnic groups. There were, however, differences among the cluster sites. All in the North England sample were white British, and most originated from the local area. By comparison, slightly fewer (71%) in the Central London cluster and just 58% in the South London one were white British. The findings are not surprising given the ethnic mix of the general population in London.

Education and employment histories

Just under half (45%) of residents said that they had an educational or vocational qualification, mainly GCSEs (General Certificates of Secondary Education), NVQs or City & Guilds certificates. Those in the North England cluster were most likely to have a qualification (68%); and those aged 50–54 years were least likely to have a qualification (29%). More than one-third of participants (37%) reported literacy difficulties, including some who had dyslexia. Reports of literacy problems were less common among the North England cluster (26%) and most common among those aged 50–54 years (48%). Several participants who reported literacy problems described the struggles that they experienced with reading and writing, including some who had left school at an early age:

I’m a bit dyslexic – have problems with writing and spelling, but not reading. I had a misspent youth. I should have gone to school but I got picked on a lot. So I’d go into school, get a mark, and then leave again.

Can’t read or write – I struggle to read a sentence. Left school aged 13.

Difficulty reading. They kicked me out of school.

The employment histories of the participants varied greatly. Three-fifths (61%) said that they had been mostly employed during their adult lives, one-quarter (25%) had been ‘in and out’ of work, and 14% had been mostly unemployed. The last group included one man who had never been in employment. Among the 59 residents who were able to provide more detailed information, 31% had worked until they were aged in their fifties or older, and 29% had been unemployed since aged in their twenties or thirties. Participants’ types of employment also varied greatly. Using the Office for National Statistics’ Standard Occupational Classification (2010),102 most residents (46%) had either been in skilled trades, such as painters and decorators, carpenters, chefs or similar, or they had been in elementary occupations (34%), as security guards, cleaners, postal workers, shelf fillers or similar (Table 7). Just one person had employment at the time of their interview.

The great variation in their employment histories is apparent from their accounts (some roles have been slightly modified to maintain anonymity):

• Worked for the civil service for 40 years, until the age of 57 years.

• Worked as a delivery man for 10 years from the age of 15 years, and then as a street cleaner for a few years. Not worked since the age of 40 years.

• Worked as a manual labourer for 16 years, then as a machine operative for 20 years, and then as a semiskilled engineer for 10 years. Retired aged 65 years.

• Worked in a supermarket filling shelves from 16 to 26 years. Not worked since.

Age first homeless

The age at which residents first became homeless varied greatly. According to 54 people who were able to give details, almost one-quarter (24%) had first become homeless as teenagers or when they were aged in their twenties. By comparison, 22% had first become homeless when they were aged in their fifties and 15% were aged ≥ 60 years at the time. As shown in Table 8, more than one-third (37%) of those aged in their early fifties had first become homeless before the age of 30 years. By contrast, most of those aged ≥ 60 years had experienced late-onset homelessness; 70% had first become homeless after the age of 50 years.

The age at which residents first became homeless differed by cluster site (Figure 2). Almost three-fifths (59%) of those in North England first became homeless after the age of 50 years. This compares with just 39% in the South London hostels and 25% in the Central London cluster. Many in the two London sites had become homeless when they were in their thirties or forties.

FIGURE 2.

Participants’ age when they first become homeless, by cluster site.

Length of time homeless

It is very difficult to collect accurate histories from homeless people about the length of time that they had been homeless, as ‘homelessness’ is often not a continuous state. Many frequently move between homelessness and insecure housing. For our study, collecting this information was more problematic given that some residents had memory problems, and several had moved between homelessness and insecure housing for many years. Data about the length of time homeless are drawn from residents’ accounts and those of hostel staff, and from hostel records, when available. In some cases, the data refer to the minimum time spent homeless, as hostel records had information only about the length of time that residents had been using hostels run by that particular organisation. Therefore, the findings have to be interpreted cautiously.

Residents’ length of time as homeless varied, but many had been living on the streets or in hostels continuously or intermittently for several years. Among 57 people for whom some details were available, 7% had been homeless for < 1 year, and nearly half (47%) had been homeless for > 10 years, including 21% who had been homeless for > 20 years. There were differences by cluster site. Almost one-fifth (19%) in the North England site had been homeless for < 1 year, compared with just 5% in South London and none in Central London (Table 9). The corresponding figures for those who had been homeless for > 10 years were 50%, 58% and 36%.

Duration of stay in current hostel

There were variations in the length of time that residents had been staying in their current hostel. Nearly half (48%) had been in the hostel for ≤ 12 months, and nearly one-quarter (23%) for > 5 years. Those in North England tended to have been in their hostel longer; nearly half of them (47%) had been so housed for > 5 years (see Table 9). By contrast, those in the two London clusters were more likely to have been so accommodated for ≤ 12 months. These differences may partly reflect the contractual arrangements imposed on homelessness organisations by local authorities rather than the needs of homeless people. Since 2010, there have been several changes to housing and welfare policies and practice in England, and the length of stay in hostels imposed by some local authority contracts has reduced substantially. Some hostels now have maximum durations of ≤ 6 months, and people are moved from one hostel to another. 103

Physical health problems

Nearly all residents (97%) had one or more physical health problems. As shown in Table 10, the most common problems were musculoskeletal conditions, followed by respiratory problems, eye conditions, hypertension and gastrointestinal problems. The reported figures are the minimum percentages of residents with each problem, as some people may have had conditions that they did not report, and data from their medical records were not always obtained or obtainable.

Head injuries

Three-fifths (61%) of residents reported having had a head injury, including 11% who described having had three or more such injuries. In some cases, the head injuries dated back years, but for others the injuries had occurred in recent years. Most were the result of accidents or assaults and were often linked to heavy drinking and fighting. A small percentage of residents (7%) had sustained a head injury during an epileptic seizure. The situations leading to head injuries were described by two men:

I used to be a very heavy drinker and drug taker. I’d regularly fight and fall about. One time I had an axe on my head because of a fight.

A guy knocked me out – I was away with the fairies . . . [another time] I keeled over and knocked my head when I was pissed.

Most residents who reported a head injury (88%) had received hospital treatment on at least one occasion, including 68% who were admitted into hospital. One man aged in his fifties, for example, first sustained a head injury as a teenager when he was playing rugby. He described being ‘knocked out’ but was not taken to hospital. As an adult, he worked on building sites and in factories, and said that he had suffered a head injury on two occasions when items fell on him at work. The first time he was in his early thirties and was taken to hospital but ‘sent home to rest’. The second injury occurred a few years later and he was kept in hospital for 4 days. He sustained a fourth head injury about 7–9 years ago during an epileptic seizure, and was taken to hospital but discharged the next day. He had subsequently experienced further head injuries during epileptic seizures.

Mental health problems

Nearly two-thirds (63%) of residents reported a history of mental health problems, and 44% described current mental health problems. These percentages do not include memory problems, which are described in Memory problems. When information from staff and from medical and hostel records is also considered, a higher prevalence of mental health problems was apparent: 81% of residents had a history of mental health problems, and 74% had current mental health problems.

The main problem both identified by residents and recorded by GPs was depression, sometimes accompanied by anxiety. Almost seven-tenths (69%) of residents were reported to have suffered from depression, including five people who had attempted suicide in the past. According to the medical records, two residents had been diagnosed with personality disorder, and six with schizophrenia or ‘alcoholic psychosis’ or ‘alcohol-induced hallucinations’. One-fifth (22%) of residents described having first experienced mental health problems when they were teenagers or young adults. For some, the problems had continued over many years. Several associated the onset of mental health problems with traumas and stressful events that had taken place before they became homeless, such as marital or family breakdown, the death of their mother or a child, or the loss of a job:

[Mental health problems] started when I was aged 14 when I was badly beaten up at school. Was in hospital 7 months. Was bullied at school.

It was when things were all going wrong for me – marital problems, work problems. I was a heroin addict at the time as well.

So much depression – I can’t forget my family . . . they were killed. When I sleep I see my mother and brothers and gran dead. I have no future.

My parents divorced when I was 15, and I was diagnosed with anxiety and depression.

Drawing on scores from the depression tool administered during the baseline interview (described in Chapter 3), although 54.1% of residents were assessed as having no depression, 6.6% had mild depression, 26.2% had moderate depression and 13.1% had severe depression. There were age differences. Those aged in their fifties were more likely to have moderate or severe depression than those above this age (affecting 42.9% aged 50–54 years, 58.8% aged 55–59 years, 30.8% aged 60–64 years and just 10% aged ≥ 65 years). Residents living in hostels in Central London were more likely to have moderate or severe depression (52.2%) than those living in hostels in South London or North England (31.6% for both areas).

Residents with moderate or severe depression were more likely than those with mild or no depression to rate their general health as ‘bad’ or ‘very bad’ (52.1% and 13.9%, respectively), and their QoL as ‘poor’ or ‘very poor’ (34.8% and 10.8%, respectively). There were also associations between depression and alcohol and drug use. Those who drank alcohol daily or almost daily (6 days per week) were more likely to be moderately or severely depressed than those who drank less frequently or not at all (47.2% compared with 28%, respectively). Likewise, 64.7% who had used illegal drugs in the preceding 3 months were assessed as having moderate or severe depression, compared with 30.3% who had not used illegal drugs during this period [χ² = 6.03; degrees of freedom (df) = 1; p = 0.014]. Further details of their drinking habits and use of drugs are provided in the following sections.

Alcohol use

Three-quarters (76%) of residents said that they had been a heavy drinker at some time or had had alcohol problems. This increased to 85% when reports from staff and information from medical records were included. Twenty-six residents (42%) said that they had started to drink heavily before the age of 20 years. Among those who had been heavy drinkers, 26% said that one of their parents had also drunk heavily, and an additional 10% said that both parents had been heavy drinkers. A history of heavy drinking was common among all age groups, but particularly prevalent among those aged 50–54 years (95% of this age group). It was also more common in the two London cluster sites, affecting 91% of participants in London, compared with 74% in North England.

When current drinking habits are examined, 13% of residents said that they did not drink alcohol and this was confirmed in hostel staff accounts. Almost three-fifths (58%), however, drank alcohol on 6 or 7 days per week (Table 11). Some had a few days of heavy drinking when they received their welfare benefits, followed by periods of light drinking or abstinence when they had no money. Although relatively few people drank spirits or wines, a high proportion (57%) regularly had super-strong lager (such as Tennent’s, which is 9% alcohol by volume) or extra-strong beers or ciders, such as White Ace (7.5% alcohol by volume). Most people (83%) who drank alcohol daily tended to consume extra-strong beers and lagers.

It is difficult to calculate the number of units of alcohol the residents consumed each week because their drinking patterns fluctuated. Some drank spirits when they had money but consumed cheap, extra-strong lager or beer at other times. Hostel staff were also unable to say how much alcohol some residents consumed as they tended to drink either in their bedroom or away from the hostel. The average number of units of alcohol consumed by residents per week has been calculated as accurately as possible from the information they provided. Among 52 residents for whom information is available, 63.5% were regularly drinking in excess of the Department of Health and Social Care recommended weekly guidelines of no more than 14 units per week. Moreover, 48% were drinking > 50 units each week, including 25% who were drinking, on average, > 100 units per week.

Those aged in their early fifties were more likely than other age groups to drink excessively. Among those aged 50–54 years, 71% had alcohol most days and 81% consumed extra-strong beers and lagers (Table 12). They were also the age group most likely to be far exceeding the Department of Health and Social Care recommended weekly guidelines for alcohol consumption; 71% were consuming > 50 units of alcohol per week (see Table 12).

There were differences in drinking habits by cluster site. Residents in the North England cluster were less likely than those in the two London clusters to consume alcohol most days (Table 13). The former were significantly less likely to drink strong beers or lagers (21%, compared with 71% in Central London and 78% in South London; χ² = 15.1; df = 2; p = 0.001). They were also less likely to exceed the Department of Health and Social Care recommended guidelines of no more than 14 units of alcohol per week (χ² = 6.7; df = 2; p = 0.034). Residents in the Central London cluster were most likely to be drinking excessively, with nearly half (48%) drinking, on average, > 100 units each week (Table 14).

Among residents whose weekly alcohol intake was > 14 units, three-quarters (76%) acknowledged that they had felt the need to cut down on their drinking, half (50%) had felt guilty about their drinking and most (84%) said that they had sometimes needed a drink first thing in the morning to steady their nerves or get rid of a hangover. There was no significant association between drinking excessively each week and current mental health problems. As described in the following accounts, however, a few participants believed that there was a link between their mental health problems and heavy drinking:

Something inside me doesn’t want to go on . . . that’s why I’m binging [drinking] more. I have made threats to kill myself. But when I’m sober I don’t want to kill myself.

My depression can’t stop. I’ve asked my GP for treatment for depression but he said I have to stop drinking first.

Use of illegal drugs

Just over half (55%) of residents said that they had used illegal drugs at some time, although many fewer (16%) reported having taken drugs during the previous 3 months. These findings increase to 57% for past drug use and 29% for drug use in the previous 3 months when reports from hostel staff are also included. Most residents who had used drugs in the previous 3 months reported having taken cannabis or crack-cocaine. Just two people said that they had taken heroin. In most cases, they had started taking drugs when they were in their teens or early twenties.

Three-quarters (76%) of residents aged in their early fifties reported a history of illegal drug use, and just over half (52%) said that they had taken drugs in the previous 3 months (Table 15). This age group was most likely to report this behaviour. There were also differences by cluster site. Two-thirds (67%) of Central London participants had used illegal drugs in the past, and 37.5% had taken drugs in the previous 3 months. Slightly lower figures were reported for the South London participants (57% and 28%, respectively), and even lower figures were reported for the North England cluster (47% and 21%, respectively).

Reports of memory problems by hostel residents

Among the 62 residents, more than half (54.8%) reported having memory difficulties, 25 (40.3%) reported no problems, and three were unsure. As shown in Table 16, those aged in their early fifties and early sixties were most likely to describe memory problems, and older residents aged ≥ 65 years were least likely to report a problem. This may be a surprising finding in the light of the strong association of such problems with advancing age; however, caution has to be taken when interpreting these findings because of small numbers.

When the 34 residents who reported memory problems were asked how they were affected, the most common responses were:

• unable to remember appointments and other tasks (41.2%)

• unable to remember names (20.6%)

• unable to remember the day and date (17.6%)

• unable to remember yesterday’s activities (14.7%)

• loses things (14.7%).

When the residents who reported memory problems were asked how long they had experienced such difficulties, 21.2% said < 1 year, 27.3% said between 1 and 5 years, 12.1% said > 5 years, and 39.4% were unable to say.

The accounts from three residents illustrate the ways in which they felt they were being affected by their poor memory:

I can go to the shops and forget what I’ve gone for and forgotten to take my money with me . . . small things like that.

I mislay things all the time. When I have a conversation about sport I know the answer but it’s slow coming to me. I used to be able to answer immediately but now my brain is not ticking over as much.

It’s [memory loss] annoying. It will stop me talking to someone as I can’t recall their name and it’s embarrassing.

Reports of memory problems by hostel staff

Hostel staff were asked whether or not they believed that certain residents had memory problems and, if so, the reasons why. Information was obtained about 60 residents from their key workers (some key workers were responsible for more than one study participant). As shown in Table 17, hostel staff believed that nearly half (46.7%) of residents had a memory problem and that 40% of residents did not have a memory problem, and they were unsure about the remainder (13.3%). Staff in Central London were less certain about their residents than were staff in the other two cluster sites.

Members of staff gave many reasons as to why they suspected or were aware that the residents had memory problems. For example, among the residents in whom memory problems were suspected, staff said that 77.8% had difficulty remembering recent happenings and appointments, 74.1% had difficulty recalling conversations a few days later, and 37% had difficulty finding the correct words. A selection of staff’s accounts illustrate this further:

He repeats himself a lot; asks everyone the same questions all the time.

He can’t remember anything. Quite severe. He can’t remember conversations we had 2 minutes ago. Tell him what he needs from the shop, ask him to repeat it 2 minutes later, and he can’t.

He doesn’t remember some things, like the other day he went out of the hostel. He could not remember this and thought he was here in the hostel all day.

Hostel staff explained that it was difficult to determine the extent to which alcohol was responsible for memory difficulties among some residents. As two workers summarised:

He often forgets what he needs to do but that could be due to alcohol. He would often ask what we had spoken about . . . but he knew when meal times were.

I give her all letters and reminders . . . is it early dementia or alcohol?

According to staff, six residents had been diagnosed as having Korsakoff syndrome, vascular dementia or alcohol-related dementia. A further six had been referred by their GP to the local memory clinic or to other services for an assessment of their memory. For some, assessments were in progress, and one resident had been initially assessed at a memory clinic but could not be assessed further because he was still drinking heavily. Another resident with memory problems was being monitored by his GP and, according to the staff, was to be referred for specialist help if his memory difficulties worsened.

Making a diagnosis of cognitive decline/dementia

The study aimed to identify individuals with memory problems. To be able to assess an individual in order to make a diagnosis of dementia or related memory problems, clinicians across different services in England follow a fairly similar care pathway. It usually starts with an individual or their family reporting concerns about the person’s memory to the person’s GP. After basic blood tests and initial memory assessments are carried out in primary care, the patient may be referred to secondary care services for further assessment and diagnosis. The majority of memory assessments in secondary care in England are carried out by CMHTs or at memory clinics for older adults; some of those assessments are carried out by geriatricians, neurologists and clinicians working in private centres. After a person is referred to a secondary care service, they will be seen in an outpatient department, or in their home, with the aim of gathering detailed information about their difficulties, and also collateral information will be gathered from their relatives, friends or carers.

The majority of people who are seen in memory services will be offered a type of brain imaging (computerised tomography scan or magnetic resonance imaging) to exclude a brain-occupying lesion or another reason for their cognitive presentation. Brain imaging data are helpful to confirm some dementia syndromes. Only a small percentage of those people require further imaging, such as a PET (positron emission tomography) or DaT (dopamine transporter) scan. This would usually be requested by a specialist. Almost all of those who attend a memory clinic will have memory assessment tests, and, if required, they will be referred to a neuropsychologist for further detailed memory and cognitive assessment. A simple memory assessment takes between 5 and 20 minutes to complete and a detailed assessment takes up to 3 hours. After the results of imaging and memory assessments are reviewed by clinicians, individuals with memory problems will be seen again to receive their results and possibly be given their diagnosis during the same session. The entire process may take up to 6 months to complete.

Formal assessments

In this study, it was not possible to follow the routine care pathway of the NHS to identify people with memory and cognitive difficulties; therefore, we tried to gather as much information as possible during visits and interviews with individuals and the staff from hostels. We used ACE-III as a memory test. 93 The tests were undertaken by the research team’s consultant psychiatrist, psychiatric registrar and psychologist.

The ACE-III test has been widely used in memory clinics to assist clinicians to assess individuals’ cognition. It has been validated, and the maximum score is 100 points. The cut-off point of 88 is 100% sensitive and 96% specific. There is also another cut-off point of 82 out of 100; in this case it is 93% sensitive and 100% specific for dementia. This test usually takes 20 minutes to be completed and requires a rater who has been trained to complete this memory assessment. 105 This scale includes five subdomains, with a total maximum score of 100 points. The five domains are visuospatial (out of 16 points), attention (out of 18 points), memory (out of 26 points), fluency (out of 14 points) and language (out of 26 points).

Cognitive scores from Addenbrooke’s Cognitive Examination III

Of the 62 residents interviewed at baseline, the ACE-III was completed with 42. Among those who did not complete the ACE-III, five left their hostel before it could be administered, one resident died, another could not complete it because of language difficulties, and another could not complete it because of literacy difficulties and therefore the MoCA96 was administered to this person. The remaining 12 residents either refused to take part or were not in the hostel when the psychiatrist visited (despite several visits).

Of the 42 residents who completed the ACE-III, total scores ranged from 29 to 89 points. Just 28.6% scored ≥ 80 points, and 30.9% scored < 60 points. The mean of their total ACE-III scores was 66.81 points, with a SD of 16.64 points. As shown in Table 18, there were differences by age group, although the findings must be treated cautiously because of small numbers. Half of those aged in their early fifties scored between 60 and 79 points, just over one-third scored ≥ 80 points and just 14.3% scored 40–59 points. None in this age group scored < 40 points. By contrast, half of residents aged 55–59 scored ≤ 59 points, including 10% who scored < 40 points. Those aged in their early sixties had the highest scores, and those aged ≥ 65 years had the lowest scores.

Classifying whether or not residents had memory problems

Owing to the small numbers of hostel residents who were interviewed at baseline, it was agreed that all who were willing and available would be included in the ‘case study’ element of the study. The next step was to classify residents according to whether or not they had memory problems. Several team meetings reviewed all of the data with the psychiatrist, and participants were allocated to groups according to whether or not there was enough information about them to make a judgement.

Based on all of the information, residents were assigned to one of three groups: ‘memory problems’, ‘borderline’ and ‘no memory problems’. We decided to use these three groups based on the data we had after reviewing each participant’s study profile, such as age, physical health problems, weekly alcohol intake and use of illegal drugs, reports and diagnoses of memory problems, and the duration of this complaint. We used these data, alongside their score on the ACE-III, to identify participants with memory problems. All data relating to individuals were reviewed by the consultant psychiatrist and his team to make sure that people were accurately placed in a subgroup. We also examined the subdomains of ACE-III scores to help us identify individuals with memory problems. We categorised each subdomain into three parts to help us analyse the data relating to individuals, before placing individuals in one of the three categories related to memory problems. Although the subcategory score of each subdomain is not validated and not tested, it has been used widely by clinicians to make decisions about an individual’s memory problems based on their scores on subcategories.

We explained earlier the process and the time needed to make an accurate diagnosis of dementia through the NHS (see Making a diagnosis of cognitive decline/dementia). We had to adapt the system to improve the accuracy of our diagnostic criteria and it was impossible to rely on only the cut-off point of ACE-III as an inclusion or exclusion criterion for memory problems. Other factors, as explained previously, played an important role in assessing this group.

Many of the residents reported having had one or several head injuries in the past. We were not in a position to request any brain imaging for this group, and therefore we relied on the data that we had to make a decision about the cognition of these participants. One of the major problems we faced when assessing cognition was that many of these participants were under the influence of alcohol and so it was not possible to conduct a memory assessment at that point. We had to repeat some of the visits to be able to complete those assessments. Such efforts are not generally undertaken in routine NHS services in our experience.

Cognitive groups and their scoring

Based on all of the information, 23 out of the 47 case study participants were allocated to the ‘memory problems’ group, seven were allocated to the ‘borderline’ group and 17 were allocated to the ‘no memory problems’ group. However, only 42 residents completed the ACE-III, of whom 14 were allocated to the ‘no memory problems’ group (mean of their ACE-III score 83.79 points, SD 4.09 points), eight were in the ‘borderline’ group (mean 71 points, SD 6.78 points) and 20 were in the ‘memory problems’ group (mean 53.25 points, SD 12.49 points) (Table 19). Their scores for the subdomains are in Table 20. One of the residents in the borderline group was subsequently evicted from the hostel and therefore could not be included as a case study.

Hostel staff’s reports of memory problems compared with formal assessments

One of the study’s objectives was to determine the extent to which hostel staff were aware of memory problems among their older hostel residents. As shown in Table 21, hostel staff identified memory problems for most residents who were formally assessed as having memory problems; out of 23 residents with memory problems, hostel staff identified memory problems in 70%. They were unsure for two residents, and did not identify memory problems in just five residents (22%) in whom there were memory problems in our judgement. At the same time, among 17 residents who did not have memory problems, the staff reported this correctly in 59%. Among the five residents whom the hostel staff suspected had memory problems but were found not to have memory problems by the psychiatrist, one was undergoing tests at a memory clinic, and another was a former heavy drinker and, according to the hostel staff, ‘may be a little bit degenerative’.

Residents’ ratings of quality of life

Residents were asked at baseline to rate their QoL; they could respond ‘very good’, ‘good’, ‘neither good nor poor’, ‘poor’ or ‘very poor’. Almost half rated their QoL as ‘very good’ or ‘good’, 31.1% described it as neither good nor poor, and 20% described it as ‘poor’ or ‘very poor’ (Table 24). There were slight differences, however, between those with and those without memory problems (see Table 24). Fewer residents with memory problems were likely to say that their QoL was ‘very good’, although a relatively high percentage described it as ‘good’. By contrast, a high proportion without memory problems rated their QoL as ‘neither good nor poor’.

There were differences by age groups in ratings of QoL, although the figures must be treated cautiously because of the small numbers in each age group. Those aged ≥ 65 years were most optimistic and four-fifths described their life as ‘good’ or ‘very good’ (Figure 5). Residents aged 50–59 years were most likely to describe their QoL as ‘neither good nor poor’. Those who were depressed were less likely than those who were not depressed to rate their QoL in positive terms. Among those with depression, 27.8% described their QoL as ‘good’ or ‘very good’ and 33% described it as ‘poor’ or ‘very poor’. The corresponding figures for those without depression were 62.9% and 11.1%. There were also marked differences in relation to whether or not they were drinking heavily. As shown in Figure 6, those who were drinking very heavily (≥ 50 units of alcohol per week) were less likely to be optimistic about their QoL.

FIGURE 5.

Residents’ ratings of QoL, by age group.

FIGURE 6.

Residents’ ratings of QoL, by weekly alcohol consumption.

Factors contributing to quality of life

When asked what factors contributed to their QoL, most residents described one or more factor. However, 19% were unable to answer or could not identify anything that added quality to their life. This included 17.6% of those without memory problems and 20.1% with memory problems (or who were borderline). There were large differences between those with and those without memory problems in relation to their responses about the factors that gave quality to their life (Table 25). The residents without memory problems were most likely to identify being healthy and engaging in various activities, such as reading, watching television and going for walks, as important factors in enhancing their QoL. By contrast, those with memory problems were most likely to say that socialising with other hostel residents and drinking alcohol were the most important factors.

The following accounts are examples of activities that residents said added to their QoL:

Able to get out and about. Chat with people. Make a fuss of people’s dogs. Chat to shopkeepers and people on the streets.

‘Borderline’ group

I love music. I used to write songs and lyrics and I used to be in a band. I like being creative and doing paintings. I’ve done several courses at college.

‘No memory problems’ group

Don’t know. Going to the pub.

‘Memory problems’ group

Factors that would improve quality of life

When the residents were asked what factors would improve their QoL, there were again differences depending on whether or not they had memory problems. The most commonly reported factor by far, in both groups, was being rehoused and moving out of the hostel (Table 26). Overall, 34.9% of residents identified this, including 58.8% of those without memory problems. Those residents without memory problems also mentioned having more money, getting a job and having better health as important factors. Those with memory problems identified having more money, having better health and stopping drinking or using drugs as important factors. Only a small proportion in both groups (4.7% overall) said that more contact with family and relatives would improve their QoL.

Sample for the economic analysis

The analysis of costs was based on complete cases. Data on service use were available for 47 (75.8%) out of the 62 study participants.

Frequency of service use

Data were recorded for a large number of services, many of which were used by small proportions of the participants (see Appendix 2). GP services were the most frequently accessed. Non-attendance at GP appointments was > 15%, and it was 44% for appointments with allied health professionals (see Appendix 3). Table 28 presents frequency of service use data.

When services were grouped (Table 29), 94% of participants had appointments for contact with a GP during the 6-month period, 60% had appointments with other general practice staff, 68% had a hospital outpatient appointment, 47% had used emergency or out-of-hours services, 38% had used alcohol and drug services and 32% had had contact with any psychiatric or secondary mental health service (such as a member of a CMHT).

Service use costs

The cost for each service for each participant was calculated by applying 2014/15 unit costs (see Appendix 2) to utilisation data, including for appointments that were not attended. Summary statistics for the service groupings are shown in Table 30. All participants incurred a cost from at least one service group. The mean (median) cost during the 6-month period was £2975 (£14,550), of which £128 (£0) was for psychiatric services; however, there is considerable variability between participants, as shown by the SDs and ranges. The most costly grouping (means of > £700 for the 6 months) were social care services, owing to 12 participants having a care worker making frequent visits to the hostel to provide personal care and support to specific residents (five participants had daily visits and seven had several visits per week), and in-hospital stays (one day case, one psychiatric/mental health stay, and nine non-psychiatric stays including one participant having two hospital episodes). The next most costly items were alcohol and drug services (mean £443) and GPs (mean £324).

Associations between memory problems and service use costs

When all services were considered individually, there was only one significant difference in costs between participants in whom memory problems were present, participants who were borderline and participants in whom memory problems were not present. This was for the visiting home care workers, for which all costs were incurred by participants with memory problems either present or borderline (data not shown). When comparisons were conducted for groups of services, the costs for people without memory problems were significantly higher than those for people in whom memory problems were present or borderline for other GP services (i.e. nurses, care navigators in general practices), but not for any other service grouping (see Appendix 4). The difference in service patterns between memory problem groups should not be interpreted as identifying specific impact of memory problems, as the three groups of homeless people also differ in a number of other characteristics.

There were 41 participants who had undergone cognitive testing with the ACE-III (range 0–100 points, with lower scores indicating worse cognition). The mean ACE-III score was 65.76 points (SD 16.85, median 66, interquartile range 29–81.5 points). There were significant correlations between the ACE-III scores and costs of three groups of services. Higher costs of GP services and emergency and out-of-hours services were associated with lower scores; higher costs of other GP services (nurses, care navigators) were associated with higher scores. There was no significant association between all GP services (GPs and other practice staff together) or any other cost grouping, or total costs, and cognition as measured by ACE-III scores (see Table 30).

Associations between participant characteristics and costs

Tests of association between participant characteristics revealed significant (positive) correlations between number of named medications and number of comorbidities and number of head injuries reported by participants. Illegal or illicit drug use in the previous 3 months was also associated with more head injuries. Location in London, rather than North England, was associated with non-white British ethnicity and more frequent alcohol consumption (see Appendix 5).

Associations between costs and other participant characteristics (sociodemographic and medical) revealed that higher total costs were associated with location in London, and having more comorbidities and more prescribed medications; higher psychiatric costs were associated with more head injuries. Higher costs of all GP services were associated with being in London, more alcohol use, and having more comorbidities, prescribed medications and head injuries. Hospital stays were associated with more comorbidities and prescribed medications. Out-of-hours costs were associated with number of medications and number of head injuries (Table 31).

Backward stepwise regression modelling was used to explore the independent predictors of each category of costs. Characteristics describing participants were entered into the models and the least significant were removed one by one (if a p-value > 0.1). Sex (gender) was excluded from the analysis because of an underlying relationship with location (there were only five women in the sample, four of whom were in the south of England). Final models are shown in Table 32. The number of medications was an important predictor of several cost categories. Memory problems were a significant factor in predicting emergency and out-of-hours care, but not any other cost category. Over one-third (37.5%) of the total service use cost (which does not include the cost of prescribed medications) was explained by number of comorbidities and number of prescribed medications. Each extra prescribed medication that a participant reported added £309 to the 6-month service use costs; each additional comorbidity added £393.

Challenges in identifying memory problems and getting a diagnosis

As presented in Chapter 5, there were discrepancies between the reports of residents about their memory problems and hostel staff’s perception of these. Some residents were aware of their memory problems; for example, participant 110 described their problems thus: ‘I put it down to my head getting old’. Participant 122 acknowledged:

Some days I wake up and I don’t know what day it is when I’ve been on a binge. So I have to go and buy a newspaper to find out what day it is. It’s when I’ve had too much booze.

Qualitative responses may shed some light on these discrepancies. Residents described a number of everyday tasks for which they received help from hostel staff, including budgeting, arranging appointments, prompting, and assisting with hygiene and self-care tasks. Some of this assistance may have supported hostel residents who, if living independently, would have otherwise struggled to cope.

‘Prompting’ was often cited by many hostel staff as the support they provided to some residents who were able to complete daily tasks fairly independently, but might forget to do them. In the case of participant 207 (borderline), for example, hostel staff maintained his daily diary of appointments, details of his medication regime and ‘things coming up’ that needed attending to. His key worker, SW14, said about participant 207 that ‘he couldn’t live on his own; he would forget to eat, or run out of clean clothes, etcetera. He would forget what day it is’, indicating the extent of participant 207’s problems with activities of daily living.

Participant 207 was himself aware of the help he was getting in supporting what may have otherwise manifested as memory problems:

If I was not living here [in hostel], it [his memory problems] would be serious; but I am living here so it’s not that serious as I’ve got help.

He referred to the hostel as ‘my home’, was very satisfied with the support he received from hostel staff, and described this support as just ‘assistance’ and not ‘care’.

Some staff felt that it was not always evident why residents needed prompting. In the case of participant 109 (borderline memory problems), his key worker, SW06, thought that he had early memory problems, which, combined with alcohol, resulted in his needing a great deal of prompting. SW06 described his view of participant 109:

Yes [he has] memory difficulties. A little. Because he’s drinking alcohol. Forgets laundry loads – hard to know why. We prompt him to do things because of hygiene and his drinking, not memory problems.

Challenges in making referrals to health and social care services

Many challenges were described by staff in several of the hostels concerning local referral systems to health and social care services for hostel residents whom staff felt needed more support. These ranged from the requirement to not be drinking at the time of assessments, how the personalised approach adopted by hostel support staff did not seem possible to replicate in formal services, the refusal of older residents to engage with services and services’ refusals to engage with residents.

Case study of ‘John’ (name anonymised)

Life skills and independence

Some residents with borderline memory problems were supported by hostel staff and managed to live mostly independently, with many of their everyday needs for care and support met in a relatively informal way. However, once individuals left the hostel and moved to more independent living situations or sheltered accommodation, this low-level but continuous support was not always available. In the case of participant 129 (memory problems), for example, staff member SW39 observed that in the hostel:

Here he’s [participant 129] fine and managing OK – but for moving-on purposes, he will need a carer [care worker], who sees him at least once a day for medication and self-care.

The hostel worker thought that this would be an unlikely development.

Similarly, staff member SW04 said about participant 104 (memory problems):

He’s been here [in the hostel] over the allotted time. I’d like to see him in a care home where he gets the support he needs. I need help from social services with this. He’s been allowed to stay on here because the manager wants him to have a positive move on. The level of support he needs will take money and that’s the problem. If he was put in a flat he wouldn’t last – he has no life skills. It’s difficult to prove his case – I’ve sent in medical assessments to X housing options but they say he doesn’t fit the criteria.

In this case, the hostel was able to continue to support participant 104 but this appeared to rely on the goodwill of the manager and of the hostel’s ability to be able to secure funding for participant 104 even though he did not meet the criteria set out in the hostel’s funding contract. This was not the only example of hostels ‘holding’ residents who they thought would deteriorate substantially if they moved on to locations that proved unable to support them.

Residents being reluctant to move

Staff in several hostels reported finding it difficult to engage residents and to encourage them to move from their present location. Several explanations were provided for such reluctance. For example, participant 108 (memory problems) could not find housing that was acceptable to him or suitable for his needs. Staff member SW08 said that this was not for want of effort; indeed, he had accompanied participant 108 to at least five locations to view potential new accommodation. However, each was rejected with the following reasons:

One place was in [the area] but was too far away; another was nice but the waiting list was too long; another was [name of house] but he didn’t like it. Also another in [name of area] but it was too independent. Also [name of house]. Nothing has worked out.

Although some of these reasons could be linked to the participant’s possible fear of not managing in more independent living and of being some distance away from his support or social networks, they were not as firmly linked to declining functional ability as in the case of participant 106. Here, hostel staff described their great and continuing difficulties in finding appropriate housing for him to move on to. Box 1 summarises what happened to participant 106 (memory problems).

Head injuries

Our findings of the level and extent of head injury among older hostel residents in this study have implications for any assessments of cognition as well as longer term effects of such injuries. It is likely that among this group not all head injuries had been reported to NHS professionals or were on patient records. Among hostel staff, their knowledge of residents’ head injuries was partial; thus, their role as possible informants in assessments or monitoring would be limited, although better than no-one.

Implications for the consideration of commissioners

• There is a need for local strategies to create and agree a plan(s) for the support of older homeless residents in hostels, which should be undertaken in partnership with the hostel staff. This might be usefully informed by public health expertise. Such specific planning and anticipation of need will help to ensure that current reactive care is part of a co-ordinated plan. Local work on integration may be the vehicle for such developments, but only if older homeless people are identified as a priority.

• Are hostels providing another form of ‘housing with care’ in the local health and care economy? The implications of our findings that some hostel residents have substantial care and support needs merits exploration of whether or not the use of such terminology might help integrate funding and provide better continuity of care and support. Furthermore, the closure of a hostel or change to its contracted funding arrangements may have substantial implications for older residents with high or growing support needs. Our study suggests that hostel provision might be encompassed in the new market shaping responsibilities of local authorities under the Care Act 2014119 and its wider responsibilities for local citizens’ well-being.

• We suggest that local health and care agencies’ training and education be welcoming to hostel staff. Such shared learning might also forge greater contact between hostel staff and statutory services to the benefit of their shared clients or residents.

Implications for the consideration of practitioners

• We do not know what social care staff supporting older hostel residents with or without memory problems find to be the most effective and acceptable elements of their practice. This study suggests that they may have a wider role in ‘bridging’ the different sectors, or at least their managers being able to communicate with other professionals if a care plan is not working well. As with hostel staff who are key workers, we do not know if and how care workers are engaged in care planning and anticipatory care. The implications of this study are that home care workers who provide care and support to hostel residents are potentially underused as a monitoring system, particularly in hostels without key worker systems.

• Living alone with dementia but not on one’s own is a situation highlighted by this study. People who are living alone or are isolated are not generally thought of as living in a congregate environment such as a hostel. However, increasing frailty and the needs of other residents may affect socialisation within a hostel. Hostel residents themselves and those who work with them may have relevant insight into the needs of people who are used to being on their own and person-centred experiences of how to help support be accessible and acceptable. These could be usefully explored by other practice communities working with older people who are lonely or unwillingly isolated. The implication of this study is that societal commitment to reducing the loneliness and isolation of older people should be inclusive of older people who have a history of homelessness.

• The practice evidence about the relative merits and benefits of different resources and supports for homeless people wishing to sustain their abilities and memory is thin. Memory support, environmental modifications and technology are variously suggested and appreciated among some people with dementia. A stronger evidence base for these in diverse settings might be developed so that health-care practitioners can better support people who are not usually seen as benefiting from them, such as older homeless people.

• This study found that many hostel staff would like more training. This is a broad topic; but in respect of ageing and memory problems, evidence of most effective and cost-effective ways to support older residents could be collated into good practice guidelines that could underpin training. Training resources for practitioners could be developed by the homelessness sector with the support of dementia practice communities.

• For clinical practice this study has identified several barriers to memory assessment that could be addressed by flexibility in terms of appointments, assessment practices and liaison with hostel staff. Records of individual health status and history are not always complete among people with histories of homelessness. One important implication of this study is that the presence of head injury or a history of trauma may merit particular consideration in assessments, but that records of such an injury may not be available or may not have been taken. Another clinical implication is our finding of high levels of depression among residents. Although this was being treated in some residents, its presence may merit greater preventative efforts, taking a public health approach.

Implications for the consideration of hostels

• Hostel staff are a key resource, and many were informed, supportive and committed to resident well-being. Providers may wish to encourage and assist them to access local training and practice development opportunities.

• Key working was a very effective way of supporting residents with high-level needs in hostels and some key workers undertook case management roles. Not all hostels run their service with key workers and their managers or owners may wish to consider this staffing arrangement.

• Hostel staff and hostel records may be the main sources of information about some residents. Practice varies widely in terms of record keeping and how such records are maintained. Hostels facing difficulties in this activity should work with partners in local authorities and health services to seek assistance so that such valuable data are not lost. The implication of this study’s finding that hostels may close quickly means that there is substantial risk of unique information about residents being lost. For those with memory problems, this is of great concern to future person-centred care that takes a life history approach.

• Hostels are not traditionally seen as dementia communities; but this research shows that they are supporting people with such conditions, diagnosed or not. The difficulties faced by hostels supporting people with memory problems are substantial and are compounded at a system level by demands for short stays and resettlement. Even those hostels whose facilities are more suitable for people with physical health conditions are not the optimal environment for living well with dementia. The implication of this study is that hostel providers will need to continue to advocate for their residents to move to more appropriate accommodation in general and that the appropriateness of this takes into account declining cognition or memory problems. Hostel providers may wish to build up local alliances or to join dementia community initiatives to improve the lives of their residents with memory problems and to supplement their advocacy on residents’ behalf.

Implications for the consideration of the research community

• Multiple research methods were used in this study. Not surprisingly, interviews were more productive than questionnaires or surveys among people whose recall levels might fluctuate. Our interviews with older homeless people seemed to be encouraging of discussion and facilitated recall in some instances; however, they were only partially successful (see Chapter 3) in providing factual information. A few records were incomplete or partial; as noted above, these could cover clinical histories but sometimes chronologies or biographies. Research with older homeless people needs more time and resource than is the case with other homeless people, and compared with studies of other housed older people with or without dementia. Further longitudinal research would be helpful in considering ways to support older homeless people with memory problems across the trajectory of multiple conditions and challenges.

• The funders of research are in a strong position to enable researchers to accommodate the efforts of data collection and may have a helpful role in the developing of researcher capacity to work in this sector when investigating health and social care questions, especially those related to addressing inequalities. The building up of experience of research participation in hostels, particularly those that are small and not part of larger charitable bodies, could be part of this capacity building. There are risks of bias in recruiting only from large hostels or day centres whose funders or central offices are committed to research but even in such locations this needs to be fostered locally. The professional or service backgrounds of researchers in homelessness studies assist greatly in understanding the subject, the approach and communications; if they are not available, then greater preparation and induction will be needed. An expert advisory group is strongly recommended, including people with lived experience of homelessness. Such capacity could be fostered and funded.