Abstract
Access to appropriate accommodation is regarded as one of the most important determinants in successfully treating people with chronic mental illness in the community. Negative outcomes of not being adequately housed and supported can include deterioration in mental health, strain on family relationships, increased risk of suicide, involvement with the criminal justice system, inappropriate hospitalization and unnecessarily long stays in hospital [1–5].
The National Survey of Mental Health and Wellbeing [6], [7] reported there was a general lack of support services for people with psychotic disorders and a need for better access to public housing, flexible supports linked to accommodation and a range of residential disability support services. Patients with stable accommodation were more likely to be linked with specialist mental health services and to have their needs met, thus attaining a better level of functioning and improved quality of life. The report also found that one in 10 people with a psychotic disorder had been in hospital continuously for 12 months during the year preceding the census, which suggests there is a persisting problem of long-stay patients whose management and care needs cannot be adequately met in the community [6], [7].
High support housing in community settings is considered to be in short supply [8] and to date there has not been a systematic assessment of the range of services available to mental health consumers in New South Wales (NSW) or any other Australian state. Moreover, there is little consistency in the language used to describe models of care and the functional objectives of such models. In recognition of this, the first aim of this study was to survey all community-based facilities in NSW providing high support, very high support and residential rehabilitation services for people with chronic mental illness, to identify the range of services available and the models of care employed.
The second aim of the study was to assess the needs, satisfaction and functioning of patients residing in community-based high support services. A ‘needs-led’ approach to planning allows the supply of services to be closely matched to those needed and is equally applicable to the individual patient and to structuring service delivery [9–11].
Method
Survey of high support services
The development and distribution of the survey for assessing high support services was completed in three phases. In the first phase, all accommodation services for people diagnosed with a severe mental illness that met NSW Health's criteria for high support, very high support and residential rehabilitation were identified. This was achieved using the NSW Association for Mental Health database, the Mental Health Co-ordinating Council Member Directory 2001, via direct liaison with the NSW Department of Community Services and by contacting intake officers from all community health centres across the state. High support services were included in the survey if they met the following criteria: (i) provide accommodation and support for six months or more to mental health patients; (ii) provide services to stabilize symptoms, maintain functioning and facilitate community participation; (iii) have staff available for at least 8 hours per day, five days per week; and (iv) able to provide on-site support for a minimum of 8 hours per day, if necessary. New South Wales Health reported that for the financial year 2001/2002, there were 412 rehabilitation beds in public psychiatric hospitals and 266 extended care beds. These were excluded because the focus of this study was on accommodation available in community settings. Boarding houses were also excluded because they do not provide high support as defined by NSW Health.
In the second phase, a survey was developed, piloted and mailed to all the services identified in phase 1. Survey questions were based on previous research describing patient preferences for housing [12], [13] and a key policy document detailing ‘best practice’ in the provision of housing and support for people with mental health problems [14]. A pilot survey was sent to seven accommodation and support services asking them to comment on its ‘user-friendliness’, the time it took to complete, how easy the questions were to understand and whether the information requested was easy to access. The final survey comprised 58 items covering housing issues, services, patient information and staffing over the financial year ending 30 June, 2002.
The survey was sent to all service providers identified in phase 1 in July 2002. If the survey was not returned within 4 weeks, a friendly reminder was sent the following month. If there was still no response, the service provider was contacted by telephone and asked to answer a subset of questions about the service. Informed consent was assumed by voluntary completion of the survey.
In the last phase, service providers were asked if the research team could visit the facility to conduct face-to-face interviews. In order to collect an unbiased sample, three methods were used to recruit patients depending on the size of the facility: (i) for small facilities (12 beds or less) all patients were invited to participate; (ii) for medium-sized facilities (12–25 patients) an alphabetical list was made and all evennumbered patients were invited to participate; and (iii) for large facilities every fourth person on the alphabetized list was invited to participate.
Face-to-face interviews
At the start of each interview, patients were asked a number of simple questions about their current living situation, accommodation preferences and past experiences of accommodation, including homelessness.
Camberwell Assessment of Need (CAN)
The CAN consists of four sections that capture the subjective perception of needs from either a patient, staff or carer perspective [15], [16]. Need status is rated across 22 areas on a three-point scale: 0 = no serious problem; 1 = no serious problem or else a moderate problem because of a continuing intervention (i.e. met need); 2 = serious problem (i.e. unmet need); and 9 = unknown. The number of needs (scores of 1 or 2) and unmet needs (scores of 2) are aggregated over the 22 items. Ratings of not known (9) were re-coded to zero (meaning no problem) since subjectively not knowing one has a need is equivalent to not having one [17]. For each need recorded as being present, further questions were asked about who was providing help and whether the patient was satisfied with the kind and amount of help received. All interviews were conducted by a member of the research team after completing interview training according to the CAN guidelines [18]. All answers supplied by the patients pertained to their needs over the last month and they were paid $20 for their time after completing the interview.
Life Skills Profile (LSP)
Key workers were asked to complete a shortened version of the LSP [19], [20] for each participant who completed the CAN. The 16-item LSP (LSP-16) focuses on observable behaviours and offers a robust measure of functioning in four domains. The domains and number of items include withdrawal (4 items), antisocial behaviour (4 items), self-care (5 items) and compliance (3 items). Each item was rated using a 4-point scale (0–3) where higher scores indicate poorer functioning. When scoring each item, key workers were asked to consider the patient's general functioning over the past three months.
Analysis
Data were analysed using the Statistical Package for the Social Sciences (SPSS) V 10. Descriptive data are presented as response rates, percentages and means. Comparisons were made between services operated by NSW Area Health or non-government organizations (NGOs), those located in urban or rural areas and those that did or did not provide 24-hour support. Urban services were defined as those located within the metropolitan districts of Sydney, Wollongong and Newcastle and rural services were those located elsewhere. Categorical variables were analysed using the Pearson chi-square test for independence and group differences for continuous variables were analysed using oneway analysis of variance (ANOVA).
Results
Survey of high support services
A total of 42 high-support services were identified, of which 32 returned a completed survey. Eight other service providers were interviewed over the telephone (95% response rate) meaning that for some questions data were only available for the 32 services that completed a full survey. Table 1 lists in descending order the number of beds for each Area Health Service in NSW together with the number of residents over the financial year ending 30 June, 2002. Three Area Health Services had no housing providers that met the high support criteria (the Far West, New England and South-west Sydney).
Number of beds, residents and patient demographics for each New South Wales Area Health Service ranked in descending order by beds
The majority of services (63%) were operated by the nongovernment sector with 60% located in urban areas. Over half the NGO-operated services provided 24-hour support (56%) compared to a third of those operated by government (33%). Property ownership was more common among government-operated services (40%) than non-government services (8%). Forty-percent of non-government services leased property from the Department of Housing with this being the case for only 20% of government-operated services. Of the 753 identified beds, the majority (64%) were in properties with three or more bedrooms (329 beds), or hostels (154 beds). The rest comprised of two-bedroom settings (120 beds) or one-bedroom settings (91 beds) and 59 places were part of a home-based outreach program. Most (84%) of the places provided by government-operated services were in group homes or hostels, while this type of accommodation made up only 53% of the places available at NGOs. Most people with an intellectual disability (n = 101 of 121; 84%) resided in facilities providing 24-hour support.
The most frequent criteria for acceptance into the service were that the person must have a diagnosis of mental illness (70%) or be within a particular age range (60%). Over a third of services (38%) indicated that the person must agree to participate in a rehabilitation program. The most commonly quoted circumstances under which someone would be excluded from the service were if they had unmanaged drug and/or alcohol abuse (40%) or if the service did not have sufficient resources to meet the support needs of the patient (35%).
Table 2 shows the housing options available and patient involvement in choosing prospective tenants in government and NGOoperated services. Half the services (50%) indicated that patients could reside there permanently (60% of NGOs and 33% of Area Health Services). However, when asked whether patients were encouraged to move to alternative accommodation when their support needs change, 81% of services replied ‘yes’. The types of support provided by services were very similar and include living-skills training, medication management and social and leisure services.
Patient participation in government and non-government organization-operated services
The methods being used by services to evaluate their service included the LSP-39 (53%), individual service plans (50%), satisfaction surveys (47%) and Mental Health Outcome and Assessment Training (MH-OAT) scales (31%). However, government-operated services were more likely to use MH-OAT (75%) compared to NGOoperated sites (5%).
Most services had at least one nurse on staff regardless of whether they were area health (83%), non-government (80%), urban (90%) or rural (69%). More NGO-operated facilities had a psychologist on staff (60%) than area health-operated facilities (25%) and services located in urban areas were more likely to have a psychologist (63%) or social worker on staff (68%) than rural services (23% and 15%, respectively).
Satisfaction with services, needs assessment and functioning
A total of 159 patients from 25 sites were interviewed, representing 14% of the total number of patients in high support settings in the last financial year. Three-quarters were male and the mean age was 43 years (range = 19–79). Sites visited were evenly distributed between urban and rural areas (13 and 12, respectively), between those providing 24-hour support and those less intensively staffed (13 and 12, respectively) and comprised 60% of all government-operated services and 64% of NGO-operated sites (9 and 16, respectively). On average, seven patients were interviewed at each site visited.
Most patients (64%) resided at their current dwelling for less than two years and 45% reported they had lived on the streets or in crisis accommodation at one time because they could not find a place to live. When asked whether they would like to continue living where they were, 58% of patients said they would like to move elsewhere. About half (49%) indicated they would like to live independently and 31% said they would like to live alone. Of those patients who said they would like to live in shared accommodation (14%), friends or partners were specified as preferred housemates.
The prevalence and severity of needs as assessed by the CAN are shown in Table 3 in descending order of met need. The mean number of needs for care was 7.6 (range = 1–16). The mean number of met needs was 5.5 (range = 1–12) and the mean number of unmet needs was 2.2 (range = 0–10). There were no significant differences in the number of needs expressed by patients residing in NGO or governmentoperated services, those living in rural or urban areas and those patients in 24-hour services compared to those residing in less intensively staffed services. Patients were most satisfied with the type and amount of help they received with accommodation and looking after the home, self-care and sexual expression. Patients were least satisfied with treatment information, basic education, intimate relationships and daytime activities. Although a high percentage of patients were dissatisfied with help they received with alcohol and drug misuse, few patients expressed needs in these areas.
Prevalence of met and unmet need for 159 high support residents and satisfaction with the type and amount of help received
Table 4 displays the subscale and total LSP-16 scores for facilities that do and do not provide 24-hour staff support. Overall, patients in facilities that do not provide 24-hour support had lower levels of functioning as indicated by higher LSP-16 scores on two of the four subscales (antisocial and compliance). Patients in NGO-operated services had higher LSP-16 totals (18.5) than patients in government operated facilities (13.4, F 1,157 = 15.00, p < 0.001) and LSP-16 total scores were higher for patients in rural services (18.3) compared to urban services (15.4, F 1,157 = 5.26, p = 0.023). There were no significant differences in LSP subscale or total scores between men and women. Moreover, there were no significant correlations between LSP subscale or total scores and age, or the number of met or unmet needs on the CAN.
Mean LSP-16 subscale and total scores in residents of facilities with and without 24-hour staff support
Discussion
Survey of high support services
The majority of high support services identified in this study were operated by NGOs which is in line with NSW Health's statement that accommodation support should be provided by the non-government sector [14]. Property ownership was more common in governmentoperated services while most NGOs were leasing properties from the Department of Housing. However, the majority of properties had three or more bedrooms rather than the one or two-bedroom units preferred by patients [12], [13], [21]. Very few people were receiving support in their own homes on an outreach basis, which is in stark contrast to the Victorian system where the majority of patients receiving accommodation services in 2000/2001 were part of a home-based outreach program. Home-based outreach is known to be preferred by many patients [13], [22], [23].
Housing choice is important to an individual's satisfaction and success with community living [24] but only 56% of high-support facilities offered prospective patients a choice of housing options. This lack of choice may be due either to some government departments allocating patients to services or because most services can only provide support to patients who live in a property they own or manage.
Unmanaged substance abuse was the most frequently cited reason to exclude someone from entering a highsupport service. As housing is crucial to any integration of services for people with a mental illness and cooccurring substance use disorder [25], further investigation into the resources needed to support these patients is warranted.
Most of the service providers offering permanent housing indicated that patients were encouraged to move to alternative accommodation when their support needs changed. This is one of the characteristics of a continuum/ transitional model of care [26], [27] which may contribute to the lack of housing stability experienced by many participants. The shortage of housing further up or down the continuum means that many services have abandoned predetermined ‘lengths of stay’ and provide housing and support until other options become available. Services may have indicated that they provide permanent accommodation to reflect this, while still encouraging people to move to an alternative setting when their support needs change.
Clear differences were evident between government and NGO-operated services in terms of the evaluation methods being used. NGOs were less likely to use formal methods and this could be due to the very limited resources available to this sector and lack of opportunities to access training. NGOs have not been actively included in the MH-OAT implementation and this is reflected in the small number of services using this tool compared to government services.
Feedback from the survey did not support the common misperception that NGOs are staffed by less qualified individuals than government-operated services and this opens up enormous potential for the types of support that could be provided by the non-government sector. However, fewer than 25% of rural services employed psychologists and social workers, indicating a need for more flexible training options to be made available to staff in these areas.
Face-to-face interviews
Patients reported they were often subject to transient housing arrangements and a disturbingly large number had experienced homelessness (45%). Most patients wanted to move somewhere else and the most favoured characteristics of future housing were living independently or living alone [28]. Tanzman [13] concluded that independent living arrangements were not construed as living without support from mental health staff but rather that supports should be available as-needed rather than all the time.
Previous studies using the CAN have reported a mean number of patient-rated needs between 4.8 and 6.6 among users of general psychiatric community services [16], [17],[29–36]. Recipients of housing support in this study expressed on average 7.6 needs and a previous study reported 8.3 needs [37]. This indicates that patients receiving housing support have a greater number of needs than those who receive standard psychiatric services. However, the types of needs reported are similar across settings with patients in this study having similar areas of unmet need to those receiving other services. Unmet needs included information about condition and treatment, company, psychological distress, intimate relationships, sexual expression, daytime activities, physical health and transport [17], [31], [34], [37]. It seems that while a high level of support ensures that patients have a clean and tidy place to live with adequate food and regular medication, areas frequently overlooked are social and psychological needs and meaningful daily activities [38]. Although many patients felt that formal services would not be able to address their needs for intimate relationships or sexual expression, further research into the ways needs interact may identify areas where services can intervene to reduce unmet need in these more complex areas.
The mean LSP-16 score obtained in the current study (16.8) was similar to that reported previously (16.2) for patients in a community setting [39]. Longitudinal evaluation would be necessary to determine if the higher levels of functioning we found in government, urban and 24-hour services are a result of interventions, or whether patients enter these services with higher functional skills.
Conclusions
To our knowledge, this is the first state-wide review conducted in Australia of government and non-government facilities that provide high-support services for people with severe mental illness in the community. The survey identified 753 high-support community beds in NSW but there was a limited range of housing options available, especially for those who did not want to live in congregate settings. The NSW Select Committee on Mental Health [40] recommended that over the next two years the number of supported accommodation places be increased by 1000 and that NSW Health report annually on outcomes of their housing and support framework [14]. This study could inform the implementation of such recommendations and hopefully inspire more research into this neglected area.
We found that community-based high-support services are succeeding in meeting patients' needs in most of the 22 domains assessed by the CAN, but there should be a greater emphasis on enhancing patients' social interactions and providing more psychological support. Evidently, shared accommodation does not meet many patients' needs for socialization and the present study illustrates that sharing brings its own set of challenges. Existing social and leisure services may be helpful but in short supply, but for those patients for whom current services are inappropriate, further research is needed to identify ways to better meet their needs. The absence of a significantly higher number of patient-assessed needs in 24-hour settings compared to less intensively staffed settings raises questions about the way accommodation and support needs are assessed at the point of referral, especially as functioning levels were also higher within this group. The poor correlation between patient perceptions of need and staff ratings of functioning has implications for the choice of measures used in any such assessment.
There are a number of studies linking unmet needs to lower subjective quality of life [16], [31], [41]. It is hoped that the identification of unmet needs within this patient group will open the way for better-targeted service provision and in turn improve quality of life for mental health consumers in high-support accommodation. Also, patient feedback on specific areas of dissatisfaction could provide a starting-point from which to open up dialogue with patients about areas of service provision that could be improved.
Footnotes
Acknowledgements
We thank the following members of the steering committee for their involvement in this project: Jenna Bateman, Helen Blum, Danielle Fisher, Allan Hall, Elena Katrakis, Lindsay Oades, Joy Said and Victor Storm. We also thank research assistants Debra Smith and Emily Evenhuis, service providers and patients who agreed to be interviewed. This study was conducted by Aftercare with funding provided by the NSW Health Department. A more detailed report can be accessed online at 
or a printed version can be obtained by sending a cheque for $25.00 to Aftercare, 3/3 Wharf Road, Leichhardt, NSW 2040.