Introduction

With high hospitalisation rates and widespread social exclusion, mental illness is a profound health and social need that requires an integrated policy approach (Shepherd 2007). In the past, mental health models of support for people with acute mental illness have focussed

on clinical symptoms, rather than addressing the multiple aspects of a person's life, such as mental health, housing stability and community participation. This article describes the NSW Mental Health Housing and Accommodation Support Initiative (HASI) model and reports the findings of an evaluation of the program. The evaluation examined housing, mental health and community participation outcomes for the people involved in the program, as well as assessing the implementation of the integration model. The article includes a discussion of the policy implications for supporting people with mental illness to improve their mental health, stabilise their tenancy and increase their participation in their community.

Transitional Housing and integrated Models of Support

The risk of people with mental illness experiencing poverty and social exclusion is an international phenomenon. People with mental illness can experience profound disability in their public and private lives. They are often excluded from employment, education and social services and are at risk of homelessness or unstable housing and limited or no social supports (ABS 2006; Bassett et al. 2003; Australian Department of Health and Ageing 2006).

Mental health policies have had limited success in decreasing the disability and exclusion of people with acute mental illness. Deinstitutionalisation was intended to improve quality of life and decrease social exclusion of people with mental illness. While the deinstitutionalisation policy and subsequent community services enabled a shift from segregated, institutional care to community based living for many people with mental illness, living in the community did not necessarily translate to stable housing or social participation (Harvey and Fielding 2003). In the early 1990s, the widely cited Burdekin Report, for example, found many people with mental illness experienced discrimination, social exclusion and a contravention of human rights (HREOC 1993).

For people with acute mental illness, housing services largely focus on offering transitional models of supported housing. After leaving hospital or institutional care, people are accommodated in different types of supported housing depending on their clinical wellbeing. As people's mental health progressively declines or improves, they are required to move through different types of housing (Hanrahan et al. 2001; Tsemberis 1999). This model is based on clinical wellbeing and it has been successful in improving mental health and decreasing hospitalisations (Rog 2004).

The transitional model of supported housing however, does not offer secure housing or ways of building participation and strong community connections. In a review of 109 studies on transitional models of support, Carling (1990) concluded that this approach potentially increases stigma, limits the development of transferable living skills and fails to effectively integrate people into the wider community.

Recent developments in support for people with mental illness focus on local, community-based, coordinated services that provide treatment at home to support both symptoms and disabilities (Thornicroft and Tansella 2003). Yet for such a model to be successful, stable housing is critical. Individualised supported housing options attempt to stabilise housing in a community setting. Support is often provided in collaboration between clinical mental health services and other sectors, such as housing and non-government organisations (Ridgway and Zipple 1990).

In Australia, many programs continue to follow a transitional model of supported housing, in which psychosocial accommodation support is provided in collaboration with clinical support (Freeman et al. 2003; NSW Health, 2002; Dadich, 2007). In only a few exceptions are individualised supported housing programs offered that provide stable, long-term housing and community and clinical support. Some of these include the Independent Living Program, Western Australia 1995, Project 300, Queensland, 1996, and HASI, NSW 2002 (NSW Health, 2002; Meehan et al. 2004; Flatau et al. 2006).

This article reports on the findings from the HASI evaluation; one of the first published comprehensive evaluations of an Australian individualised supported housing program offering housing, community and clinical support to people with high levels of psychiatric disability. (1) It describes the program, presents results about client outcomes (housing, mental health, service use and participation) and discusses the policy implications.

HASI Model

HASI is a partnership between NSW Health, Department of Housing (Doll) and non-government organisations (NGOs), and is jointly funded by the two Departments. The program follows psychosocial rehabilitation principles and has a recovery focus. HASI aims 'to assist people with mental health problems and disorders requiring housing (disability) support to participate in the community, maintain successful tenancies, improve quality of life and most importantly to assist in the recovery from mental illness' (NSW Health and NSW Doll 2005: 1).

HASI Stage One was implemented in 2002. It provides approximately 100 people with high levels of psychiatric disabilities three types of support. These include social housing from the Doll and Office of Community Housing; support within and outside the home to develop living skills and participate in the community from NGO accommodation support workers; and clinical mental health case management from Area Mental Health Services. A coordinated, client-centred approach is used to provide this holistic support. HASI Stage One is offered in nine locations in six NSW Area Health Services: Greater Western, Hunter/New England, Northern Sydney/Central Coast, South Eastern Sydney/ Illawarra, Sydney South West and Sydney West. The program has since been extended to other clients and locations (Muir et al. 2007b).

Methodology

The Social Policy Research Centre was commissioned to evaluate HASI. The aims of the study were to determine whether HASI:

1. enables clients with high levels of psychiatric disability to maximise their participation in the community, sustain successful tenancies, and improve their mental health;

2. has appropriate and effective governance arrangements in place to support the establishment and ongoing development of the program; and

3. enhances access to specialist and generalist support services including housing, mental health, disability and other human services through processes of partnership and planning.

To examine these domains, a longitudinal, multi-method study was used. Spanning two years with three data collection phases at six-month intervals, the project involved 633 interviews and surveys with all consenting clients involved in the initiative; family members and carets (when appropriate); managers and direct support providers from the non-government accommodation support agencies; managers and case managers from the government mental health services; personnel from the government and non-government housing providers; managers from both government funding bodies; and consumer advocates. The aims for consulting with each cohort varied accordingly (Table 1).

Client outcomes were examined by inviting clients and those who supported them to participate in a semi-structured interview and complete survey material. For clients, this included a number of standardised psychometric tools, including the Global Assessment of Functioning. Service providers also provided quantifiable client information pertaining to demographics, health service use, living skills, social and community participation, and educational and employment pursuits. This information was collected using a client database developed for the study, as well as standardised assessment tools, including the Living Skills Profile and Health of the Nation Outcome Scale. Furthermore, hospitalisation and tenancy data were sourced for each consenting client who participated in the evaluation from administrative datasets maintained by the two government funding bodies. This enabled the researchers to triangulate the data from different stakeholders in different forms, which helped strengthen findings (Blaikie, 2000; Mathison, 1988; Onwuegbuzie & Leech, 2005). To examine internal and external governance arrangements, both senior personnel and direct support providers participated in a semi-structured interview and completed survey material. Questions clustered around the structures, functions and practices used to achieve program aims (Gill, 2001: Hudson, 1999). (2)

Recruitment

All clients actively supported through HASI at the time of the first phase of the evaluation (n=92) were invited to participate in the study. Given their established relationship with the clients, the accommodation support workers were important in assisting with recruitment. They informed clients about the study, explained its aims and determined client interest in participating at each phase of the evaluation. Of the 92 HASI clients invited to participate, 71 consented to be interviewed in phase one of the study and information on 90 clients was provided by accommodation support workers via the client information database (Table 2). Longitudinal interview and survey data was collected on 55 clients and longitudinal client information data was completed on 69 clients. It was only possible to obtain NSW Health hospital and assessment data from clients who consented as part of the evaluation, whose profiles could be matched by NSW Health in the administrative datasets and whose records were complete (n=67 and n=25 respectively). NSW Housing provided tenancy information on 105 HASI clients who they provided housing for throughout the evaluation.

Client Profile

Clients are typically men (67.3 percent), born in Australia, with a primary diagnosis of schizophrenia (74.3 percent) and a secondary diagnosis (64.3 percent; Table 3). They have histories of long-term hospitalisation, unstable tenancies, minimal living skills, low levels of occupational, social and psychological functioning, high levels of psychological and significant behaviour, impairment, symptoms and social problems.

It is difficult to determine whether people supported through HASI proportionately represent those in the community requiring such support. Levels of psychiatric disability by demographic details are not available. However, according to national demographic data on the prevalence of mental health problems, hospitalisation and mental health service use, women, people from culturally and linguistically diverse backgrounds and Indigenous Australians may be underrepresented in HASI.

While approximately one-third of HASI clients are women, the same proportion of men and women experience mental illness. The prevalence of certain mental illnesses differs by gender, but the most common diagnosis among HASI clients is schizophrenia and men and women are equally likely to experience this condition, albeit at different ages (AIHW 2005). Women are also more likely to have a mental health-related hospital admission than their male counterparts (AIHW 2005).

Some HASI staff believed women are underrepresented in the program because men generally have less support from family and/or friends and poorer living skills, and are therefore more likely to need" the program. While this may be the case, the prevalence of mental illness, the hospitalisation rates and the fact that women face other barriers accessing mental health services (Mowbray et al. 2003) suggests that the under representation of women in HASI requires further consideration.

People from culturally and linguistically diverse backgrounds also remained underrepresented in HASI, when compared to mental health populations. Although 14.5 percent of people with mental and behavioural problems living in Australia speak a language other than English (ABS 2001), only six percent of HASI clients spoke a language other than English. While people born outside Australia are slightly less likely than Australian born people to experience mental health disorders ('health migrant effect'; Andrews et al. 1999: 10), people from culturally and linguistically diverse backgrounds still account for 19.8 percent of mental health care service contacts (AIHW 2005).

Indigenous Australians are well represented in HASI as a proportion of the population, but their retention in the program is poor. In phase one of the evaluation, 8 percent of clients identified as Indigenous Australians. While recruitment of Indigenous people was strong, the proportion of Indigenous Australians in the program decreased significantly (p<0.05) to four percent. The poor retention rate of Indigenous people was not only statistically significant as an independent variable, but also the strongest indicator of a person exiting the program when regressed across time and other factors were controlled. (3)

Client outcomes were measured longitudinally. Where data were available, the following comparisons were made:

* Client experiences pre-HASI with client experiences during the program;

* Client experiences at the three phases of data collection; and

* Client experiences with normative data.

Housing Tenancy

Most clients had come from vulnerable housing situations. They were living in hospitals, boarding houses, refuges, crisis and other temporary housing or living with family members in situations that were unstable and unsuitable. With housing supplied by Doll or the Office of Community Housing and appropriate support to maintain the tenancies, clients were provided with affordable, secure social housing.

Most clients (70 percent) maintained stable housing since starting the program (Table 4). A further 15 percent had moved properties (either to exit the program or to move into more appropriate housing), but remained with the same housing provider, demonstrating a positive tenancy record.

It is difficult to account for the remaining 15 percent who were no longer with the same housing provider. (4) Evidence provided by the accommodation support workers and housing providers suggests that some clients left the program with positive rental histories that have or will assist them to obtain public or private rental housing. The housing situation of others who have left the program was unknown. It is likely that some returned to situations of housing instability and vulnerability.

Tenancies were successful for several reasons. Client mental health, personal preference, as well as access to resources and social networks were well considered when matching clients to appropriate housing. The sustained tenancies also reflect high levels of support from the accommodation support workers and housing providers. Furthermore, client satisfaction resulted in good property care, consistent rental payments (assisted by the use of an automatic rental payment system) and amicable neighbour relations. The ongoing intervention and prevention strategies from housing providers and accommodation support workers were critical to the stable tenancies.

Housing options for people with high levels of mental illness require significant forethought with regard to the type of housing; the social setting within the complex, street and suburb; and accessibility to social networks (family, friends and carers), services and resources (such as shopping facilities, doctors and hospitals). Half of the HASI clients accommodated by housing providers live in a unit or an apartment (52 percent), with the remainder accommodated in townhouses, villas, duplexes (31 percent) and houses (18 percent). This distribution is converse to the types of housing occupied by NSW public housing tenants, where 70 percent live in separate houses, 18 percent in a flat, unit or apartment and nine percent in a semi-detached, terrace or townhouse (ABS 2001 in Doll 2002). By phase three of the evaluation, housing providers were slightly less likely to place HASI clients in units or apartments. Interviews with stakeholders suggest that close living within units may increase the potential for conflict with neighbours. Standalone houses have also been problematic given the responsibility for ground maintenance. While a couple of clients avoided this burden by moving to townhouses or villas, the problem was also addressed through support strategies and subsidising gardeners.

The housing providers either leased or owned the properties used within the program. Leasehold properties provide a level of flexibility and choice for program participants that capital properties do not. Conversely, capital properties offer reliable tenancy security, which is not available from leased properties.

Stakeholders debated the appropriateness of clustered housing for people with mental illness. Some stakeholders considered it stigmatising, negating the aim of independent, integrated, community living. Others recognised it was beneficial in fostering social relations between tenants, but only if a maximum of three or four properties were grouped and people were carefully selected.

Finding housing that was accessible to social networks, services and resources was sometimes difficult, particularly when a person had to move away from family and friends to participate in the program. In most cases, it was possible to match client needs and preferences with housing. Consequently, 94 percent of clients were satisfied or very satisfied with their homes and tenancy retention rates remained high.

One of the limitations with the model is the prohibition of shared tenancies. Almost half of the clients (42 percent) wanted to share their housing with a friend, relative, partner or flatmate. They wanted to share with people who they trusted and with whom they had a meaningful relationship, rather than with strangers or acquaintances (as in group homes). Some clients contravened the rules and allowed other people to reside in their home. In a few instances, the co-tenancy was exploitative.

Mental Health

Mental health improved for most clients. Evidence for this was provided by the clients, their family members and carers, accommodation support workers, case managers, and housing providers. It was also strongly supported by a sharp decrease in the frequency and duration of hospitalisations (NSW Mental Health Outcome and Assessment Tool (MH-OAT) data also supported these findings, but as data quality was poor these results are not included within this article; Muir et al. 2007b).

NSW Health provided hospital data for 67 consenting clients whose records could be retrieved for five financial years (2000/01-2004/05). Data about psychiatric-related hospitalisations from psychiatric and emergency units included the dates of admission and discharge, unit type on admission, and facility. Data were categorised as pre-HASI or in-HASI. Hospitalisation in each period was analysed as the proportion of days hospitalised compared to the total days in the period.

Eighty-four percent of clients (n=56 of 67) spent less time in hospital since joining HASI, compared to the immediate years prior to HASI. Only 12 percent (eight clients) spent a greater proportion of time ha hospital since participating in the program (hospital records did not exist or were not available for the three remaining clients). Overall, the average time in hospital as a proportion of total time decreased from 24.3 percent of time pre-HASI to 4.6 percent since starting the program. When annualised, this equates to 88.7 hospital days per person per year prior to HASI and 18.9 days since; an 81 percent reduction (Table 5).

The average number of days clients were hospitalised per admission also dropped substantially (77.6 percent, an average of 23.2 fewer days per admission). The average hospital stay (6.8 days) was half the average of all mental health-related hospitalisations in 2003-04 (AIHW 2005).

At an individual level, some clients experienced significant decreases in the proportion of time spent in hospital. One client's hospitalisation decreased from 94 percent of their time in hospital to no admissions. Clients experienced a decrease in both the frequency and duration of hospital admissions since joining the program. However, the average length of time they spent in hospital decreased at a greater rate than the average number of occasions they were admitted (duration decreased by 81.1 percent, while frequency dropped by 16.7 percent; Table 5). This was primarily because one person used 73 of the 189 in-HASI admissions (38.6 percent).

The decrease in hospitalisation resulted in a significant drop in hospital expenditure on these clients. Between 1 July 2000 and client entry into HASI, the NSW government spent almost $7.5 million on client admissions to psychiatric units and emergency departments. Annualised per client, admission costs for clients decreased from $42,959 pre-HASI to $7,831 since the program started. The hospital dollars avoided for HASI clients using 2004-05 costs (NSW Health 2004) is estimated to be $41,589 per client; an 84 percent decrease in hospitalisation costs per client (Table 6). This represents over $4 million per year avoided on the hospitalisation costs of approximately 100 HASI clients, thereby increasing capacity for other patients. These are costs avoided for these participants rather than costs saved by Health.

Personal Wellbeing

Despite decreased hospitalisation and improved mental health, the personal wellbeing measurement of clients who participated in all three phases of the evaluation (n=55) indicated that there was a slight decrease in overall satisfaction in wellbeing between the first and third phases (Table 7; Personal Wellbeing Index; Cummins 2005).

These decreases could be related to shortcomings in the HASI service delivery model, but qualitative data from the interviews with clients and service providers suggests that it may be a result of changes in experiences, expectations and aspirations over time. The first score was measured when clients entered HASI--their experiences were new and their lives were considerably changed. They had a new home and new furniture; a renewed sense of security and safety; and an opportunity to re-establish, re-build or strengthen personal relationships. By the second phase of the evaluation, the feelings of elation and excitement were more subdued because they had moved from the 'honeymoon' period to one of stability and routine. The third measure of satisfaction was largely a reflection of clients' increasing awareness of life's possibilities; this may have been coupled with feelings of disappointment, frustration or helplessness about missed opportunities in the past and/or the ability to attain those possibilities--such as the ability to attain aspirations of paid mainstream work and/or education, and access mainstream social activities outside of their budget.

Importantly, dissatisfaction with health was more likely to be in relation to poor physical, rather than mental health problems. While there was generally a drop in satisfaction across all domains, the median scores have remained fairly stable and for every measure, except health, clients were more satisfied than not (scores remained over 50). The overall drop in wellbeing indicates, however, the need for HASI and similar service models to track, plan for and try to address decreases in wellbeing.

General Health

Of the 55 clients who participated in all three phases of the evaluation, 60 percent reported improved physical health (Muir et al., 2007a). This was due to regular contact with medical services; access to specialist services, where required; improved diet and increased physical activity; and general participation in the wider community. However, numerous clients had ongoing comorbid physical health problems, such as Hepatitis C, asthma and other respiratory problems, incontinence and arthritis.

Almost one in five clients reported declining physical health in phase three of the evaluation, and they were more dissatisfied with their physical health than their mental health (Table 7). It is therefore not surprising that client perceptions of their general health are poorer than the general population. Clients were more than twice as likely as the Australian adult population (18-64 years) to describe their general health as fair or poor (Table 8; ABS 2006). Despite physical health problems, according to the clients interviewed, HASI has helped to recognise, identify and treat these problems.

Access to Services

Clients access a range of health, specialist and generalist community services. Almost all (96 percent) have accessed clinical mental health professionals, psychiatrists (94 percent) and general practitioners (94 percent; Table 9). Consistent access to health professionals and ongoing monitoring of mental and physical health are a valued part of the program and are likely to have influenced reports of improved heath.

These results are likely to be an improvement relative to client experiences before the program, and relative to other people with mental illness. It is difficult to compare client service contacts with contacts made by people with mental illness in the general Australian population because of differences in data collection (ABS 2001). In the general population, 29.2 percent of people with mental and behavioural problems consulted a health professional within the two weeks prior to completing the survey and 36.9 percent consulted a doctor (ABS 2001). Client access rates are also high when compared to Western Australians with mental disorders (ABS 1999). Within a twelve-month period, 39 percent of Western Australians with a mental disorder had accessed a health service and 32 percent, a general practitioner. People with severe levels of psychiatric disability had higher levels of health service access (59 percent of women and 56 percent of men) than other Western Australians with more moderate or mild levels of psychiatric disability. Comparatively, 96 percent of HASI clients had used at least one health service. Throughout the evaluation, clients who saw a general practitioner (94 percent) had an average of 22 consultations (1.2 visits per month; Table 9).

Access to non-medical, general community services varied across sites and between clients. Clients accessed four different types of services:

* Work and financial services: employment agencies, government income support, Centre Pay and Office of the Protective Commissioner for financial management, and non-government organisations for welfare relief;

* Domestic services: home care cleaning, meal services, and garden/lawn maintenance;

* Social/community services: day programs, psychosocial and community support groups for people with mental illness, gymnasiums, sporting organisations, leisure centres, and art groups; and

* Educational services: TAFE, universities, community colleges and public libraries.

Access to community, welfare, educational and social services steadily increased throughout the program. This was largely attributed to client goals and the facilitation, assistance and networking of accommodation support workers and case managers. In some situations, however, more resources and direction are required to overcome barriers to accessing some services.

Community Participation

The final HASI objective for clients is to maximise community participation. Many increased their participation in the community after joining the program (Table 10). Statistically significant increases in participation included paid and voluntary work, training and education, as well as social and community activities. HASI enabled a shift from social exclusion towards social inclusion for many clients.

Clients started the program with limited social networks (23 percent had no friends) and almost all were excluded from work (8 percent were working, most of whom were engaged in supported employment) and education (only 2 percent were involved in study or training) (Table 10). By phase three of the evaluation, most clients had increased their social networks (94 percent had established friendships) and 83 percent were actively participating in their communities in at least one of three ways. They accessed shopping centres, dined out, accessed local libraries, attended social groups or church, studied, worked, played sport, participated in leisure activities, or exercised.

Seven in ten clients (73 percent) were involved in social and community activities and 43 percent were working and/or studying. Work included unsupported, paid positions (50 percent of people working); supported employment (33 percent); and volunteer roles (22 percent). Of those studying, the majority (57 percent) attended a Technical and Further Education (TAFE) college, 29 percent attended a community college or education centre, and 14 percent were enrolled at university. For some clients, support from employment agencies and educational institutions was pivotal, helping them to assume new roles and experience personal and professional development.

Recreational activities were also important. They provided opportunities to build social skills, increase confidence and enhance a sense of independence. Clients who had access to organised activities had greater opportunities for community participation than those who did not. This was particularly the case for clients supported by NGOs that organised social events. These events provided a suitable alternative to disability programs (which were often deemed as stigmatising) and mainstream programs (which were sometimes difficult to participate in).

Increased social contact also facilitated trust. In phase three of the study, client trust in other people was comparable to that of other Australians. The Australian Social Attitudes survey (2003) found 36 percent of 18-34 year olds believed most people could be trusted; this compares to 32 percent of interviewed clients. Fluctuating trust levels at an individual level (p<0.05) reveal that some clients were vulnerable to exploitation. HASI however, helped to improve client awareness of exploitation and reduced its likelihood.

While community participation has increased, feelings of loneliness have persisted for approximately half of the clients. Historically, most clients were isolated from their communities, irrespective of geographic location. For some, isolation and loneliness were further compounded by the need to move away from family and/or social networks to join the program. There is thus an important need to address client loneliness.

One of the most profound outcomes associated with HASI is the improved connection between clients and communities. HASI has eased tensions, reconnected some clients with estranged family members, and improved relationships for others. For many, the frequency and quality of family contact had improved since joining HASI. At phase three of the study, 81 percent reported being satisfied or very satisfied with family relationships. They enjoyed functional relationships with their families, partly consequent to improved mental health and housing stability.

Policy Implications

These findings demonstrate that the HASI approach of individualised supported housing can assist people with high levels of psychiatric disability. The approach can help people to secure stable housing, improve their mental health and increase their community participation. These findings also have important implications for future programs aiming to support people with these needs.

At the core of the model is stable housing in the community. Success in this outcome is associated with finding housing suitable for the person's needs and providing early responsive housing support to prevent tenancy breakdown. Factors included personal preferences, social connections, access to community opportunities, compatibility in the neighbourhood and flexibility to relocate when needs change or the housing proves to be unsuitable. Coordination between the housing, accommodation support and mental health partners is also instrumental to the success of the model.

Community mental health case managers provide early intervention when client wellbeing deteriorates. Clinical advice from case managers also complements the expertise of the accommodation support workers. In cases where the active participation of case managers is missing, the accommodation support workers struggle to achieve successful outcomes. Thus management structures to facilitate continued collaboration between clinical and community services are critical to the success of this type of program.

Good community support practice relies on case planning processes that are client-driven; involve other professionals supporting the client; and focus on incremental, achievable goals. Due to the episodic nature of some mental illness, the support is most successful when it can be flexible or structured, as required. Pathways from recreation to work or study can take considerable time. It is important that the incremental steps are based on healthy risks to ensure client confidence is maintained, skills are continually developed, and support is continued.

NGO workers were instrumental to referring, linking, transporting, organising and assisting clients to access community services and resources. These include educational facilities, employment agencies, welfare services, domestic services, as well as social and community services. NGOs require trained and motivated workers to identify and organise affordable activities, and motivate clients.

For most clients, HASI is not a transitional program. Given the fragility of mental health, the program offers continued support to clients, until they are no longer part of the program. The strength of the HASI model is that it acknowledges the integral nature of ongoing partnerships between the three service partners and commits sufficient resources to facilitating its success.

Conclusion

Mental health services have traditionally had a medical focus and because of their 'siloed' structures they have had little capacity to address the broader needs of clients, such as secure housing and community participation. Transitional models of supported housing move people from institutional congregate care through to smaller therapeutic settings, and then typically to more independent, small group housing in the community. As such, housing is dependent on client wellbeing. This approach has been criticised because it fails to provide housing certainty--there is little opportunity for sustained tenancies or community connections.

Akin to other initiatives (Meehan et al. 2004; Flatau et al. 2006), HASI has provided an alternative approach, where consumers with severe and chronic mental illness access long-term independent housing in the community, along with mental health case management and accommodation support. The model helps clients with tenancy matters, mental health issues, and disability support. The evaluation of HASI has demonstrated that this model can assist clients to maintain stable housing and connections with neighbourhood (despite histories of homelessness or marginal housing); significantly improve their mental health (with substantial reductions in days in hospital); improve their physical health by increasing their use of primary and specialist health care services; and decrease social isolation by increasing participation in social and community activities, education and employment. HASI has demonstrated that with appropriate housing and support, many of the exclusionary effects of chronic mental illness can be overcome.

Acknowledgements

Thank you to the participants in the HASI Evaluation; NSW Department of Housing and NSW Health and other Evaluation Reference Group members; Alan Morris; Michael Bleasdale and people who have commented on the research and article.

References

Australian Bureau of Statistics (ABS, 1999) Mental Health and Wellbeing: Profile of Adults, Western Australia, 1997-98, No. 4326.5, ABS, Canberra, ACT.

--(2001) National health survey: Mental health, No. 4811.0, ABS, Canberra, ACT.

--(2006) National Health Survey: Summary of Results 2004-0,5, No. 4364.0, ABS, Canberra, ACT.

Andrews, G., Hall, W., Teesson, M., & Henderson, S. (1999) The mental health of Australians, Canberra, ACT: Mental Health Branch, Commonwealth Department of Health and Aged Care.

Australian Department of Health and Ageing (2006) Homelessness and Mental Health Linkages: Review of National and International Literature, May 2005, Canberra.

Australian Indigenous HealthInfoNet (2005, 25th Jul.), Mental health, Accessed 18 July 2006, http://www.healthinfonet.ecu.edu.au/index.htm.

Australian Institute of Health and Welfare (2005) Mental health services in Australia 2003-04, Australian Institute of Health and Welfare, Canberra, ACT.

Bassett H., Lloyd C. & King R. (2003) 'Food Cents: Educating Mothers with a Mental Illness about Nutrition', The British Journal of Occupational Therapy, 66(8), 369-375, 1 August.

Blaikie, N. (2000) 'Using triangulation and comparative analysis to advance knowledge in the social sciences: The role of four research strategies'. Paper presented at the 5th International Conference for Methodologists in Social Sciences, Cologne, Germany.

Carling, P.J. (1990) 'Major mental illness, housing, and supports: The promise of community integration', American Psychologist, 45(8), 969-75.

Cummins, R. (2005) Australian Unity. Wellbeing Index, Survey 14, Report 14.0, Part B: Appended Tables 'The Wellbeing of Australians--Personal Relationships', Australian Centre on Quality of Life, Melbourne, http://www.deakin.edu.au/research/acqol/index_wellbeing/index.htm, accessed 21 July 2006.

Dadich, A. (2007), Telephone consultations with Australian State Government Departments and mental health services (Northern Territory Mental Health Coalition, QLD Office of Public Advocate, SA Mental Health Unit, Tasmania DHHS Mental Health Services, Victorian Department of Human Services, Mental Health Branch, WA Department of Health, Mental Health Division), April.

Disability Services Queensland (2008), Overview of Future Directions: Transitional Recovery Program, Queensland: Disability Services Queensland, Queensland Government.

Flatau, P., Martin, R., Zaretzky, K., Haigh, Y., Brady, M., Cooper, L., et al. (2006) The effectiveness and cost-effectiveness of homelessness prevention and assistance programs. Perth, WA: AHURI (Australian Housing and Urban Research Institute).

Freeman, A., Hunt, G., Evenhuis, E., Smith, D. & Malone, J. (2003) High Support Accommodation for People with Psychiatric Disabilities--a survey of high support, very high support and residential rehabilitation services in NS W and assessment of the needs and satisfaction of consumers residing there. Aftercare: Sydney.

Gill, M. (2001) Governance do's & don'ts: Lessons from case studies on twenty Canadian non-profits: Final report, Ontario, Canada: Institute on Governance.

Hanrahan, P., Luchins, D.J., Savage, C. & Goldman, H.H. (2001) 'Housing satisfaction and service use by mentally ill persons in community integrated living arrangements, Psychiatric Services, 52(9), 1206-9.

Harvey, C & Fielding, M. (2003) 'The configuration of mental health services to facilitate care for people with schizophrenia', Medical Journal of Australia, 178, $49-52.

Hudson, M. (1999) Managing without profit: The art of managing third-sector organizations, London: Penguin.

Human Rights and Equal Opportunity Commission (1993) Report of the National Inquiry into the Human Rights of People with Mental Illness, Canberra: Australian Government Publishing Service.

Jones, S. H., Thornicroft, G., Dunn, G., & Coffey, M. (1995) 'A brief mental health outcome scale: Reliability and validity of the Global Assessment of Functioning (GAF)', British Journal of Psychiatry, 166(5), 654-659.

Mathison, S. (1988) 'Why triangulate?' Educational Researcher, 13-14.

Meehan, T., O'Rourke, P., Morrison, P. & Drake, S. (2004) Evaluation of Project 300:Results at 3 years post-discharge, January 2004, Queensland Health and University of Queensland.

Morris, A., Fisher, K.R., Abello, D., Norris, K., Sutherland, K., Yallop, S., Dadich, A. & Muir, K. (2006) Housing and Accommodation Support Initiative, Evaluation Plan, report prepared for the NSW Department of Health, February 2005, SPRC Report Series 8/06, Social Policy Research Centre, University of New South Wales, Kensington.

Mowbray, C., Nicholson, J. & Bellamy. C. (2003) 'Psychosocial Rehabilitation Service Needs of Women', Psychiatric Rehabilitation Journal, Fall 2003, 27(2), 104-13.

Muir, K. (2006), Dadich, A., Abello, D., Bleasdale, M., Morris, A. and Fisher, K., Housing and Accommodation Support Initiative: Evaluation Report II, report prepared for the NSW Department of Health, SPRC Report Series 10/06, http://www.sprc.unsw.edu.au/reports/HASI_Report_II.pdf, Social Policy Research Centre.

Muir, K., Dadich, A., Abello, D., Bleasdale, M. & Fisher. K.R. (2007a) Housing and Accommodation Support Initiative: Evaluation Report III, http://www.sprc.unsw.edu.au/reports/HASI_Evaluation_Plan.pdf, SPRC Report Series 2/07, Social Policy Research Centre.

Muir, K., Fisher, K.R., Dadich, A., Abello, D. & Bleasdale, M. (2007b) Housing and Accommodation Support Initiative Evaluation: Final Report, report prepared for the NSW Department of Health, September, SPRC Report Series 10/07 and www.health.nsw.gov.au/pubs/2007/hasi_evaluation.html.

NSW Department of Housing (2002) Housing Today, NSW DOH, Ashfield.

NSW Department of Housing (2005) Annual Report 2004-2005, http://www.housing.nsw.gov.au/About+Us/Reports+Plans+and+Papers/ Annual+Reports/2004-05, accessed 5 September 2006.

NSW Health (2004) NSW Health Costs of Care Standards 2004/05, Sydney, NSW Health.

NSW Health, & NSW Department of Housing. (2005) Housing and Accommodation Support Initiative (HASI) resource manual (draft version 1.3). NSW, Sydney, NSW Health and NSW Department of Housing.

NSW Health (2002) Framework for Housing and Accommodation Support for People with Mental Health Problems and Disorders, Sydney, NSW Health.

NSW Health & NSW Department of Housing (2005) Housing and Accommodation Support Initiative (HASI) resource manual (draft version 1.7), Sydney, NSW Health and NSW Department of Housing.

Onwuegbuzie, A. J., & Leech, N. L. (2005) 'On becoming a pragmatic researcher: The importance of combining quantitative and qualitative research methodologies', International .Journal of Social Research Methodology, 8(5), 375-387.

Ridgeway, P. & Zipple, A.M. (1990) 'The paradigm shift in residential services: From the linear continuum to supported housing approaches', Psychosocial Rehabilitation Journal, 13, 11-32.

Rog D. (2004) 'The evidence on Supported Housing', Psychiatric Rehabilitation Journal, 27, 334-344.

Shepherd, J. (2007) 'The production and management of evidence for public service reform'. Evidence & Policy, 3(2), 231-251.

Thornicroft, G. & Tansella, M. (2003) What are the arguments for community-based mental health care? Copenhagen, WHO Regional Office for Europe (Health Evidence Network report), http://www.euro.who.int/document/E82976.pdf.

Tsemberis, S. (1999) 'From streets to homes: An innovative approach to supported housing for homeless adults with psychiatric disabilities', Journal of Community Psychology, 27(2), 225-241.

Wilson, S., Meagher, G., Gibson, R., Denemark, D. & Western, M. (eds) (2003) Australian Social Attitudes: the first report, UNSW Press, Sydney.

(1) Project 300 was evaluated by Meehan et al. in 2004. The WA Centre for Mental Health Policy Research was asked to review the Independent Living Program in 2006. Phase two of this evaluation, which focuses on client outcomes, has not yet commenced.

(2) This article focuses only on the outcomes for clients. Further detail about the methodology can be found in Morris et al. 2006 and further information regarding governance findings are available in Muir et al. 2007a, 2007b.

(3) This is based on population data from the client information dataset. The qualitative research suggests that most Indigenous clients who exited HASI left because the program did not meet their needs, rather than because they had improved mental health (for further discussion on people who exited the program, see Muir et al. 2006 and 2007a). Further research should be undertaken regarding the cultural appropriateness of this model for Indigenous Australians.

(4) The researchers could not follow up clients who exited the program and left their housing provider because contact details were unknown.

Table 1. Aims for consulting each cohort

                                   Evaluation   Evaluation   Evaluation
Stakeholder                          Aim 1        Aim 2        Aim 3

Clients                             [check]                   [check]
Family members and carers           [check]                   [check]
NGO accommodation support           [check]      [check]      [check]
workers
NGO accommodation support           [check]      [check]      [check]
managers
Case managers from government       [check]      [check]      [check]
mental health services
Managers from government mental                  [check]      [check]
health services
Housing providers                   [check]      [check]      [check]
Managers from government                         [check]
funding bodies
Consumer advocates                               [check]

Table 2: Number of clients included in the evaluation by method

Method                             Phase 1   Phase 2   Phase 3

Clients interviewed and surveyed     71         79        69
Client information database          90         90        96
NSW Health Hospitalisation data      --         --        --
NSW Health Assessment tool data      --         --        --
Housing data                         73         87        85

                                   Longitudinal
Method                             (all phases)   Total

Clients interviewed and surveyed        55         85
Client information database             69        113
NSW Health Hospitalisation data         --         67
NSW Health Assessment tool data         --         25 *
Housing data                            54        105

Note: * Only 41 consenting clients had completed assessment data and
only 25 of these had data completed prior to and after joining HASI.

Table 3. Characteristics of HASI clients (phases 1-3 of the
evaluation)

                    Description                             Percent

Primary diagnosis   Schizophrenia                              74.3
(n=109)             Schizoaffective disorder                   11.9
                    Bipolar                                     2.8
                    Depression                                  1.8
                    Other                                       9.2

Gender (n=110)      Male                                       67.3
                    Female                                     32.7

Age (n=106)         Under 35 years                             55.7
                    35-39 years                                 7.5
                    40+                                        36.8

Multiple            At least a dual diagnosis                  64.3
diagnoses (n=98)    History of substance use disorder          45.9
                    Substance use disorder                     29.6
                    Physical disability                        17.3
                    Intellectual or cognitive disability       31.6
                    Secondary mental illness                   11.2

Birthplace          Australian born                            93.6
(n=110)             Born overseas--English speaking             1.8
                    Born overseas--non-English speaking         4.5
                    country

Aboriginal          Aboriginal                                  6.4
(n=109)             Non- Aboriginal                            93.6

First               English                                    94.2
language (n=109)    Other than English                          5.8

Note: There was very little change In client demographics between
evaluation phases The proportion with schizophrenia remained
consistent; the percentage of males decreased slightly from 70 to 67
percent of the client group: and the proportion of clients under 34
years of age also decreased slightly from 61 to 59 per cent (see Muir
et al., 2007a).

Table 4. Sustained tenancy (entry to phase 3, n=105 *)

                                                              HASI
                    Explanation                           participants

Sustained tenancy   Remained in same home                     70%
                    (12 months or more)
                    Remained with housing provider            85%
                    in or out of HASI

Rental arrears      Proportion of tenants without             83%
                    rental arrears

Relations with      Proportion of tenants with no             70%
neighbours          complaints from neighbours
                    throughout the program

Satisfaction        Self-reported participant                 94%
with tenancy        satisfaction measured with a
                    five-point Liken: scale **

Property care       Housing provider managers reported         --
                    property care by the majority of
                    HASI participants were equal to or
                    better than general tenants

Notes: * Based on administration data from housing providers

** Based on responses from clients surveyed from the evaluation M=84)

Table 5. Hospitalization in psychiatric units and emergency
departments (n=67) *

                                                       Change since
                              Pre-HASI   In-HASI      entering HASI

Time in hospital as a            24.3%      4.6%     -19.7 percentage
proportion of total                                       points
time (percent)

Average number of days            88.7      16.8     81.1% reduction
hospitalised per person
per year

Average number of                  3.0       2.5       -0.5 days **
admissions per person                               (16.7% reduction)
per year

Median number of admissions        1.7         0   -1.7 less admissions
per person per year                                  (100% reduction)

Average number of days            29.9       6.7          -23.2
hospitalised per admission                          (77.6 % reduction)

Notes: * Based on data received from NSW Department of Health
hospitalisation records (2000/01-"004/05). Total hospital admission
days pre-HASI (days from first admission after 1 July 2000 to day prior
to HASI entry) = 15,449 over 63,606 days and in-HAS' (day entered HASI
to 30 July 2005) = 1267 admission days over 27,554 days in the program.
Total number of admissions pre-HASI 516 and in-HASI 189.

** This drop is small because 38.6 percent of all in-HAS! Admissions
were by one person (n=73 of 189).

Table 6. Estimated cost of psychiatric unit and emergency department
hospitalisations (n=67)

                               Pre-HASI     In-HASI        Change

Average cost per day           $136.04      $22.10        -$114.00
per person

Average cost per year         $49,654.63   $8,065.65     -$41,588.98
per person

Proportion of pre-HASI cost                   16%       84% reduction

Total cost for 100 people     $4,965,463                 -$4,158,898
per year

Source: NSW Health, 2004.

Notes: Based on 58,652,965.50 on 67 people over 63,606 days pre-HASI
($129,148.74 per person) and $608,874 00 on 67 people over 27,554
days in-HASI ($9p87.67 per person).

Table 7. Personal wellbeing (n=55)

Satisfaction with                 Phase 1   Phase 2   Phase 3

Life as a whole          Mean        59.4      58.0      54.4
                         Median        60        70        60
Standard of living       Mean        65.2      63.2      58.7
                         Median        70        65        70
Health                   Mean        56.0      53.7      49.2
                         Median        60        60        60
Achieved in life         Mean        64.2      56.6      58.1
                         Median        60        70        70
Personal relationships   Mean        63.2      60.4      54.2
                         Median        70        70        70
Feeling safe             Mean        70.4      68.2      60.9
                         Median        70        75        80
Part of the community    Mean        57.7      64.3      55.8
                         Median        60        70        70
Future security          Mean        68.0      61.2      57.0
                         Median        70        70        70
Mental health *          Mean        66.7      56.1      58.2
                         Median        70        60        70
Physical health *        Mean        56.9      51.4      53.4
                         Median        60        60        60

Note: * These are not part of the PWI, but were added in the
evaluation of HASI to separate clients' perceptions of their mental
and physical health.

Table 8. General health (percent)

                      HASI participants      National Health
                       (Phase 3, n=51)    Survey * (18-64 years)

Excellent/very good         27.5                   58.6
Good                        37.3                   28.0
Fair/poor                   35.3                   13.4

Note: * ABS 2006:16

Table 9. Health service use (n=109)

                             General                     Psychologist/
                           practitioner   Psychiatrist    counsellor

Proportion who had a           94%            94%             32%
consultation

Median number of visits         8              12              4
(of people who had at
least one consultation)

Average number of visits        22             15             11
(of people who had at
least one consultation)

                           Specialist,    Clinical health
                              other       professional

Proportion who had a           51%            96%
consultation

Median number of visits         3              18
(of people who had at
least one consultation)

Average number of visits        5              36
(of people who had at
least one consultation)

Table 10. Community participation (percent)

                    No friends         Social and       Work (paid,
                     (n=69) *           community        voluntary;
                                       activities      supported or
                                        (n=69) *       open; n=55) **

Entry to HASI           23                 --                8
Phase 1                 15                 73                18
Phase 2                 7                  84                31
Phase 3                 6                  73                26

                   Education or        Work and/or
                training (n=55) **   study (n=55) **

Entry to HASI           2                   9
Phase 1                 22                 37
Phase 2                 18                 41
Phase 3                 20                 43

Notes: * 69 people had participated in HASI from Phase 1-3 of the
evaluation.

** Longitudinal interviews with people who participated in
the evaluation, Phase 1-3.

All longitudinal differences from Phase 1-3 are
statistically significant at p<0.05.