[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Next Article in Journal
Youth, Transferability, and Sport-Based Interventions: Reopening and Rethinking the Debate on the “What” and the “How”
Next Article in Special Issue
Prioritizing Prevention: Examining Shelter Diversion as an Early Intervention Approach to Respond to Youth Homelessness
Previous Article in Journal
Unlocking Male Youth Soccer Players’ Peak Performance Potential: Exploring the Impact of Maturation, Age, and Physical Demands on Neuromuscular Injury Risk and Recovery Following Competitive Matchplay
Previous Article in Special Issue
Diverse Profiles of Homeless Young Adults: Implications for Tailored Prevention Strategies—Insights from Belgian Homelessness Counts
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Youth
Volume 4
Issue 3
10.3390/youth4030082
youth-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The ‘Community of Schools and Services’ (COSS) Model of Early Intervention: A System-Changing Innovation for the Prevention of Youth Homelessness

by
David MacKenzie
1,*,
Tammy Hand
1 and
Peter Gill
2
1
Gonski Institute for Education, University of New South Wales, Sydney, NSW 2052, Australia
2
Institute for Health and Sport, Victoria University, Melbourne, VIC 3000, Australia
*
Author to whom correspondence should be addressed.
Youth 2024, 4(3), 1305-1321; https://doi.org/10.3390/youth4030082
Submission received: 25 May 2024 / Revised: 11 July 2024 / Accepted: 20 August 2024 / Published: 29 August 2024
(This article belongs to the Special Issue Youth Homelessness Prevention)
Browse Figures
Figure 1
<p>The Albury Project Governance Structure.</p> "> Figure 2
<p>Proportion of students identified as at risk of homelessness: Albury 2019–2023.</p> "> Figure 3
<p>The dynamics of risk of homelessness, Albury 2019–2023.</p> ">
Versions Notes

Abstract

:
Prevention and early intervention have become part of the Australian policy discourse; however, the prevention and early intervention of youth homelessness remain significantly underdeveloped and underfunded in practice. Consequently, too many young people experience homelessness. This article presents the ‘Community of Schools and Services’ (COSS) Model as an innovative approach to the prevention of youth homelessness. The COSS Model is an Australian place-based collective impact approach that uses data gathered via population screening in secondary schools to identify and then support adolescents at risk of homelessness and also reorganizes the local support system available to vulnerable young people and their families. This paper is not the result of a research project. Rather, this paper presents the findings of the Embedded Development and Outcomes Measurement (EDOM) report, which is a feature of the COSS Model. This paper is limited to findings from the COSS Model implementation in Albury, NSW, known as the Albury Project, from 2019 to 2023. The Albury Project has demonstrated significant reductions in the risk of homelessness and entry into the local homelessness service system. Findings reveal that: (1) when COSS Model support is delivered to identified at-risk students, 40–50% of individuals are no longer at such high risk of homelessness 12-months later; (2) only 3–5% of students identified as at risk of homelessness and supported through the COSS Model sought assistance from local homelessness services in the following two years; and (3) the flow of adolescents (12–18 years) into the local homelessness services was reduced by 40% from 2019 to 2023. As an evidence-based, complex innovation, there are major policy, funding, and implementation challenges in scaling the model to multiple community sites.

1. Introduction

Homelessness remains a persistent ‘wicked’ social problem in most Western countries, including Australia. The wickedness lies in homelessness being a complex issue with multiple causes that cannot be addressed by simple targeted programs and instead requires a whole-of-government approach with cross-departmental initiatives sustained over the long term [1]. The Australian policy discourse on homelessness has begun to highlight the importance of early intervention and prevention [2], yet the Australian homelessness service system remains largely a crisis management response.
In many countries, homelessness is defined more broadly than people sleeping in public places [3,4,5,6]. In Australia, since the early 1990s, homelessness has been defined more broadly than ‘rooflessness’ to include people staying temporarily with friends and relatives, residing in supported accommodation provided by homelessness services, and those people living in various housing situations without security of tenure [7]. The current definition of homelessness used by the Australian Bureau of Statistics (ABS) has also added severely overcrowded dwellings as situations of homelessness reported in official statistical estimates of homelessness using Census data [8].
In Australia, the official homelessness statistical estimates in the 2021 Census identified 122,494 individuals in situations of homelessness in a population of 25,400,000 people. About one quarter (23%) were young people (12–24 years) and another 14% were children part of a family unit [9]. About 14% of the 274,000 individuals who approached homelessness services in 2022–2023 were young people aged 15–24 years presenting alone [10].
Young people who experience homelessness are at risk of being disproportionately negatively impacted for the remainder of their lives, with issues including but not limited to lower education attainment, under and unemployment, and poverty. Experiencing homelessness during adolescence increases the chance of re-experiencing homelessness later in life. In an Australian study of intergenerational homelessness, two thirds of the homeless respondents (66%) reported that they had first experienced homelessness before the age of 18 [11]. Canadian surveys have produced similar findings [12]. In addition to the negative and possible long-term, and even life-long, consequences for individuals, the status quo of the response to youth homelessness carries a considerable cost burden for the Australian community and government [13]. Such findings point to the strategic value of early intervention to prevent youth homelessness and its reoccurrence later in life.
Youth homelessness and early school leaving have been described as ‘the twin peaks of youth disadvantage’ [14]. Students who become homeless while attending secondary school are highly likely to leave school before completing Year 12 unless their family issues are resolved. Early school leavers may not necessarily be at risk of homelessness at the time when they leave school, but if their home situation becomes more problematic, the disadvantage that results from leaving school early means that these young people are more likely to experience a period of homelessness in their later adolescence or early adulthood.
In this article, we present the ‘Community of Schools and Services’ Model of early intervention (COSS Model) as an innovative approach to the prevention of youth homelessness using population screening in secondary schools to identify adolescents at risk of homelessness who can then be supported to reduce that risk. The COSS Model is an Australian place-based collective impact model that reorganizes the local support system available to vulnerable young people and their families. A prototype site for the COSS Model was developed in the City of Greater Geelong, through which proof of concept evidence was obtained from 2013–2016. However, the first homelessness prevention funding for the COSS Model came under the Universal Screening and Support (USS) Pilot Program under the NSW Homelessness Strategy 2018–2023. The article focuses on the COSS Model implementation of the lighthouse COSS Model site, the Albury Project, located in the Regional City of Albury, 550 km from Sydney, in the State of New South Wales. The paper presents the emerging outcomes and impact data on reducing youth homelessness achieved through the Albury Project.
Section 2 is an overview of the COSS Model innovation and its core foundations, linking the model and its key principles to theory and the literature. This is followed by a discussion of the materials and methods used in the evaluation of the effectiveness of the COSS Model. In the fourth section, the results from the COSS Model implementation in the Albury Project are presented. The emerging evidence highlights the effectiveness of the COSS Model in terms of a reduction in the risk of homelessness for students in the identified at-risk cohort and reductions in the flow of identified at-risk young people into the local homelessness service system. In the final discussion section, we highlight the scale-up policy and funding challenges and the policy context for moving from an evidence-based innovation to multiple COSS community sites.

2. The Community of Schools and Services (COSS) Model

The ‘Community of Schools and Services’ (COSS) Model is an innovative early intervention place-based collective impact service-delivery and reform-oriented model for addressing and supporting vulnerable young people and their families to reduce disengagement from education and early school leaving and to help where family issues are heading towards a crisis whereby a young person becomes homeless [15,16]. The COSS Model takes a place-based systems approach [17] and represents a raft of innovations to realize a more effective early intervention system for supporting vulnerable youth. As such, the COSS Model is not an ‘off-the-shelf’ social program deployed alongside other programs with similar foci in the same community. The COSS Model is an exemplar of collective impact, see Table 1 (a more extensive discussion of the COSS Model as a collective impact initiative is outside the scope of this paper).
Collective impact is defined as ‘the commitment of a group of important actors from different sectors to a common agenda for solving a specific social problem’ [18]. Collective impact initiatives are different from the status quo of targeted programs by having a centralized infrastructure, a dedicated staff, and a structured process that leads to a common agenda, shared measurement, continuous communication, and mutually reinforcing activities among all participants [19].
Collective impact prioritises a focus on outcomes and addressing complex or ‘wicked’ social problems (such as youth homelessness) that potentially require change/reform at several levels. The local community level implementation of the COSS Model is relatively well-developed, but there are questions and issues about what needs to be done at the government and bureaucratic levels for a systemic whole-of-government approach [20].
Upstream Australia is the innovator of the COSS Model. Upstream Australia is purpose-designed for providing the backbone support as one of the five core features of collective impact. Since the collection and sharing of data on individuals is key to enabling early interventions for the identified at-risk students in secondary schools, Upstream Australia acts as a data custodian and supports the schools and community agencies to share data in a way that safeguards privacy [21,22]. In terms of a broader range of activities and responsibilities, Upstream Australia can be characterized as a ‘field-building intermediary organization’ working within the community sector to bring about systemic change by developing the capacity to achieve impact at scale around preventing youth homelessness and disadvantage [23].
The COSS Model has four key foundations: community collaboration; early identification; the practice framework and early intervention support work with youth and families; and robust, embedded longitudinal monitoring and measurement of outcomes. Each of these foundations will be discussed in turn in the following sections.

2.1. Foundation 1: Collaboration for Collective Action

The concept of a ‘Community of Services and Schools’ is about a deep, formalized collaboration between homelessness services, relevant community-based youth programs, and secondary schools in a definable functional community context [24]. The process of building this collectivity between schools and community services requires dedicated work over a 6–12 month period to the point where the various actors understand the Model and the community wants the local system changes associated with the COSS Model, and a planning group has been formed.
This deep collaboration is key, as early intervention for adolescents requires the participation of schools as universal institutions where young people can be reached. However, when problems arise, such as family issues leading to homelessness, young people are less likely to complete their secondary education. Homelessness then becomes an issue for homelessness services. So, while school-based social programs are necessary, there are some serious structural limitations when it comes to highly disadvantaged and vulnerable students.
The inter-sectoral collaboration is operationalized through specific COSS governance structures that need to be organized in each COSS community and that operate cohesively to build and sustain local community-level support and service system change. One of the foundational innovations is the creation of new institutional forms through which participants from schools, homelessness agencies, and health services can cross boundaries and work together in genuinely deep and formalized collaboration, sharing the power to make decisions that have implications for the vulnerable youth in their communities.
The collaboration between schools and services that makes possible a more integrated service system of support for at-risk young people requires a formal organization, which in Albury is the Albury Community Collective. Figure 1 outlines the structure, membership, and basic functions of the components of the Albury COSS Community Collective.
The Executive is the leadership of the collective; the Operational group consists of active workers from various partners who decide and organize the way various activities are coordinated. The school-based teams are an inter-disciplinary team that meets regularly to focus on what needs to be done for identified at-risk young people in particular schools. Upstream Australia is a collaborating partner, providing data management for the COSS community as well as a range of additional support, including, for instance, support to ensure fidelity to the Model.

2.2. Foundation 2: Early Identification of Risk—Population Screening

The COSS Model is an early intervention and secondary prevention approach that uses population screening so that interventions can be provided before crises. The terminology of ‘early intervention’ and ‘prevention’ has been derived from the health sector [25]. The generic practice typology of primary, secondary, and tertiary prevention is a continuum across a particular issue, operationalized in terms of actual interventions. The Australian movement around youth homelessness prevention has some unique features but is largely similar to prevention thinking in other countries [26,27].
Primary prevention or ‘universal prevention’ consists of programs or measures directed at an entire population or a whole-population cohort [28,29]. Secondary prevention focuses on intervening to support young people who can be identified as most at risk of becoming homeless. This may be ‘selected prevention’, or measures directed toward young people who are members of an at-risk group, such as all young people aging out of the Out-of-Home-Care system, including foster care. However, it can also be ‘indicative prevention’ or ‘targeted prevention’ that focuses on identifiable at-risk individuals and interventions directed to individuals because of characteristics known to place them in the high-risk category [16]. This is the ‘early intervention’ approach used in the COSS Model where risk is assessed by an annual individual-level population screening survey in secondary schools (see below) using at-risk indicators followed by a face-to-face interview to further ascertain the young person’s circumstances. A preventative response to youth homelessness is not primarily about access to housing, but the challenge of dealing preventatively with a range of family and individual risks and protective factors. That is, the COSS Model involves a universal multi-stage population screening process to identify young people at risk so that the COSS early intervention workers can intervene supportively to resolve issues before the onset of crises [17].
The Australian Index of Adolescent Development (AIAD) survey, developed by Upstream Australia, is used to screen all students from participating secondary/high schools for risk, not just a select ‘at-risk’ group. The AIAD survey collects identifiable student data on eight indicators—at risk of homelessness, disengagement from school, wellbeing, resilience, self-esteem, connectedness with family, school, teachers and friends, and psychological stress and anti-social behavior/risk of offending. The identification of at-risk students in the Albury Project is based primarily on the at risk of homelessness indicator and considers conjointly the indicators for disengagement from school and mental health [30].
The risk of homelessness measure is a proprietary five-item Likert scale assessing risk across five dimensions—attitude, disposition, behavior, relationships, and environment. Students who score in the 7–10 range on the at risk of homelessness indicator are followed up with an interview to determine what level of early intervention support is warranted. The at risk of homelessness scale was developed as part of a 1996 survey research project of some 40,000 secondary students in 63 schools across three Australian state jurisdictions [31]. Bearsley-Smith and colleagues (2008) subjected the at risk of homelessness scale to an external validation test, comparing the psychosocial profile of adolescents reporting elevated risk factors for homelessness with a sample of homeless adolescents. They concluded that the at risk of homelessness indicator ‘detects a significant subpopulation of adolescent students who are suffering significant emotional and family distress’ and concluded that the ‘five-item measure appears a valuable screening tool for further research in relation to adolescent risk of homelessness, depression and family difficulties’ [32].
There are issues with referral-based programs for young people. A systematic review of school-based programs for the identification of children and young people with mental health difficulties concluded that ‘evidence suggests that overall, universal screening may be the most effective method of identification’ but commented that the evidence base is not yet systematic and robust [33]. Compared to less formal processes (such as teacher or parent identification, referrals, or self-referrals), systematic school-based approaches detect a greater proportion of children and young people with mental health difficulties [34,35,36,37,38]. Teachers often report that they are not equipped to perform such systemic approaches and consistently under-identify early symptoms of various mental health disorders [39,40,41]. Research comparing the effectiveness of a teacher-rated universal screening instrument to typical teacher-referral methods for identifying youth at risk of emotional and behavioral problems found that only about half of the at-risk youth were identified correctly by the teachers [41].
In contrast, universal screening as a methodology for early identification, as used in the COSS Model, is more efficient. It does involve some additional costs, specialist data analysis, and reporting; however, it is an effective method of identification that removes the issues of relying on self or direct referrals and enables hidden populations of risk to be identified [42,43].
The AIAD survey is used as a key part of annual screening for risk. There are a range of consents, from the initial opt-out option for parents as to whether they object to their children taking the survey to active consent by students as to whether they want to take the survey and further consent for an interview or case management support [44]. Given that the purpose of the process and the COSS Model is to prevent youth homelessness and ameliorate cognate risks, ethics approvals have been granted for Upstream Australia to collect and manage identifiable data for the COSS Community Collective. Care and strict operational practices are followed to safeguard privacy, but efficient and effective early intervention depends on being able to identify vulnerable individuals to prevent homelessness crises from occurring.
The whole school population screening process enables hidden populations of risk to be identified and then supported, reducing any potential stigma attached to undertaking the survey as ‘everyone completes the survey’. The COSS Model screening methodology using the AIAD survey is designed to identify a young person at risk so that support and intervention can be delivered pro-actively and systematically before the onset of crises.

2.3. Foundation 3: Practice Framework: Youth-Focused Family-Centered Support

A flexible and responsive practice framework, the foundation of the COSS Model, has three levels of response—‘active monitoring’, ‘short-term support’, and ‘wrap around’ case management for complex cases. The work undertaken by the COSS early intervention workers is described as youth-focused and family-centered early intervention practice with a strong emphasis on a trauma-informed approach. This is work with identified at-risk students as well as with their families. Various types of intervention and support work are undertaken with identified young people and their families, depending on their needs. A 2021 research project undertaken by a team not connected to the COSS work nor Upstream Australia sought to answer the question: ‘What are the practices undertaken by early intervention workers with young people at risk, within the COSS model?’. The research project found that COSS early intervention practice spans four key settings—the young person, school, family/community, and internal and external service providers—and that COSS early intervention practice can be described as ‘an evidenced-based mode of practice … [with] … a dynamic, complex, creative and solution-focused way of working’ across multiple domains such as ‘mediator/coach’, ‘significant other/trusted adult’, advocate’, and ‘bridge builder’. Extended casework support was not required for every young person where family issues were evident and where there was a level of risk of homelessness or leaving school early before completing Year 12. Ideally, all identified at-risk students are regularly but unobtrusively monitored [45].
Early intervention requires an approach that is flexible, adaptive, a ‘whatever it takes’ approach, and a preparedness to not stay restricted within narrow professional boundaries. Clinical decisions about what is an appropriate response rest with the youth and family workers and the lead agency. A feature of the framework is the flexibility to move from one level or type of support to another as necessary, with no imposed restrictions on the duration of the support, and efficiency, in that interventions are undertaken for only as long as needed. Young people can always re-access support if needed, and identified at-risk young people are monitored even when case work is not active to enable re-engagement with former clients if/as necessary.
This work is undertaken with identified at-risk young people, and importantly, intervention is also provided to the young people’s siblings and parents/families if and as required. Typically, this work deals with a wide range of complex issues and means that working with a young person also means working with their family members. The support work involves the young person, their family, schools, and agencies working together under the same care plan. The amount of support provided to clients varies over time and as needed. This support is provided in a range of settings—inside schools, outside schools, in the community, in service provider organizations, and inside young people’s homes [46]. Family-based interventions are effective for a range of youth problems, with a caveat that in some cases where sexual abuse or continuing family violence are issues, young people may not be able to remain safely in the same household [47]. As such, the early intervention practice provided under the COSS Model is flexible and responsive and has the capacity to work with whole-family units across a range of issues and in different locations, and the lead agency in the community ideally has the capacity to provide supported accommodation should that be necessary.
The role and practice undertaken by the COSS early intervention workers are complimentary but different from what is currently undertaken by in-school welfare staff. In schools, wellbeing/welfare staff are vital and do important support work. However, a major limitation is that in-school workers cannot undertake deep family interventions in the way that community sector workers can. A critical innovation in the COSS Model is for in-school workers and COSS early intervention workers to, as far as possible, become a seamless workforce, funded through different funding streams but collaborating on the ground around early interventions to reduce and avoid crises. In Albury, this has been accomplished through the development of school-based teams (see Figure 1) involving in-school wellbeing/welfare staff, early intervention workers, crisis support workers, and other specialist workers as required who meet together on a regular basis for the purpose of ensuring that all identified young people are being supported and working towards meeting their case goals.

2.4. Foundation 4: Outcomes Measurement

One of the four core foundations of the COSS Model is a sophisticated embedded outcomes monitoring and measurement regime that includes an embedded triennial outcomes evaluation. Outcome measurement is the process of setting goals and defining approaches that measure performance towards the stated goals. In COSS Model sites, the ultimate goal is the prevention of youth homelessness and the amelioration of the disadvantages of young people and their families at the community level.
Upstream Australia provides ongoing cyclical monitoring and measurement of outcomes data to lead COSS agencies and community collectives as one of the five key characteristics of collective impact—i.e., data sharing. This is the collection of data and the measurement of outcomes in an ongoing way that directly and cyclically informs the intervention support practice of the lead COSS agency and the community collective.
There are several different but linked purposes for outcome measurement and monitoring under the COSS Model, including:
  • Data for accountability: On one level, outcome measurement is about accountability to assess the overall effectiveness of the COSS Model’s implementation in a specific community site and whether the investment of public funds is worthwhile and justifiable in terms of cost-effectiveness, as well as social return on investment or cost-benefit.
  • Data to inform practice: At the level of the individual, to inform practice, identifiable data are managed by Upstream Australia acting as the ‘data custodian’, observing strict ethical and data-sharing practices while ensuring privacy is safeguarded. The data from the AIAD Survey are pivotal to this work.
  • Data to monitor and measure outcomes: Under the COSS Model, data to monitor and measure outcomes occur across the individual, the identified at-risk cohort, and the community levels. A whole-community approach to outcomes for young people looks at the entire community cohort of vulnerable youth and monitors what has been achieved over time. The AIAD survey is administered on an annual basis and other monitoring data (including but not limited to local homelessness services data) are collected between the annual population screening process. This contrasts with the current agency-focused approach to service delivery contracts that assesses agencies against key performance indicators that are not usually constructed nor adequately resourced to redress the extant need in a community overall.
The next section discusses the materials and methods used to build the evidence-base of the effectiveness of the COSS Model in reducing youth homelessness.

3. Materials and Methods

The findings presented in this paper are based on the embedded evaluation of the COSS Model site, the Albury Project. An embedded, ongoing evaluation of client outcomes is a key component of the COSS Model. Embedded evaluations are undertaken by the Upstream Australia team responsible for the development of the COSS Model, which provides backbone support and data management/outcomes measurement for communities implementing the COSS Model.
Campbell (1979) questioned the presumed objectivity of external program evaluations when he asked ‘how objectivity in science is obtained in spite of the partisan bias of scientists’ arguing that program evaluations would benefit from adopting the scientist’s model of ‘experimenter-evaluator’ [48]. Similarly, embedded evaluations are arguably more appropriate for developmental projects and community-level social innovations. As the innovation developer and backbone support partner, the Upstream Australia team is an embedded ‘internal’ participant in the COSS community collectives but external in the sense that management of the data and identification of risk, monitoring, data matching, and measurement of outcomes must be undertaken on the basis of high standards of research and evaluation knowledge and expertise [49]. In addition, and of note, government-commissioned external evaluations of COSS Model sites in two Australian state jurisdictions have been undertaken [50,51].
Interrupted time series designs (ITS) are particularly well-suited for community interventions, collective impact projects, and evaluations across population-level interventions over specific time periods that target population-level health or social outcomes [52,53,54,55,56,57,58]. ITS can be usefully described as follows:
A time series is a continuous sequence of observations on a population, taken repeatedly (normally at equal intervals) over time. In an ITS study, a time series of a particular outcome of interest is used to establish an underlying trend, which is ‘interrupted’ by an intervention at a known point in time.
([58], p. 349)
Randomized Controlled Trials [RCT] are generally considered the ‘gold standard’ for the experimental evaluation of new interventions or treatments because randomization allows for the differences in outcomes between the treatment and control groups to be attributed to the intervention or treatment. However, in situations where only quasi-experimental designs are feasible, ITS becomes a preferred approach.
Interrupted time series (ITS) analysis is arguably the strongest quasi-experimental research design. ITS is particularly useful when a randomized trial is infeasible or unethical.
([59], p. S38)
A place-based collective impact model reaching for community-level impact, such as the COSS Model, is one such case. Thus, the embedded triennial outcomes evaluation in the COSS Model uses an ITS evaluation design. The ultimate test of the COSS Model’s effectiveness in preventing youth homelessness is a natural experiment in a community context whereby a cohort of at-risk students is identified and supported, and the entry of homeless young people into the Specialist Homelessness Services system is monitored longitudinally. The adopted time series approach uses annual measurements of risk and data on young people entering the homelessness service system prior to the fully developed implementation of the COSS Model in a community (pretesting) and continued annual measurements from that point thereafter (post-testing). Thus, the reduction in youth homelessness can be measured and assessed. The Albury Project from 2019–2023 provided the first opportunity to undertake this test.
This paper presents three main outcomes of the current implementation of the COSS Model at the lighthouse COSS Model site, the Albury Project. The three outcomes and data sources used to make these assessments are detailed as follows.
First, the extent to which the risk level of the cohort of at-risk secondary students in Albury has been reduced over a 12-month period of support. In the COSS Model, the first outcome measure is monitoring the risk level of young people identified as at risk of homelessness via the AIAD Survey over a 12-month period. Measured annually using data from the AIAD survey, this provides a longitudinal picture of the dynamic pattern of homelessness risk. Most young people identified as at risk of homelessness, together with their parents, will have consented to participate in some type of service/support/intervention; only a small proportion of young people will decline an offer of support or their parents will not consent [44].
Second, the extent to which the number of individual identified at-risk students has entered the Specialist Homelessness Service (SHS) system in the two years following their identification and COSS support. Although the overall level of risk of homelessness in all the participating schools in a COSS community may be relatively stable over time, that picture may conceal the real dynamics of risk. A young person’s risk status can change over time because family dynamics and situations fluctuate over time. A young person’s identified risk status one year after intervention/support may be reduced below the risk threshold in the next year. Other young people not at-risk in one year may be identifiably at-risk the following year. In terms of an outcome from the COSS intervention with identified students, the key change is what happens to students who are assessed as above the risk threshold in one year when they are reassessed in the next year. This analysis excludes Year 12 students who complete Year 12 and leave school a year later and the new students in Year 7, as they were not at secondary school the previous year.
Thirdly, the extent to which young people in Albury aged 12–18 years have entered the homelessness service system from 2019 to 2023. The community-level measure of homelessness prevention is a reduction in the flow of adolescents presenting to the local youth homelessness crisis service. Data from the local youth agency that provides the local specialist homelessness crisis services (that is, the youth refugee/crisis service and also the local lead agency that works with Upstream Australia to provide the youth and family work to COSS clients) are examined. This client data are cross-tabulated with AIAD Survey data to examine if the young people entering the local youth crisis service have ever been identified via the AIAD Survey as at risk of homelessness.
The following section details the results of the Albury Project from 2019 to 2023.

4. Results

The first section shows the proportion of the Albury school population identified as at risk of homelessness. The second section displays and explains the dynamics of risk. The third section follows identified individuals over the two years from the time they were first identified as at-risk to determine the extent to which identified and supported young people present to the local homelessness service. The fourth section examines the statistics on adolescents of secondary school age to determine if the incidence and prevalence of youth homelessness have changed. The emerging evidence from the Albury Project has begun to quantify the achievable effectiveness of the COSS Model [60]. There is evidence of reductions in individual supported young people’s risk of homelessness and reductions in the flow of identified young people into the homelessness service system, as well as a cohort-level reduction in the number of 12 to 18 year olds being supported in the local homelessness crisis service.

4.1. Risk of Homelessness Profile, Albury 2019–2023

The risk of homelessness in Albury was assessed annually from 2019 to 2023, and continues. Screening the entire secondary school population in the three local public secondary schools produces an identifiable cohort of students who are at risk of homelessness. The population screening is able to report on changes in risk for each identified individual and for the annual identified cohort a year later. The annual response rate for students completing the AIAD survey between 2019 and 2023 is: 2019, 74.40%; 2020, 87.27%; 2021, 86.39%; 2022, 83.40%; and 2023, 76.25%. Figure 2 shows the proportion of students identified as at-risk via the AIAD survey each year between 2019 and 2023.
During the COVID-19 pandemic, despite lockdowns, the Albury project workers and their in-school support staff partners managed to undertake the annual AIAD surveys with high response rates and continued their support of identified young people and their families using a range of creative methods of non-face-to-face engagement and support [61].

4.2. Reduction of an Individual’s Risk of Homelessness

The first outcome measure is the extent to which early intervention for students identified as at risk of homelessness can reduce their risk. The dynamics of risk are presented in Figure 3, which shows that in a disadvantaged community, the risk status of young people fluctuates from time to time as the exigences of everyday family life are experienced.
Figure 3 shows that of the AIAD-identified at-risk students who are engaged with COSS support, in the year that they are first identified as at risk, about 40–52% are no longer above the risk threshold a year later. The inference is that the support that is delivered has had an impact on the at-risk cohort. Another group remains at risk of homelessness because issues can be complex and not resolvable in the short term. However, and importantly, all these young people remain at school, and most continue living with their families. There is another group of young people who were not at risk of homelessness one year but are AIAD-identified as at risk of homelessness the following year. Problems escalate, and these young people are at risk a year later—life goes up and down.
This seems to be the ongoing pattern of risk incidence and risk amelioration within the Albury Project and the COSS Model, and this underpins the necessary practice of risk identification, monitoring, and outcome measurement as well as ongoing support practices by the in-school and COSS early intervention workers.

4.3. Reduced Flow of Identified and Supported Young People into Homelessness

The second outcome measure is the extent to which students in the identified at-risk cohort become homeless and present for assistance at the local youth crisis service (local specialist homelessness services (SHS)). This is a measure of what happens to the individuals identified as at risk of homelessness via the AIAD survey. This measure involves matching individual-level data from the AIAD survey with SHS client data for the two years after young people have been identified as at-risk and supported. The results are shown in Table 2 below.
As shown in Table 2, very few AIAD-identified young people go on to become homeless and clients of the local Specialist Homelessness Services system within two years of their AIAD identification as at risk of homelessness. Indeed, 94.9–96.6% of the young people identified via the AIAD survey and then supported via the COSS Model did not become homeless nor present to the local homelessness services within two years of the identification as at risk.

4.4. Reduced Numbers of Young People Supported in the Local Homelessness Crisis Service

Interrupted time-series design includes establishing an underlying trend, or baseline, for a particular phenomenon, then assessing the trend or baseline after implementation of the intervention. The historical youth homelessness pattern in Albury can be monitored and measured in terms of the number of individuals who are provided with supported accommodation in the local crisis refugee program. Admission to the Albury youth crisis service, which provides supported accommodation, is an indicator of SHS client numbers in the regional City of Albury.
In an ITS study, a time series of a particular outcome of interest is used to establish an underlying trend that is potentially ‘interrupted’ or changed by an intervention at a known point in time. As a requirement of the Specialist Homelessness Services program, extensive statistical data are collected on all clients of homelessness services, so there are data on the annual number of young people provided with supported accommodation in the local crisis service each year from 2016 (pre-COSS implementation) to 2019 (the start of COSS implementation) and subsequently to 2023. The average local SHS crisis refuge (Broughton House) intake from 2016–2019 was 48 individuals each year. COSS Model implementation began in Albury in 2009. The average youth crisis intake for 2020–2023 was 25 individuals, which is a decrease of 48 percent.
It must be noted that the COSS Model implementation in Albury worked through the COVID-19 pandemic, and it is unknown what impact this may have had on these results. That will only become apparent in the next few years post pandemic. However, a systematic review of individual and local SHS case records suggests that the support of at-risk young people and families via the COSS Model is likely to be the major factor.
A counter-factual point of comparison is that across the state of New South Wales (NSW) over the same period, the decrease in youth homelessness from 13,700 individual young people presenting to NSW SHS services for support in 2019–2020 to 12,994 individual young people in 2022–2023 was a 5% decrease, compared to a decrease of 48% in Albury (see Australian Institute of Health and Welfare (AIHW) annual SHS reports from 2019–2020 to 2022–2023).

5. Discussion

Implementation of the COSS Model as a place-based collective impact reform of local community youth support systems has demonstrated the achievability of a significant reduction in youth homelessness in Albury. The main interventions are provided to the young people identified as at-risk and their families, providing support to resolve or ameliorate whatever issues are present via flexible youth and family work that extends throughout secondary school and potentially post-school.

5.1. Limitations and ITS Evaluation Design

A limitation of the outcomes data presented in this paper is that they are based on a single COSS community. There are other Australian COSS communities at an early stage of development, and youth homelessness prevention initiatives are attempting the COSS Model approach in several other countries.
The COSS Model is a complex, real-world, system-changing innovation seeking to achieve a community-wide impact in reducing youth homelessness. The outcomes evaluation of the Albury Project and the COSS Model uses an interrupted time series (ITS) design.
Random Control Trials (RCT) are appropriately considered the most rigorous method for evaluating the effectiveness of interventions. However, RCTs have limiting factors for evaluating complex, real-world interventions, including: (a) high costs; (b) unsuitability for developing generalized theoretical principles about community change; (c) problems with innovative practices diffusing from intervention to control communities; and (d) obscuration of unique features of different communities to which they can be added. Although a RCT is a good method for testing replicability, it is a poor method for achieving replicability ([52], p. 32).
Future evaluations of multiple COSS Model sites will be able to attempt a quasi-experimental cross-site design for outcomes evaluation, possibly including community sites where risk data have been obtained but the COSS Model has not been implemented, thus strengthening the assessment of counter-factual inferences and the measured real-world significance of the COSS Model. An associated analysis of effectiveness is cost-effectiveness and this type of analysis, unreported in this paper, will be strengthened by the developmental and outcome measurement of additional COSS Model sites.

5.2. The Challenge of Scaling-Up Innovation Systemically

Readying communities for local youth support system reform and the organizing of COSS community collectives are foundational premises for the successful implementation of population screening and effective support services. The building of readiness for change and a community collective with a plan for early intervention takes at least 12 months. Thus far, the Albury Project has set a benchmark for what can be achieved even under the adverse conditions of the COVID-19 pandemic when all aspects of the COSS Model were implemented with strong fidelity.
A challenge for social innovations, even with a strong evidence base, is scaled-up implementation with fidelity and impact. Deigelmeier and Greco (2018) found that barriers to scaling up were most difficult and critical between the piloting phase and the scaled-up phase of an innovation, which they called the stagnation chasm, ‘where proven ideas get stuck before they are able to maximize their impact’. Implementation is about ‘the supports required to purposefully and reliably produce full and effective use of innovations in practice’, and many implementation efforts fail to achieve impact at scale and sustainability [62]. Complex innovations are more likely to fail to be successfully scaled up than simpler innovations [63,64]. Successful implementation depends on attending systematically to individual, organizational, and system-level factors [65,66].
Uptake of the COSS Model and implementation of the model in new sites are subject to broader policy reforms and associated funding. Although prevention and early intervention of youth homelessness have entered the Australian policy discourse, prevention and early intervention of youth homelessness remain significantly underdeveloped and underfunded in practice. Australia’s homelessness service system, like that of many other Western countries, remains largely in a state of crisis management, with a slowly but steadily increasing number of people seeking assistance each year.
There is evidence of the beginnings of change. During the COVID-19 pandemic, there were two government inquiries into homelessness and a review undertaken by the Productivity Commission. The Inquiry into homelessness in Victoria report (March 2020) recommended a shift away from crisis management to prevention and early intervention and a radical expansion of social and affordable housing ([67], pp. 158–166). A parallel Inquiry into Homelessness in Australia report (August 2021) recommended the development of a new 10-year national strategy on homelessness and identified three main areas for reform: prevention and early intervention ([68], pp. 165–166); housing first; and addressing the shortfall in social and affordable housing. The Productivity Commission review report, In need of repair (2022), recommended that early intervention and prevention programs should be a strategic focus for the next National Housing and Homelessness Agreement and National Housing and Homelessness Plan in 2025 ([69], p. 186). The practical significance of this policy shift has yet to be determined.
Currently, there are recommendations for the expansion of the COSS Model in two state jurisdictions. The COSS Model and the early intervention agenda are well represented in the homelessness policy discourse [2]. Australian governments make annual budget decisions, and at the time of this paper’s publication, we are still dealing with the financial consequences of the COVID pandemic, so the extent of investment in prevention has yet to be finalized. In addition, the shift to a place-based collective impact paradigm from a siloed departmental program approach (i.e., the status quo) arguably requires reform at several levels and a whole-government strategy. Australia may be poised to begin that reform process, but based on historical experience with other social policy reforms, even under the most optimistic scenario, it will take at least a decade-long change process to become an impactful part of the service system infrastructure.

Author Contributions

Writing—review and editing, all authors. All authors have read and agreed to the published version of the manuscript.

Funding

The development of the COSS Model has received funds from the National Homelessness Research Program (2010–2014), the Victorian Department of Human Services Innovations Action Projects program (2012–2013); the Lord Mayor’s Charitable Foundation (2015–2018), and the New South Wales Government (2019–2024).

Institutional Review Board Statement

The study was conducted in accordance with the National Statement on Ethical Conduct in Human Research 2023 and approved by the Human Research Ethics Committee of University of South Australia (Ethics Protocol code 201990, on a 12 monthly renewable extension, current until 28 July 2025).

Data Availability Statement

The datasets presented in this article are not publicly available because these are part of the identifiable agency data records used in client management systems. Requests to access datasets should be directed to the corresponding author.

Acknowledgments

We acknowledge the support provided to Upstream Australia by statistician Jascha Zimmerman from Swinburne University and the dedicated cooperation of the Albury Project YES Unlimited workers and Albury school principals and welfare/wellbeing staff.

Conflicts of Interest

There are no conflicts of interest to report.

References

  1. Head, B.W. Wicked Problems in Public Policy: Understanding and Responding to Complex Challenges; Springer Nature: Berlin/Heidelberg, Germany, 2022; p. 176. [Google Scholar]
  2. Hand, T.; MacKenzie, D. The Community of Schools and Services Model in Homelessness Policy Discourse: 2018–2023, A Policy Brief; Upstream Australia: Melbourne, Australia, 2023. [Google Scholar]
  3. Gaetz, S.; Barr, C.; Friesen, A.; Harris, B.; Hill, C.; Kovacs-Burns, K.; Pauly, B.; Pearce, B.; Turner, A.; Marsolais, A. Canadian Definition of Homelessness; Observatory on Homelessness Press: Toronto, ON, Canada, 2012. [Google Scholar]
  4. FEANTSA. European Framework for Defining Youth Homelessness; European Federation of National Organisations Working with the Homeless: Brussels, Belgium, 2020. [Google Scholar]
  5. Busch-Geertsema, V. Defining and Measuring Homelessness [Chapter 1]. In Homelessness Research in Europe; Edger, W., Doherty, J., Eds.; European Observatory on Homelessness: Brussels, Belgium, 2010; pp. 19–39. [Google Scholar]
  6. MacKenzie, D. Homelessness: Definitions. In International Encyclopedia of Housing and Home; Elsevier: Amsterdam, The Netherlands, 2012; pp. 25–35. [Google Scholar]
  7. Chamberlain, C.; MacKenzie, D. Understanding contemporary homelessness: Issues of definition and meaning. Aust. J. Soc. Issues 1992, 27, 274–297. [Google Scholar] [CrossRef]
  8. Australian Bureau of Statistics. Information Paper—A Statistical Definition of Homelessness; ABS: Canberra, Australia, 2012. [Google Scholar]
  9. Australian Bureau of Statistics. Estimating Homelessness: Census—Estimates of People Who Were Experiencing Homelessness or Marginally Housed as Calculated from the Census of Population and Housing; ABS: Canberra, Australia, 2023. Available online: https://www.abs.gov.au/statistics/people/housing/estimating-homelessness-census/latest-release (accessed on 22 March 2023).
  10. Australian Institute of Health and Welfare. Specialist Homelessness Services Annual Report 2022–23; AIHW: Canberra, Australia, 2024. Available online: https://www.aihw.gov.au/reports/homelessness-services/specialist-homelessness-services-annual-report/contents/clients-services-and-outcomes (accessed on 13 February 2024).
  11. Flatau, P.; Conroy, E.; Spooner, C.; Eardley, T.; Forbes, C. Lifetime and Intergenerational Experiences of Homelessness in Australia: AHURI Final Report No. 200; Australian Housing and Urban Research Institute Ltd.: Melbourne, Australia, 2013; pp. 26–33. Available online: https://www.ahuri.edu.au/research/final-reports/200 (accessed on 24 August 2024).
  12. Gaetz, S.; O’Grady, W.; Kidd, S.; Schwan, K. Without a Home: The National Youth Homelessness Survey; Canadian Observatory on Homelessness Press: Toronto, ON, Canada, 2016; pp. 5–8. [Google Scholar]
  13. MacKenzie, D.; Flatau, P.; Steen, A.; Thielking, M. The Cost of Youth Homelessness in Australia: Research Brief; Swinburne University: Melbourne, Australia, 2016; Available online: https://apo.org.au/sites/default/files/resource-files/2016-04/apo-nid63479.pdf (accessed on 24 August 2024).
  14. MacKenzie, D.; Hand, T. Youth homelessness and early school leaving: The twin peaks of youth disadvantage. Parity 2022, 35, 22–24. [Google Scholar]
  15. MacKenzie, D. Interim Report: The Geelong Project 2016–2017; Swinburne University: Melbourne, Australia, 2018; pp. 17–30. [Google Scholar]
  16. MacKenzie, D.; Thielking, M. The Geelong Project: A Community of Schools and Services Model of Early Intervention; Department of Families, Housing, Community Services and Indigenous Affairs: Canberra, Australia, 2014. [Google Scholar]
  17. MacKenzie, D.; Hand, T. Place matters... place-based ‘collective impact’: A new service delivery paradigm. Parity 2019, 32, 4–5. [Google Scholar]
  18. Kania, J.; Kramer, M. Collective Impact. Stanf. Soc. Innov. Rev. 2011, 9, 36–41. Available online: https://ssir.org/articles/entry/collective_impact# (accessed on 26 August 2024).
  19. Kania, J.; Hanleybrown, F.; Juster, J.S.; Edmondson, J.; Hecht, B.; Bartczak, L.; White, E. Collective insights on collective impact. Stanf. Soc. Innov. Rev. 2014, 12, 1–24. Available online: https://ssir.ebookhost.net/ssir/digital/29/ebook/1/download.pdf (accessed on 26 August 2024).
  20. Hyde, J. How to make the rhetoric of joined-up government really work. Aust. N. Z. Health Policy 2008, 5, 22. [Google Scholar] [CrossRef] [PubMed]
  21. Lawson, H.; Briar-Lawson, K. Connecting the Dots: Integrating School Reform, School-Linked Services, Parent Involvement and Community Schools; Danforth Foundation & The Institute for Educational Renewal at Miami University: Oxford, OH, USA, 1997. [Google Scholar]
  22. Hand, T.; MacKenzie, D. Structures to Support Collective Impact—Backbone Organisation and Multi-Levelled Working Groups Concept Brief; Upstream Australia: Melbourne, Australia, 2021. [Google Scholar]
  23. Social Ventures Australia. Insights on Australian Field Building Intermediaries and Their Funding Journeys towards Sustainable Impact; A report funded by the Paul Ramsay Foundation; Social Ventures Australia Consulting: Melbourne, Australia, 2022; Available online: https://mspgh.unimelb.edu.au/__data/assets/pdf_file/0005/4377893/Insights-on-field-building-intermediaries.pdf (accessed on 23 August 2024).
  24. Claiborne, N.; Lawson, H.A. An intervention framework for collaboration. Fam. Soc. 2005, 86, 93–103. [Google Scholar] [CrossRef]
  25. Kennedy, F. Beyond ‘prevention is better than cure’: Understanding prevention and early intervention as an approach to public policy. Policy Des. Pract. 2020, 3, 351–369. [Google Scholar] [CrossRef]
  26. Fitzpatrick, S.; Mackie, P.; Wood, J. Advancing a five-stage typology of homelessness prevention. Int. J. Homelessness 2021, 1, 79–97. [Google Scholar] [CrossRef]
  27. Dej, E.; Gaetz, S.; Schwan, K. Turning off the tap: A typology for homelessness prevention. J. Prim. Prev. 2020, 41, 397–412. [Google Scholar] [CrossRef]
  28. Durlak, J.A.; Wells, A.M. Primary prevention mental health programs for children and adolescents: A meta-analytic review. Am. J. Community Psychol. 1997, 25, 115–152. [Google Scholar] [CrossRef] [PubMed]
  29. McLeroy, K.R.; Norton, B.L.; Kegler, M.C.; Burdine, J.N.; Sumaya, C.V. Community-based interventions. Am. J. Public Health 2003, 93, 529–533. [Google Scholar] [CrossRef]
  30. Hand, T.; MacKenzie, D. The Australian Index of Adolescent Development, Concept Brief; Upstream Australia: Melbourne, Australia, 2021. [Google Scholar]
  31. Chamberlain, C.; MacKenzie, D. School student at risk. Youth Stud. Aust. 1996, 15, 11–18. [Google Scholar]
  32. Bearsley-Smith, C.A.; Bond, L.M.; Littlefield, L.; Thomas, L.R. The psychosocial profile of adolescent risk of homelessness. Eur. Child Adolesc. Psychiatry 2008, 17, 226–234. [Google Scholar] [CrossRef]
  33. Anderson, J.; Ford, T.; Soneson, E.; Coon, J.; Humphrey, A.; Rogers, M.; Moore, D.; Jones, P.B.; Clarke, E.; Howarth, E. A systematic review of effectiveness and cost-effectiveness of school-based identification of children and young people at risk of, or currently experiencing mental health difficulties. Psychol. Med. 2019, 49, 9–19. [Google Scholar] [CrossRef]
  34. Garland, A.F. Pathways to Adolescent Mental Health Services: Adolescent Help Seeking and Teacher Identification and Referral. Ph.D. Thesis, Yale University, New Haven, CT, USA, 1994. [Google Scholar]
  35. Eklund, K.; Renshaw, T.L.; Dowdy, E.; Jimerson, S.R.; Hart, S.R.; Jones, C.N.; Earhart, J. Early Identification of Behavioral and Emotional Problems in Youth: Universal Screening versus Teacher-Referral Identification. Calif. Sch. Psychol. 2009, 14, 89–95. [Google Scholar] [CrossRef]
  36. Scott, M.A.; Wilcox, H.C.; Schonfeld, I.S.; Davies, M.; Hicks, R.C.; Turner, J.B.; Shaffer, D. School-based screening to identify at-risk students not already known to school professionals: The Columbia suicide screen. Am. J. Public Health 2009, 99, 334–339. [Google Scholar] [CrossRef]
  37. Dowdy, E.; Dever, B.V.; Raines, T.C.; Moffa, K. A preliminary investigation into the added value of multiple gates and informants in universal screening for behavioral and emotional risk. J. Appl. Sch. Psychol. 2016, 32, 178–198. [Google Scholar] [CrossRef]
  38. Kieling, R.R.; Kieling, C.; Aguiar, A.P.; Costa, A.C.; Dorneles, B.V.; Rohde, L.A. Searching for the best approach to assess teachers’ perception of inattention and hyperactivity problems at school. Eur. Child Adolesc. Psychiatry 2014, 23, 451–459. [Google Scholar] [CrossRef]
  39. Caldarella, P.; Young, E.L.; Richardson, M.J.; Young, B.J.; Young, K.R. Validation of the systematic screening for behavior disorders in middle and junior high school. J. Emot. Behav. Disord. 2008, 16, 105–117. [Google Scholar] [CrossRef]
  40. Bruhn, A.L.; Woods-Groves, S.; Huddle, S. A preliminary investigation of emotional and behavioral screening practices in K-12 schools. Educ. Treat. Child. 2014, 37, 611–634. [Google Scholar] [CrossRef]
  41. Cunningham, J.M.; Suldo, S.M. Accuracy of teachers in identifying elementary school students who report at-risk levels of anxiety and depression. Sch. Ment. Health 2014, 6, 237–250. [Google Scholar] [CrossRef]
  42. Albers, C.A.; Glover, T.A.; Kratochwill, T.R. Where are we, and where do we go now? Universal screening for enhanced educational and mental health outcomes [Editorial]. J. Sch. Psychol. 2007, 45, 257–263. [Google Scholar] [CrossRef]
  43. Mangione, C.M.; Barry, M.J.; Nicholson, W.K.; Cabana, M.; Chelmow, D.; Coker, T.R.; Davidson, K.W.; Davis, E.M.; Donahue, K.E.; Jaén, C.R.; et al. Screening for depression and suicide risk in children and adolescents: US preventive services task force recommendation statement. JAMA 2022, 328, 1534–1542. [Google Scholar] [CrossRef] [PubMed]
  44. MacKenzie, D.; Hand, T. Informed Consent and the COSS Model, Concept Brief; Upstream Australia: Melbourne, Australia, 2020. [Google Scholar]
  45. Corney, T.; Narelle, B.; Brett, W.; Rachael, H. Early Intervention Practices in Youth Homelessness Prevention: A Study of Early Intervention Workers in the Community of Schools and Services Model; Victoria University: Melbourne, Australia, 2021. [Google Scholar]
  46. MacKenzie, D.; Hand, T. COSS Practice Framework Tier Ratings’, Concept Brief; Upstream Australia: Melbourne, Australia, 2021. [Google Scholar]
  47. Robinson, E.; Power, L.; Allan, D. What works with adolescents? Family connections and involvement in interventions for adolescent problem behaviours. Fam. Matters 2011, 88, 57–64. [Google Scholar]
  48. Campbell, D.T. Assessing the Impact of Planned Social Change. Eval. Program Plan. 1979, 2, 67–90. [Google Scholar] [CrossRef]
  49. Barry, D.; Kimble, L.E.; Nambiar, B.; Parry, G.; Jha, A.; Chattu, V.K.; Massoud, M.R.; Goldmann, D. A framework for learning about improvement: Embedded implementation and evaluation design to optimize learning. Int. J. Qual. Health Care 2018, 30, 10–14. [Google Scholar] [CrossRef]
  50. Planigale, M.; Rosauer, K.; Read, L.; Bottrell, D.; Goldzieher, M.; Yin, M. Evaluation of the Geelong Project: Report 4. Final Report; Department of Education & Training: Melbourne, Australia, 2021. [Google Scholar]
  51. Blunden, H.; Wong, E.; Bates, S.; Cheung, S.; Wing, H.; Katz, I. Evaluation of the Universal Screening and Support Program: Final Report; UNSW Social Policy Research Centre: Sydney, Australia, 2023. [Google Scholar] [CrossRef]
  52. Biglan, A.; Ary, D.; Wagenaar, A.C. The value of interrupted time-series experiments for community intervention research. Prev. Sci. 2000, 1, 31–49. [Google Scholar] [CrossRef]
  53. Lopez Bernal, J.; Cummins, S.; Gasparrini, A. The use of controls in interrupted time series studies of public health interventions. Int. J. Epidemiol. 2018, 47, 2082–2093. [Google Scholar] [CrossRef]
  54. Soumerai, S.B.; Starr, D.; Majumdar, S.R. How Do You Know Which Health Care Effectiveness Research You Can Trust? A Guide to Study Design for the Perplexed. Prev. Chronic Dis. 2015, 12, 150187. [Google Scholar] [CrossRef]
  55. Bonell, C.P.; Hargreaves, J.; Cousens, S.; Ross, D.; Hayes, R.; Petticrew, M.; Kirkwood, B.R. Alternatives to randomisation in the evaluation of public health interventions: Design challenges and solutions. J. Epidemiol. Community Health 2011, 65, 582–587. [Google Scholar] [CrossRef] [PubMed]
  56. Victora, C.G.; Habicht, J.P.; Bryce, J. Evidence-based public health: Moving beyond randomized trials. Am. J. Public Health 2004, 94, 400–405. [Google Scholar] [CrossRef] [PubMed]
  57. Wagner, A.K.; Soumerai, S.B.; Zhang, F.; Ross-Degnan, D. Segmented regression analysis of interrupted time series studies in medication use research. J. Clin. Pharm. Ther. 2002, 27, 299–309. [Google Scholar] [CrossRef] [PubMed]
  58. Bernal, J.L.; Cummins, S.; Gasparrini, A. Interrupted time series regression for the evaluation of public health interventions: A tutorial. Int. J. Epidemiol. 2017, 46, 348–355. [Google Scholar] [CrossRef]
  59. Penfold, R.; Zhang, F. Use of Interrupted Time Series Analysis in Evaluating Health Care Quality Improvements. Acad. Pediatr. 2013, 13, S38–S44. [Google Scholar] [CrossRef]
  60. Mackenzie, D.; Hand, T. The Albury Project: Embedded Developmental and Outcomes Measurement (EDOM) Report; The Albury Project: Albury, Australia, 2023. [Google Scholar]
  61. McGrath, K.; Glen, R.; Cattell, E.; Pain, M.; Brosolo, K.; MacKenzie, D.; Hand, T. The COVID-19 Pandemic in Albury: Weathering the Storm. Parity 2022, 35, 20–22. [Google Scholar]
  62. Deiglmeier, K.; Greco, A. Why Proven Solutions Struggle to Scale Up. Stanford Social Innovation Review, 10 August 2018. Available online: https://ssir.org/articles/entry/why_proven_solutions_struggle_to_scale_up (accessed on 25 August 2024).
  63. Demircioglu, M.A.; Audretsch, D.B. Conditions for complex innovations: Evidence from public organizations. J. Technol. Transf. 2020, 45, 820–843. [Google Scholar] [CrossRef]
  64. Chalmers, D. Social innovation: An exploration of the barriers faced by innovating organizations in the social economy. Local Econ. 2013, 28, 17–34. [Google Scholar] [CrossRef]
  65. Fixen, D.; Blasé, K.; Metz, A.; Van Dyke, M. Implementation Science. In International Encyclopedia of the Social & Behavioural Sciences, 2nd ed.; Elsevier: Amsterdam, The Netherlands, 2015; Volume 11, pp. 695–702. [Google Scholar]
  66. Rapport, F.; Clay-Williams, R.; Braithwaite, J. (Eds.) Implementation Science: The Key Concepts; Routledge: Milton Park, UK, 2022; pp. 38–41. [Google Scholar]
  67. Legal and Social Issues Committee. Inquiry into Homelessness in Victoria: Final Report; Parliament of Victoria: Melbourne, Australia, 2021; pp. 158–166. [Google Scholar]
  68. House of Representatives Standing Committee on Social Policy and Legal Affairs. Inquiry into Homelessness in Australia: Final Report; Commonwealth of Australia: Canberra, Australia, 2021; pp. 165–166. [Google Scholar]
  69. Productivity Commission. In Need of Repair: The National Housing and Homelessness Agreement; Commonwealth of Australia: Canberra, Australia, 2022; p. 186. [Google Scholar]
Youth 04 00082 g001
Figure 1. The Albury Project Governance Structure.
Figure 1. The Albury Project Governance Structure.
Youth 04 00082 g001
Youth 04 00082 g002
Figure 2. Proportion of students identified as at risk of homelessness: Albury 2019–2023.
Figure 2. Proportion of students identified as at risk of homelessness: Albury 2019–2023.
Youth 04 00082 g002
Youth 04 00082 g003
Figure 3. The dynamics of risk of homelessness, Albury 2019–2023.
Figure 3. The dynamics of risk of homelessness, Albury 2019–2023.
Youth 04 00082 g003
Table 1. How the COSS Model operationalizes collective impact.
Table 1. How the COSS Model operationalizes collective impact.
Collective Impact CharacteristicsCOSS Model Operationalization of Collective Impact
1. Common Agenda
▪
All stakeholders have a shared vision for a changed local service system, including a common understanding of the problem and a joint approach to solving it through agreed actions.
▪
A shared vision requires a community- building process of forging relationships, creating a collective structure, planning for change, and formalizing intersectoral and collaborative agreements.
▪
The process of community building takes six to twelve months and then requires ongoing maintenance to sustain the movement for system change.
2. Shared Measurement
▪
An annual Australian Index of Adolescent Development [AIAD] population survey is used to monitor and measure risk and some outcomes.
▪
Data matching with the Specialist Homelessness Services [SHS] * client and other data at the community level and using identifiable data provide for longitudinal analyses by the Upstream Australia organization, which is the data custodian for the Community of Schools and Services (COSS) community collectives.
3. Mutually
Reinforcing Activities
▪
Stakeholder activities must be differentiated while still being coordinated through mutually common plan/s of action.
▪
Early intervention for youth homelessness needs to be able to mobilize the capacity to address multiple and interrelated issues. The lead COSS agency should ideally have the capacity to do much of the support work and intervention with identified youth and their families and have ready access to crisis accommodation.
▪
Organized and structured coordination within the community collective is how activities can be efficiently and effectively delivered as well-coordinated interventions.
4. Continuous
Communication
▪
Continuous communication amongst the COSS stakeholders supports a shared understanding of objectives, practices, and development plans.
▪
Knowledge sharing through participation in the Upstream Community of Practice, includes webinars, the circulation of information or documents such as Concept Briefs and inter-site mutual assistance.
5. Backbone Support
▪
Creating and managing the COSS Model as a collective impact requires a separate organization with staff and a specific set of skills to serve as the backbone for the entire initiative and participating organizations and agencies.
▪
Upstream Australia, a purpose designed not-for-profit organization, was created to fulfil the role of backbone support that includes data management, community-building assistance, advocacy on policy and funding, further innovation and development, and facilitation of the movement for change.
* Specialist Homelessness Services [SHS] are government funded services in Australia that provide supported accommodation and assistance to people experiencing homelessness.
Table 2. AIAD-identified youth who become homeless within 2 years post identification.
Table 2. AIAD-identified youth who become homeless within 2 years post identification.
Year of AIAD
Identification		AIAD-Identified Clients Who Become Homeless within 2 Years
of Identification
[N] 		AIAD-Identified Clients Who Become
Homeless within 2 Years
of Identification
[%]
201954.5%
202075.1%
202164.2%
202243.4%
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

MacKenzie, D.; Hand, T.; Gill, P. The ‘Community of Schools and Services’ (COSS) Model of Early Intervention: A System-Changing Innovation for the Prevention of Youth Homelessness. Youth 2024, 4, 1305-1321. https://doi.org/10.3390/youth4030082

AMA Style

MacKenzie D, Hand T, Gill P. The ‘Community of Schools and Services’ (COSS) Model of Early Intervention: A System-Changing Innovation for the Prevention of Youth Homelessness. Youth. 2024; 4(3):1305-1321. https://doi.org/10.3390/youth4030082

Chicago/Turabian Style

MacKenzie, David, Tammy Hand, and Peter Gill. 2024. "The ‘Community of Schools and Services’ (COSS) Model of Early Intervention: A System-Changing Innovation for the Prevention of Youth Homelessness" Youth 4, no. 3: 1305-1321. https://doi.org/10.3390/youth4030082

APA Style

MacKenzie, D., Hand, T., & Gill, P. (2024). The ‘Community of Schools and Services’ (COSS) Model of Early Intervention: A System-Changing Innovation for the Prevention of Youth Homelessness. Youth, 4(3), 1305-1321. https://doi.org/10.3390/youth4030082

Article Metrics

Back to TopTop