Protocol for a Systematic Review of Policies, Programs or Interventions Designed to Improve Health and Psychosocial Outcomes for Young People Leaving the Out of Home Care System

Background: Relative to their counterparts in the general population, young people who leave, or transition out of, out-of-home (OOHC) arrangements, commonly experience poorer outcomes across a range of indicators, including: higher rates of homelessness, unemployment, reliance on public assistance, physical and mental health problems, and contact with the criminal justice system. The age at which young people transition from OOHC varies between and within some countries, but for most, formal support ceases between the ages of 18 and 21. Programs designed to support transitions are generally available to young people toward the end of their OOHC placement, although some can extend beyond. They often encourage the development of skills required for continued engagement in education, obtaining employment, maintaining housing and general life skills. Little is known about the effectiveness of these programs, or of extended care policies that raise the age at which support remains available to young people after leaving OOHC. This systematic review will seek to identify programs and/or interventions that improve outcomes for youth transitioning from the OOHC system into adult living arrangements.


Background
Youth who experience abuse and neglect by their parents or carers can be placed in out-of-home care (OOHC) in jurisdictions where such formal systems exist. OOHC takes three major forms: foster carewhere care services are provided by individuals not necessarily known to the recipient; kinship carewhere those providing care are connected to the recipient through blood or kin ties; and residential carewhere care is provided in an institutional setting. Youth can experience one or more of these care types while in OOHC. Whilst the three forms of OOHC are quite different, they also have similarities: children in homes were often the victims of childhood trauma (1,2); minimal standards of care are required; and nancial support, if provided, ceases when youth reach a certain age (3).
OOHC is a policy area of considerable contemporary cross-national interest (4 left care in FY2017, because they reached the legal age of adulthood in their state (7).
Young people who leave or transition out of OOHC arrangements commonly experience poorer outcomes across a range of indicators relative to their counterparts in the general population, including higher rates of homelessness, unemployment, reliance on public assistance, physical and mental health problems, and contact with the criminal justice system (8)(9)(10)(11)(12)(13). These poorer outcomes may be due to pre-existing mental health problems and other challenges arising from their experiences pre-care or whilst in care. They may also be due to insu cient life skills knowledge or training, or may simply be related to the fact that they must fend for themselves at a much earlier age than peers who can rely on their birth families for personal and material support (14).
The age at which young people transition from OOHC varies between and within some countries -for most, formal support ceases between the ages of 18 and 21 (15). The type and mode of support to care leavers varies between jurisdictions, but can involve formal life skills training programmes, personal adviser or key worker support, mentoring or peer support programs, cash assistance, and housing support. Transition support programmes are generally available to young people toward the end of their care placement, although some extend beyond. They often encourage the development of skills required for continued engagement in education, obtaining employment, maintaining housing and general life skills (14,16).
Young people transitioning from care are often ill-equipped for independent living, and the type and amount of support they receive is insu cient to prevent adverse outcomes (17,18). Considering this, in some jurisdictions, one policy response has been to raise the age at which young people can transition, thereby extending the support available (19,20).
Since then, reviews have explored various aspects of policies, programs or interventions for youth transitioning from care, but have suffered from a range of weaknesses in either methodology or scope.
Some reviews have limited their scope, either to particular geographies (16), or to interventions delivered while youth were in care (14,21), or to independent living programmes (14,22,23). Some focused on a narrow range of outcomes (23,24). Others have methodological weaknesses, such as not conducting a systematic search (18), not applying a methodological lter or addressing the risk of bias of included studies (21,25), or they failed to critically appraise the effectiveness of the policy or practice interventions included (22,23,(26)(27)(28).
This review seeks to update this previous work and to include research published in the proceeding period, including that investigating the impact of raising the age at which young people transition.

Aim
The aim of this systematic review is to assess the effectiveness of programs and/or interventions designed to improve outcomes for youth transitioning from the out-of-home care system into adult living arrangements. The proposed systematic review question is: What programs, interventions or services are effective at improving health and psychosocial outcomes for young people leaving the out of home care system?

Methods
The systematic review has been registered with the International Prospective Register of Systematic Reviews (PROSPERO, http://www.crd.york.ac.uk/PROSPERO, registration number: CRD42020146999).
The review will follow the recommendations as laid out in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA-P) (29) (see Additional le 1). Findings will be reported as per the PRISMA guidelines (30).

Intervention
Programs, interventions or services that: provide support and/or assistance to help youth prior to leaving care and/or as they transition and/or after they leave care; are delivered in the community; support young people transition from their country's statutory out-of-home care systems into adult living.

Comparator
The following comparisons will be included: intervention/services as usual, another intervention, no intervention, or wait-list control.

Outcome(s)
Outcomes of interest include the following, at least three months following the age at which eligibility for out of home care terminates in the jurisdiction in which the study takes place. We will also measure outcomes at the point at which any policy, program or intervention ceases. Primary: Homelessness -we will include measurement that allows us to determine whether or not an individual has or does not have a permanent place to live; Health -we will include health outcomes or service usage, including emergency department presentations, hospitalisations, mental health outcomes and sexual health test results; Education -we will include measurements of high school completion, high school grades, enrolment in or attainment of a trade quali cation and enrolment in or attainment of a university quali cation; Employment -we will include measurements of whether an individual has a job, their wages, is unemployed or uses unemployment bene ts; Exposure to violence (as either victim or perpetrator) -we will include any measurement of crime perpetration i.e., whether or not an individual has been arrested, convicted, spent time in a locked setting (jail/prison) or crime victimisation; Risky behaviour -we will include measurement of risky behaviour including illicit drug use, alcohol use, risky sexual activity and either the onset or delay of teen pregnancy. Secondary: Relationships -we will include measurements of whether individuals have attained or maintained supportive relationships with others, including mentors, peer mentors or supportive peers; Life skills -we will include measurements of the attainment of competencies required for independent living; these include, but are not limited to, learning how to budget, attain essential services and perform essential household tasks.
Study design(s) The following experimental and quasi-experimental study designs will be included: Randomised Controlled Trials (RCTs) including individual RCTs, cluster RCTs, Step-Wedge designs with random time allocation, and non-equivalent control group designs using parallel cohorts that adjust for baseline equivalence. Difference-in-Difference estimation, synthetic control group methods, studies based on covariate matching, propensity score-based methods, doubly robust methods, regression adjustment, regression discontinuity designs, and instrumental variable estimation. Economic evaluation methodologies: cost-bene t analysis, cost-utility analysis, cost-effectiveness analysis, cost-analysis.
Economic evaluations and qualitative studies will be included if they are conducted as part of a qualifying study and will be used only to inform or deepen our understanding of the quantitative ndings.

Search strategy
The following databases will be searched for studies published between 1990 and October 30 2020 using strategies included in (Additional le 2): Cochrane Authors of each included study will be contacted by email to ascertain if they are aware of any additional literature.

Language of publication
Any study with an abstract, or translation, available in English, Spanish, Portuguese, Swedish, Norwegian, Danish, German, Japanese, Mandarin or Cantonese will be included during title and abstract screening. If deemed relevant, full text screening and data extraction will be undertaken by a reviewer pro cient in that language. In the event that any additional studies are identi ed in additional languages, the review team will endeavour to locate reviewers pro cient in that language to screen and extract data.

Data collection and management
Citations identi ed from the search strategy will be imported into the online systematic review application Covidence (31) for screening. Following PRISMA guidelines, a owchart will be provided demonstrating the inputs and results of each stage of the review process.

Study selection
Title and abstracts will be reviewed separately by two authors working independently, with a third reviewer resolving con icts when they arise. Subsequently, the full-text versions of all potentially eligible studies will be retrieved and reviewed for eligibility, independently, by two members of the team (with a third review resolving any con icts) using a priori eligibility criteria. These are included in Table 1 below. Qualitative studies and economic evaluations will be included if they are conducted as part of a qualifying study and will be used only to generate hypotheses, inform us about the interventions and populations, and inform or deepen our understanding of the quantitative ndings.

•
Studies without a valid counterfactual, including designs that do not include a parallel cohort that establish or adjust for baseline equivalence, including:

Population
Youth aged between 16 and 25 Youth in OOHC for reasons of child maltreatment, neglect or risk of child maltreatment, relinquishment, or lack of provision of support. OOHC settings include: • foster care • guardianship Youth in OOHC settings for reasons other than child maltreatment, neglect, risk of child maltreatment, relinquishment, or lack of provision of support.
Youth who are not in OOHC.
Youth who are currently incarcerated, including in youth justice settings.
Youth aged less than 16 and greater than 25.
• residential care • congregate care Intervention Interventions, programs or services delivered in the home / community.
Interventions, programs or services delivered in other settings, for example: custodial settings.
Comparison Treatment as usual, another intervention, no intervention, or wait-list control.
Studies using other comparators.

Outcome
Primary Outcomes: • Homelessness • Health • Education • Employment • Exposure to violence (as either victim or perpetrator) •

Risky behaviour
Secondary Outcomes: • Relationships

Life skills
Studies looking at other outcomes.

Setting
Countries where a statutory care system for child maltreatment exists.
Countries where a statutory care system for child maltreatment does not exist.

Data extraction
Data extraction will be undertaken by teams of two reviewers, with one reviewer checking the results of the other. Data will be extracted into an online form developed for this review. The following information will be obtained: study design, study outcomes, sample size, location, study timeframe, study population, population demographics, study timeframe, intervention details, effect sizes, implementation outcomes as well as information required to conduct risk of bias assessments.

Data analysis
In accordance with Campbell and Cochrane guidelines, we will try to maximise the likelihood of being able to quantitatively synthesise primary and secondary outcomes across studies using a meta-analysis.
If more than one study with suitable data can be identi ed, we intend to perform a meta-analysis if the population, intervention, comparison, and outcomes are similar enough to be reasonably combined or can be standardised for comparison. This analysis will be conducted using the R Project for Statistical Computing (32).
For each outcome, we will explore heterogeneity by preparing box plots, forest plots and examining the I 2 statistic. If limited heterogeneity is found, we will use a xed effects model. If there is signi cant heterogeneity, we will use a random effects model. These models will allow us to estimate the pooled effect size and its 95% con dence interval for each outcome. For both approaches we will perform a meta-analysis using the longest follow-up period possible for each outcome.
Data from randomised and non-randomised trial designs will not be pooled. Nor will we pool data from non-randomised studies of different study designs.
If the conditions for a meta-analysis are not met, we will describe and synthesise trial ndings narratively and/or use descriptive measures (median effect size and range) to describe their effects.
We will synthesise data around each of our primary outcomes (homelessness, health, education, employment, exposure to violence and risky behaviour). If su cient studies are identi ed, subgroup analysis will be undertaken based upon the following characteristics: a) study design: randomised control trials vs. non-randomised control trials, b) age at which statutory out of home care support ceases: 18 vs. greater than 18 and c) gender.
For binary outcomes -we will calculate a risk ratio and a 95% con dence interval. For continuous dataif a consistent outcome measure is used, we will calculate the mean difference.
Where different measures are used to examine primary outcomes, we will calculate the standardised mean difference if possible.
Where data from the same outcome are reported in different studies as dichotomous or continuous data, we will transform these (if appropriate) to enable pooled estimates of effect. If missing/illegible we will use the Campbell Collaboration ES calculator (33).

Evidence assessment
Risk of bias of included randomised controlled trials will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool (34). Non-randomised studies will be assessed using the ROBINS-I tool (35). The quality of any economic evaluations will be assessed using guidelines suggested by the Cochrane and Campbell Economics Methods Group (36). Outcomes across studies will be assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) methodology (37).

Discussion
Re ecting its place as an emergent policy issue, the volume of research published on care leavers has increased substantially in the last twenty years, but this literature has not been the subject of high-quality synthesis. Therefore, it is an appropriate time for a systematic review to inform ongoing policy and practice discussions about how best to support young people transitioning from care and to shape future research.
This review has a number of advantages over previously published work on this topic, including a) a systematic search of a wide range of databases with no language restrictions, b) the use of inclusion criteria that prioritise studies with high-quality methodologies, c) the selection of outcomes that re ect health and psychosocial wellbeing, d) the inclusion of policy alongside individual or group-level interventions and e) the use of the GRADE approach to assess the quality of evidence, report results and support knowledge translation.
The results of this review will inform policy discussions in this area in a number of jurisdictions, as they evaluate the merits of providing different modes of transitions support, raising the age at which OOHC can be provided to young people or a combination of both. Availability of data and materials Not applicable.