Open access Protocol

BMJ Open: first published as 10.1136/bmjopen-2019-031362 on 11 September 2019. Downloaded from http://bmjopen.bmj.com/ on December 1, 2022 by guest. Protected by copyright.
Interventions and practice models for
improving health and psychosocial
outcomes of children and young people
in out-of-home care: protocol for a
systematic review
Emma Galvin,‍ ‍ 1
Renée O'donnell,1 Helen Skouteris,1,2 Nick Halfpenny,3
Aya Mousa1

To cite: Galvin E, O'donnell R, Abstract

Skouteris H, et al. Interventions Strengths and limitations of this study
Introduction  Children and young people placed in
and practice models for out-of-home care (OoHC) are often affected by a history
improving health and ►► The current study employs rigorous international
of trauma and adverse childhood experiences. Trauma
psychosocial outcomes of gold-standard methodology and a comprehensive
in early childhood can impact on children’s health and
children and young people in search strategy.
out-of-home care: protocol for psychosocial development, whereas early interventions
►► Limitations of this study include the potential for
a systematic review. BMJ Open can improve children’s development and placement
publication bias since the systematic review will in-
2019;9:e031362. doi:10.1136/ stability. Although several interventions and practice
clude only published data;
bmjopen-2019-031362 models have been developed to improve health and
►► This study includes the potential that studies may
psychosocial outcomes for children and young people
►► Prepublication history and be too heterogeneous to obtain combined effect
in OoHC, there remains a lack of rigorous research
additional material for this estimates.
paper are available online. To examining the impact of these interventions in OoHC
view these files, please visit settings, as there are no systematic reviews examining
the journal online (http://​dx.​doi.​ the impact these interventions and practice models have
org/​10.​1136/​bmjopen-​2019-​ on the children and young people they serve. We aim to PROSPERO registration number  CRD42019115082.
031362). conduct a comprehensive systematic review to examine
the effectiveness of interventions and practice models for
Received 30 April 2019 improving health and psychosocial outcomes in children Introduction
Revised 19 July 2019 Children and young people in out-of-home
and young people living in OoHC and to identify relevant
Accepted 25 July 2019
knowledge gaps. care (OoHC) are some of the most vulner-
Methods and analysis  Major electronic databases able groups in society, often having experi-
including Medline, Medline in-process and other non- enced substantial harm, abuse or neglect.1 2
indexed citations, Embase, Cumulative Index to Nursing OoHC refers to the short-term or long-term
and Allied Health Literature, PsycInfo, Sociological care of children and young people up to 18
Abstracts and all Evidence-Based Medicine Reviews
© Author(s) (or their
or 21 years of age (depending on country)
incorporating: Cochrane Database of Systematic Reviews,
employer(s)) 2019. Re-use who are unable to live with their families due
American College of PhysiciansJournal Club, Database of
permitted under CC BY-NC. No
Abstracts of Reviews of Effects,Cochrane Central Register to child protection orders and/or as a result
commercial re-use. See rights of parents being unable to provide adequate
and permissions. Published by of Controlled Trials, CochraneMethodology Register, Health
BMJ. Technology Assessment and National Health Service care or protection.3 In Australia, 47 915 chil-
1
School of Public Health and Economic Evaluation Database, will be systematically dren <18 years lived in OoHC in 2017, a
Preventive Medicine, Monash searched for any studies published between 2008 and rate of 8.7 per 1000 children, reflecting an
Centre for Health Research 2018 of interventions and practice models developed to increase from 46 448 and 40 549 children in
and Implementation, Monash improve health and psychosocial outcomes for children 2016 and 2013, respectively.2 4 These rising
University, Melbourne, Victoria, and young people in OoHC. Two independent reviewers rates are concerning, since children and
Australia will assess titles and abstracts for eligibility according
2
School of Business, University
young people placed in OoHC are often
to prespecified selection criteria and will perform data
of Warwick, Coventry, UK characterised as having severe cognitive,
extraction and quality appraisal. Meta-analyses and/or
3
Policy and Research, MacKillop emotional, behavioural and social prob-
metaregression will be conducted where appropriate.
Family Services, Melbourne,
Ethics and dissemination  This study will not collect lems,1 5 coupled with complex histories of
Victoria, Australia
primary data and formal ethical approval is therefore maltreatment and neglect.5 6 This history of
Correspondence to not required. Findings from this systematic review will trauma is believed to have short-term and
Professor Helen Skouteris; be disseminated in a peer-reviewed publication and long-term effects on brain development,
​helen.​skouteris@​monash.​edu conference presentations. from childhood through to adulthood, and

Galvin E, et al. BMJ Open 2019;9:e031362. doi:10.1136/bmjopen-2019-031362 1

Open access

BMJ Open: first published as 10.1136/bmjopen-2019-031362 on 11 September 2019. Downloaded from http://bmjopen.bmj.com/ on December 1, 2022 by guest. Protected by copyright.
often culminates into complex behavioural, psycholog- and case managers’ best efforts to provide this informa-
ical and social challenges.5 6 Children and young people tion, the high percentage of placement breakdowns and
in OoHC report poorer outcomes across a number of constant change in caregivers and service providers create
health and well-being indicators compared with those gaps in information pertaining to the individuals’ social
who remain with their biological family.7 Since children or family circumstances and medical and mental health,
and young people usually enter care having experienced and there is risk of this information being lost.6
trauma and neglect, deviant behaviour and mental Most interventions have also not been properly evalu-
health problems are particularly prevalent among chil- ated, and there remains a lack of rigorous research exam-
dren and young people in OoHC, and this appears to ining the impact of these interventions in improving
be associated with both age at first placement and type health and/or psychosocial outcomes for children and
of care.8 A 2006 study found that up to 60% of children young people in OoHC.21 Indeed, a recent systematic
and young people in OoHC have a current mental health review investigating the empirical evidence of trauma-in-
diagnosis including depression, attachment and conduct formed, organisation-wide models implemented in resi-
disorders, and attention deficit hyperactivity disorders,9 dential OoHC settings identified three models including
and those placed in residential care tend to have higher The Sanctuary Model, Children and Residential Expe-
rates compared with those in foster care, whereas indi- riences programme and the Attachment Regulation
viduals in kinship care report the fewest rate of mental and Competency framework, and concluded that the
health disorders.8 10 Children and young people in OoHC evidence base is limited, making it difficult to accurately
also report a significantly higher incidence of substance evaluate outcomes of trauma-informed models.22 Recent
abuse, suicide ideation and suicide attempt,11–13 as well studies have outlined the health and psychosocial needs
as attachment difficulties, problematic sexual behaviour, of children and young people in OoHC and the inter-
eating disorders, delinquent behaviour and reduced ventions and practice models that have been designed to
educational attainment compared with children and meet these needs; however, to date, very little research
young people residing with their biological families.8 14 has been focused on evaluating the effectiveness of these
The poor outcomes that children and young people in interventions.15 To our knowledge, no previous system-
OoHC face, some of which are mentioned above, play a atic reviews have examined interventions or practice
significant role in the complexity of their overall health, models designed to respond to the physical and psycho-
making it more challenging to identify their health needs social health needs of children or young people in resi-
and develop appropriate health management plans. dential, foster and kinship settings of OoHC. To develop
Evidently, children and young people in OoHC require effective, evidence-based interventions in OoHC, we first
more intensive intervention and support, as we are need to understand which interventions and practice
dealing with complex, multifaceted issues, that require models work and how their structures and processes can
a number of strategies that can effectively support their be implemented and sustained in practice. Otherwise,
health and well-being. the cycle of disadvantage and poor health outcomes
In light of the poor health and psychosocial outcomes will not be broken, and children and young people in
experienced by children and young people living in OoHC will remain at increased risk of adverse health and
OoHC, effective and sustainable interventions for psychosocial outcomes. To this end, we aim to conduct
improving these outcomes are urgently needed. Over a comprehensive systematic review which will: (1) assess
the last few years, a number of practice models and inter- the effectiveness of interventions and practice models for
ventions have been developed with the aim of directly improving the health and psychosocial outcomes of chil-
addressing the impact of trauma on health and psychoso- dren and young people living in OoHC (all types of place-
cial outcomes for children and young people in OoHC.15 ments); (2) examine whether a particular intervention or
Some of these models, such as the Sanctuary Model, practice model is more effective than another and (3)
Therapeutic Residential Care and Treatment Foster Care delineate which components of these interventions are
are shifting towards needs-based care and incorporating associated with the greatest improvement in outcomes for
a trauma-informed, therapeutic care approach within the these children and young people.
OoHC placements.16 In Australia, the UK and the USA,
it is expected that children and young people entering Systematic review questions
OoHC have the appropriate health assessments (statu- ►► Are certain interventions or practice models effective
tory) and that healthcare records and management plans in improving health and/or psychosocial outcomes
are in place.17–20 Unfortunately, this process is not always for children and young people in OoHC compared
met, and even when a child is provided with a health with usual care?
management plan, their needs may not be incorporated ►► Are certain interventions or practice models more
and the plan may not be followed.18 Collecting the neces- effective than others in improving health and/or
sary information can be difficult as health professionals psychosocial outcomes for children and young people
must rely on parents to provide medical histories, explain in OoHC?
health and behavioural concerns and consent to the ►► Which elements are critical in determining the success
assessment and treatment of their child.18 Despite carers’ of interventions, and for whom?

2 Galvin E, et al. BMJ Open 2019;9:e031362. doi:10.1136/bmjopen-2019-031362

 Open access

BMJ Open: first published as 10.1136/bmjopen-2019-031362 on 11 September 2019. Downloaded from http://bmjopen.bmj.com/ on December 1, 2022 by guest. Protected by copyright.
Table 1  PICO for study inclusion
Participants (P) Intervention (I) Comparison (C) Outcomes (O) Study type Limits
Inclusion Children Any intervention No intervention; All health and Randomised English
criteria 0–21 years of (eg, treatment usual care; other psychosocial controlled trials; language only;
age and living in foster care; interventions in outcomes including non-randomised or peer-reviewed;
OoHC therapeutic OoHC; but not limited to: uncontrolled trials; published in the
►► School aged residential care, children who intellectual; systematic reviews; last 10 years
►► Youth sanctuary model, remain with their behavioural; cohort studies; (2008–2018).
►► Adolescents etc) delivered in biological/ foster psychosocial; cross-sectional;
►► Children an OoHC setting families. mental; suicidal longitudinal.
►► Infants (eg, foster care, ideation;
kinship care, psychological
residential care). functioning; social
skills; emotional;
educational
attainment;
relationships; illicit
drug use; smoking;
alcohol; eating
disorders.
Exclusion Adults Adoption; Editorial; Literature
criteria >21 years of age rehabilitation; commentary; published
orphanages narrative review; before 2008;
expert opinion Languages
other than
English.
OoHC, out-of-home care; PICO, Population, Intervention, Comparison, Outcomes.

Methods and analysis Reviews, American College of Physicians Journal Club,

This systematic review uses rigorous international gold Database of Abstracts of Reviews of Effects, Cochrane
standard methodology23 24 and conforms to the reporting Central Register of Controlled Trials, Cochrane Method-
standards of the Preferred Reporting Items for Systematic ology Register, Health Technology Assessment, National
Reviews and Meta-analyses25 (PRISMA; see online supple- Health Service Economic Evaluation Database. Bibliog-
mentary file). raphies of relevant studies as well as systematic reviews
identified by the search strategy will be screened for iden-
Selection criteria
As outlined in table 1, a Population, Intervention, tification of additional studies. Where required data are
Comparison, Outcomes framework was established a not presented, the corresponding authors of included
priori to screen studies and determine their eligibility studies will be contacted to provide deidentified aggre-
for inclusion in the systematic review. Interventions and gate data for the purpose of meta-analyses if deemed
practice models developed for reunification have been necessary.
excluded; however, interventions and practice models
that incorporate participants who transitioned into a
different type of care setting are included. Box 1  Sample of search terms used in electronic search
Search strategy
Concept 1: Out of home care
A systematic search, based on the selection criteria Foster care, foster, out of home, kinship, trauma informed, resident*,
(table 1) and combining medical subject heading terms guardian care, family based care, family centered, home based, child
and text words, was developed using the Ovid platform protection, child welfare, non biological care, group home, group house,
and translated to other databases as appropriate (see ‘OoHC’
online supplementary file). The search terms are outlined
Concept 2: Participants
in box 1.
Looked after child*, young person, young people, infan*, baby, babies,
Relevant articles will be sourced through electronic toddler, preschool*, adolescen*, teen*, minor, youth
databases including: Medline, Medline in-process and
other non-indexed citations, Embase, PsycINFO, Cumu- Concept 3: Intervention
lative Index to Nursing and Allied Health Literature, Model, outcome, evaluation, framework, theor*, intervention, program*,
process*, prevention, treatment, strategy*, therap*, trauma informed,
Sociological Abstracts and all Evidence-Based Medicine
trauma focused, trauma service
Reviews incorporating: Cochrane Database of Systematic

Galvin E, et al. BMJ Open 2019;9:e031362. doi:10.1136/bmjopen-2019-031362 3

Open access

BMJ Open: first published as 10.1136/bmjopen-2019-031362 on 11 September 2019. Downloaded from http://bmjopen.bmj.com/ on December 1, 2022 by guest. Protected by copyright.
Screening of search results aspects such as risk of bias, inconsistency, indirectness,
Search results will be managed using the Endnote V.X.8.0 imprecision and publication bias. Based on this evalu-
reference management software. Two reviewers will ation, a quality score (high, moderate, low or very low)
assess the titles, abstracts and keywords of every article will be assigned to each outcome. Disagreement will be
retrieved by the search strategy according to the selec- resolved by discussion to reach consensus.
tion criteria described in table 1. Full text of the articles
will be retrieved for further assessment if the informa- Data analysis and synthesis
tion provided suggests that the study meets the selection Data will be presented in summary form and narratively
criteria or if there is any doubt regarding eligibility of the as well as in tables (where possible) to describe the study
article based on the information given in the title and designs, populations and findings and to address each
abstract. Where there is more than one article describing research question. Data will be summarised statistically
the same study and reporting different outcomes, these using meta-analysis of aggregate effect measures if avail-
articles will be combined and considered a single unique able and if studies are deemed sufficiently homogeneous
study. Articles excluded after full-text assessment will be to combine. The meta-analysis will be performed on
tabulated with reasons for their exclusion, as per PRISMA studies in which a baseline and follow-up effect is available
guidelines.26 (ie, randomised control trial and quasi-experimental) and
wherein the same outcome of interest has been reported
Data extraction (ie, anxiety, depression, self-harming behaviour, delin-
Two independent reviewers will formally screen the quent behaviour, obesity) along with a change in effect.
included studies against the selection criteria and As the outcomes of interest will likely be assessed using a
perform data extraction using a specifically designed data diverse range of instruments, a random effects model will
extraction form. Extracted data will include general study be estimated accounting for the heterogeneity between
characteristics (author, year, country, setting, inclusion/ the studies. Review Manager V.5 software will be used
exclusion criteria), population characteristics (gender for meta-analysis, and results will be expressed as relative
and age distribution and other relevant features), risks or ORs with 95% CIs for dichotomous outcomes and
intervention and control characteristics (type/model, weighted mean differences with 95% CI for continuous
duration, frequency and so on), outcome assessments outcomes. Statistical homogeneity will be assessed using
(physical and psychosocial outcomes and tools used to the I2 test where I2 values over 50% indicate moderate to
assess these) and results (point estimates and measures of high heterogeneity.28 Statistical significance will be set at
variability for continuous outcomes and frequency counts a two-tailed p<0.05. For studies with qualitative designs or
or absolute numbers of episodes or relative measures of have insufficient data for pooling, a descriptive analysis
risk (risk ratio or ORs with CIs) for dichotomous vari- will be presented.
ables, numbers of participants, intention-to-treat analysis)
and any other relevant validity results. Missing data will be Subgroup analysis
obtained from corresponding authors wherever possible, Subgroup analysis, and where appropriate, meta-regres-
and two reviewers will check all computed data entries sion will be performed if possible based on study char-
for meta-analysis if applicable. Any disagreement will be acteristics and results from the search. Where there is
resolved by discussion to reach a consensus. sufficient data, these analyses will be conducted based
on prespecified subgroups/covariates including age at
Assessment of risk of bias and quality of the evidence placement, age at intervention, gender, ethnicity (indig-
Methodological quality of included studies will be enous vs non-indigenous), placement type (residential
assessed at the study-level by two independent reviewers or group home vs foster family and kinship vs non-kin-
using a risk of bias assessment template according to study ship placement), types of abuse/reason for placement
design. Individual quality items will be investigated using (maltreatment/abuse vs behavioural problems), types
a descriptive component approach which will include of intervention (psychological, social, behavioural),
assessment of key aspects such as methods of outcome duration of intervention and length of follow-up. Other
assessment and reporting, statistical analysis components factors presumed to cause variations in the outcomes may
including study power and dealing with confounding, be determined during the review process, and these will
attrition rates and conflicts of interest of authors. Using be included in additional exploratory subgroup analyses.
this process, a risk of bias rating (high, moderate or low)
will be assigned to each study. Sensitivity analysis
Quality of the evidence for the effects of interventions Sensitivity analysis will be performed to explore the
in improving health and psychosocial outcomes for chil- influence of heterogeneity (I2>50%) and determine the
dren and young people in OoHC will be assessed by two robustness of the observed effect size. Specifically, the
independent reviewers using the Grading of Recom- primary analysis will be repeated by altering the dataset to
mendations, Assessment, Development and Evaluations only include medium and high-quality studies to examine
framework.27 This will be used to appraise quality at the their influence on the results. If the findings are robust,
outcome level and, where appropriate, will incorporate then the studies of all quality will be retained, if there are

4 Galvin E, et al. BMJ Open 2019;9:e031362. doi:10.1136/bmjopen-2019-031362

 Open access

BMJ Open: first published as 10.1136/bmjopen-2019-031362 on 11 September 2019. Downloaded from http://bmjopen.bmj.com/ on December 1, 2022 by guest. Protected by copyright.
changes in the findings, then further examination of this understanding of interventions which improve health
will be performed. Where there are sufficient numbers and psychosocial outcomes for children and young
of studies, visual inspection of funnel plots and Begg and people in OoHC and the key contributing factors within
Egger et al29 30 statistical tests will be used to assess publi- these interventions. These findings will be disseminated
cation bias and small study effects. through peer-reviewed publications and at conference
meetings to inform future research and to guide the
Patient and public involvement development and real-world implementation of sustain-
This systematic review will not collect primary data, and able interventions in OoHC settings.
therefore patients and the public were not involved in the
design, conduct or reporting of the research. Acknowledgements  EG is supported by a PhD scholarship provided by MacKillop
Family Services. AM is supported by an Early Career Fellowship provided by
the National Health and Medical Research Council (NHMRC) of Australia. HS is
supported by a NHMRC senior research fellowship. We thank Dr Marie Misso for her
Discussion input and expertise in developing the search strategy and helping to run the initial
Children and young people in OoHC have typically been database searches.
exposed to a multitude of psychologically distressing Contributors  EG developed the search strategy, wrote the first draft of the review
and adverse experiences that manifests into childhood protocol and lead the data collection and analysis. AM contributed to the design
and scope of the search strategy, guided the review process and revised and
trauma.31 Childhood trauma is an important public edited the manuscript. RO and NH contributed to the revision and editing of the
health concern as adverse childhood experiences can manuscript. HS determined the design and scope of the review, revised and edited
have substantial health, social and economic implica- the manuscript, supervised in the review process and is the guarantor for ensuring
tions which extend throughout the lifespan.31 Therefore, the integrity and accuracy of the review data.
there is a need for health and psychosocial interventions Funding  The authors have not declared a specific grant for this research from any
to be implemented to prevent further traumatic and funding agency in the public, commercial or not-for-profit sectors.
adverse childhood experiences as early as possible, as Competing interests  None declared.
these interventions may reduce the negative outcomes Patient consent for publication  Not required.
of adverse childhood experiences. Existing interventions Provenance and peer review  Not commissioned; externally peer reviewed.
and practice models aim to directly address the impact Open access  This is an open access article distributed in accordance with the
of trauma on a child’s health or psychosocial outcomes, Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which
typically through trying to reduce symptoms or facilitate permits others to distribute, remix, adapt, build upon this work non-commercially,
recovery.31 32 However, many of these interventions have and license their derivative works on different terms, provided the original work is
properly cited, appropriate credit is given, any changes made indicated, and the use
not been properly evaluated or have limited evidence is non-commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/.
of their effectiveness in improving the health and/or
psychosocial outcomes for children and young people
in OoHC. To develop effective interventions for those in
OoHC, we need to understand which interventions work
and how their effects can be sustained and embedded (ie, References
1. Wall L, Higgins D, Hunter C. Trauma-informed care in child/family/
implemented) into practice. The proposed systematic welfare services. CFCA paper no 37. Melbourne: Australian Institute
review aims to address these gaps by examining how inter- of Family Studies, 2016.
2. Australian Institute of Health and Welfare. Children protection
ventions and practice models can be applied to organisa- Australia 2016-17. Canberra: Australian Government, 2018.
tions and carers to improve the physical and psychosocial 3. Jones R, Everson-Hock ES, Papaioannou D, et al. Factors
associated with outcomes for looked-after children and young
health of children and young people placed in OoHC. people: a correlates review of the literature. Child Care Health Dev
Using rigorous methodology, prespecified criteria and a 2011;37:613–22.
predetermined search strategy, this review will capture 4. Heyes N, Smart J, Walton B, et al. Children in care. Melbourne:
Australian Institute of Family Studies, 2018.
and synthesise existing quantitative and qualitative 5. Leloux-Opmeer H, Kuiper C, Swaab H, et al. Characteristics of
evidence on interventions in OoHC to establish their children in foster care, Family-Style group care, and residential care:
a scoping review. J Child Fam Stud 2016;25:2357–71.
impact in improving health and psychosocial outcomes 6. Barnett ER, Boucher EA, Neubacher K, et al. Decision-Making
and to disentangle the specific elements which contribute around psychotropic medications for children in foster care:
to their success. Findings from this review will provide perspectives from foster parents. Child Youth Serv Rev
2016;70:206–13.
much needed evidence to build the current knowledge 7. Meltzer H, Corbin T, Gatward R, et al. The mental health of young
base and to inform the implementation of effective inter- people looked after by local authorities in England. London: The
Stationery Office, 2003.
ventions in OoHC in an effort to alleviate the poor health 8. Tarren-Sweeney M. The mental health of children in out-of-home
and psychosocial outcomes of children and young people care. Curr Opin Psychiatry 2008;21:345–9.
in OoHC. 9. Tarren-Sweeney M, Hazell P. Mental health of children in foster and
kinship care in New South Wales, Australia. J Paediatr Child Health
2006;42:89–97.
10. Trout AL, Hagaman J, Casey K, et al. The academic status of children
and youth in out-of-home care: a review of the literature. Child Youth
Ethics and dissemination Serv Rev 2008;30:979–94.
This study does not require ethical approval as it does 11. Evans R, White J, Turley R, et al. Comparison of suicidal ideation,
suicide attempt and suicide in children and young people in care
not involve primary data collection. We anticipate that and non-care populations: systematic review and meta-analysis of
findings from this review will contribute to an improved prevalence. Child Youth Serv Rev 2017;82:122–9.

Galvin E, et al. BMJ Open 2019;9:e031362. doi:10.1136/bmjopen-2019-031362 5

 Open access

BMJ Open: first published as 10.1136/bmjopen-2019-031362 on 11 September 2019. Downloaded from http://bmjopen.bmj.com/ on December 1, 2022 by guest. Protected by copyright.
12. Taussig HN, Harpin SB, Maguire SA. Suicidality among 23. Higgins J, Green S, eds. Cochrane Handbook for systematic reviews
preadolescent maltreated children in foster care. Child Maltreat of interventions: the Cochrane library. Cichester, UK: John Wiley &
2014;19:17–26. Sons, Ltd, 2006.
13. Trout AL, Hagaman JL, Beth Chmelka M, et al. The academic, 24. Apsche JA, Bass CK, Zeiter JS, et al. Family mode deactivation
behavioral, and mental health status of children and youth at entry to therapy in a residential setting: treating adolescents with conduct
residential care. Resid Treat Child Youth 2008;25:359–74. disorder and multi-axial diagnosis. Int J Behav Consult Ther
14. Sullivan DJ, van Zyl MA. The well-being of children in foster care: 2008;4:328–39.
exploring physical and mental health needs. Child Youth Serv Rev
25. Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for
2008;30:774–86.
systematic review and meta-analysis protocols (PRISMA-P) 2015
15. Petersen AC, Joseph J, Feit MN. New directions in child abuse and
neglect research. Institute of Medicine (U.S.), National Research statement. Syst Rev 2015;4.
Council (U.S.), & National Research Council (U.S.), 2014: 245–83. 26. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for
16. McLean S, Price-Robertson R, Robinson E. Therapeutic residential systematic reviews and meta-analyses: the PRISMA statement. PLoS
care in Australia: taking stock and looking forward, 2011. Med 2009;6:e1000097.
17. Bartlett JD, Rushovich B. Implementation of trauma systems 27. Grading quality of evidence and strength of recommendations. BMJ
Therapy-Foster care in child welfare. Child Youth Serv Rev 2004;328.
2018;91:30–8. 28. Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring
18. Vimpani GV, Webster SM, Temple-Smith MJ. Improving the health inconsistency in meta-analyses. BMJ 2003;327:557–60.
of Australian children entering out-of-home care. Med J Aust 29. Begg CB, Mazumdar M. Operating characteristics of a RANK
2012;196:91–2. correlation test for publication bias. Biometrics 1994;50:1088–101.
19. Hill CM, Watkins J. Statutory health assessments for looked-after 30. Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis
children: what do they achieve? Child Care Health Dev 2003;29:3–13. detected by a simple, graphical test. BMJ 1997;315:629–34.
20. Mooney A. Promoting the health of Looked after children: a study
31. Gimson K, Trewhella A. Developing a Trauma-Informed therapeutic
to inform revision of the 2002 guidance. Department for Children,
service in the Australian Capital Territory for children and young
Schools and Families (DCSF), 2009.
21. Barth RP, Greeson JKP, Zlotnik SR, et al. Evidence-based practice for people affected by abuse and neglect. Australian Capital Territory
youth in supervised out-of-home care: a framework for development, Government: Australian Capital Territory, 2014.
definition, and evaluation. J Evid Based Soc Work 2011;8:501–28. 32. Australian Centre for Posttraumatic Mental Health and Parenting
22. Bailey C, Klas A, Cox R, et al. Systematic review of organisation- Research Centre. Approaches targeting outcomes for children
wide, trauma-informed care models in out-of-home care (OoHC) exposed to trauma arising from abuse and neglect - Evidence,
settings. Health Soc Care Community 2019;27:e10–22. practice and implications, 2013.

6 Galvin E, et al. BMJ Open 2019;9:e031362. doi:10.1136/bmjopen-2019-031362