Improving engagement with services to prevent Sudden Unexpected Death in Infancy (SUDI) in families with children at risk of significant harm: A systematic review of evidence

This paper reports part of a wider systematic review commissioned by the English National Safeguarding Panel on Sudden Unexpected Death in Infancy (SUDI). The wider review covered three areas: interventions to improve safer sleep practices in high-risk families, interventions to improve engagement with services and decision making by parents at high risk of SUDI about infant sleep environments. Here, we report the qualitative and quantitative studies reviewed under the engagement strand. Parental engagement is understood to be a multidimensional task for health and social care professionals comprising attitudinal, relational and behavioural components. Following a PROSPERO registered systematic review synthesizing the three strands outlined, 28 papers were found to be relevant in the review of interventions to improve engagement with services in families with children at risk of significant harm through abuse or neglect. No studies were found that specifically focused on engagement of families at high risk for SUDI, so these wider engagement studies were included. The different types of intervention reported in the included studies are described under two broad themes: Enablers (including parental motivation and working with families) and Barriers. Given the focus in the studies on interventions that support parental engagement, the Enablers theme is more extensive than the Barriers reported although all studies noted well-understood barriers. The evidence underpinning these interventions and approaches are reviewed in this paper. We conclude that effective engagement is facilitated by experienced professionals given time to develop supportive non-judgemental relationships with families in their homes, working long-term, linking with communities and other services. While these conclusions have been drawn from wider studies aimed at reducing child maltreatment, we emphasize lessons to be drawn for SUDI prevention work with families with children at risk of significant harm.


| INTRODUCTION
Each year in England and Wales, there are around 300 cases of Sudden Unexpected Death in Infancy (SUDI) (Office for National Statistics, 2020), SUDI is defined as the death of an infant that was not predicted as a possibility in the 48 h prior to the death or to the collapse that led to death irrespective of the final diagnosis (Fleming et al., 2000). SUDI may be due to acute medical conditions or external causes (accidents and injuries), but most remain unexplained and are labelled as Sudden Infant Death Syndrome (SIDS) (International Classification of Diseases version 10 ICD-10 R95), although many are certified instead as 'unascertained' (ICD-10 R99). There are specific hazardous sleep circumstances increasing the risk for SIDS: infants sleeping next to carers who smoke, who have consumed alcohol or drugs or who share inappropriate surfaces such as sofas (Blair et al., 2009;Rechtman et al., 2014).
Since the 1990s, there have been highly successful safe sleep campaigns, and the rate of unexplained infant deaths declined dramatically from a peak of 2.3 per 1000 live births in 1988 to 0.50 in 2004 (Garstang & Pease, 2018) to 0.30 in 2018, equivalent to around 200 unexplained infant deaths annually (Office for National Statistics, 2020). The proportion of SIDS cases occurring in socially deprived families has however increased with the decline in overall numbers (Blair et al., 2006), with highest rates in mothers under 20 years old (Office for National Statistics, 2020). In England when a child dies or suffers significant harm from abuse or neglect, a Child Safeguarding Practice Review (CSPR) is held, and the National Child Safeguarding Practice Review Panel notified.
During 2018-2019, there were 40 notifications of SUDI cases with significant safeguarding concerns, most relating to parents co-sleeping with infants in hazardous sleep environments involving parental alcohol or substance misuse. As a result, a national review of SUDI in families where children are at risk of significant harm was undertaken, and we conducted a systematic literature review to support this (Child Safeguarding Practice Review Panel, 2020). The literature review considered qualitative and quantitative research in three separate but linked areas: interventions to reduce the risk of SUDI in high risk families, understanding how high risk parents make decisions around infant sleep and interventions to improve engagement with professionals in families with children at high risk of abuse and neglect.
Parental engagement was considered key as a previous thematic analysis of 27 SUDI cases with significant safeguarding concerns, subject to Serious Case Review (SCR-the predecessor to CSPR) found that non-engagement with health, social care or substance misuse services was a prominent feature in 18/27 families (Garstang & Sidebotham, 2019). This paper reports on the findings on parental engagement interventions. Given the paucity of studies that have specifically focused on engagement of families in SIDS and SUDI risk reduction, the papers included in this part of the review were broadened out to include engagement interventions with families considered to be at high risk of maltreatment more generally. Later in this paper, we look at lessons from these studies for SUDI prevention in families with children at risk of significant harm.

| Engagement of parents in services
Parental engagement is a mantra of child protection practice. Indeed, one of the core values of social work practice in the United Kingdom is that of human relationships and the need to work in partnership with people (National Association of Social Workers, 2017). In the context of child protection, this means engaging families in services and interventions. While this is a well understood value position, the practicalities of how to engage are less clear (Horwitz & Marshall, 2015).
Parental engagement in child protective services has been described as a multidimensional phenomenon, having attitudinal, relational and behavioural components (Platt, 2012). Given that 'Gaining parental cooperation is a fundamental factor affecting social work interventions, treatment and decision-making' (Platt, 2012, p. 138), it is imperative to understand and address the barriers within these components. Platt's (2012) integrated model of parental engagement emphasizes the need for a workable theoretical model to enable critical examination by workers in how they engage parents in services, and he outlines both internal and external determinants that need to be taken into account. Internal determinants are identified as cognitive, affective, behavioural, identity and motivation and volition; while external determinants include circumstances, resources, support, the programme that is used and the worker who delivers the intervention. It is within this multidimensional model that we discuss the included studies below.

| Aim and research questions
The aim of this section of the wider systematic review is to inform recommendations for how professionals can best engage families where children are at risk of significant harm to ensure that safer sleep advice can be clearly understood and embedded in parenting practice. The specific research questions are as follows: 1. What interventions are effective at improving engagement between support services and families with children at risk of significant harm? 2. How do effective interventions to improve engagement between support services and families with children at risk of significant harm work?

Key messages
• Improving engagement with SUDI prevention in families with children at risk of significant harm involves longterm, face-to-face working between parents and professionals they trust. The quality of the relationship between professionals and parents is key to engagement.
• Services should be locally based and easy to access, combining parenting support with other relevant services. The exclusion criteria used were as follows: • Studies describing interventions for the general population with no high-risk targeting (wrong population) • Interventions that took place pre 2005 (too dated) • Studies describing effectiveness of alcohol/drug services without primary outcomes relating to engagement (no engagement outcomes) • Studies based in countries other than Western Europe, North

| RESULTS
A total of 3506 titles were sourced via the online database searches.
A further 42 studies were sourced using grey literature searches and snowball searches. A PRISMA flow diagram with exact numbers included and excluded at each stage is shown in Figure 1.
There was a 97% agreement rate between reviewers. A total of 28 papers on interventions to improve engagement with support services were included. There were no papers reporting on interventions to improve engagement with families at high risk of SUDI, only papers on families with children at risk of abuse or neglect.

| Description of included papers
There were 17 quantitative and 11 qualitative studies. The vast majority (26/28) were from the United States or Canada, with two from the United Kingdom. The quality of quantitative studies was generally high with 8/17 scoring >75% on QATSDD; qualitative studies quality was mixed with 5/11 scoring >75%. No papers were excluded based on quality assessment. Details of quantitative studies are shown in Table 2 and qualitative studies in Table 3.

| Findings
There were two main themes: enablers and barriers to engagement, subthemes for enablers included working with families, community support and parental motivation. These are illustrated in Figure 2.

| Enablers
Working with families Ease of service access. Ten papers (four qualitative and six quantitative) reported on easy access to services including home visiting. The benefit seen of DR in Cameron and Freymond (2015) may be partly due to DR being offered more locally. Home visiting and regular contact with a case worker improved engagement in a supported housing programme for families with both child protection concerns and unstable living arrangements (Farrell et al., 2012). Families were more receptive to home visiting and home-based delivery of services in a study of Early Childhood Connections (ECC), a service integration process attempting to coordinate an evidence-supported home visiting programme with usual child welfare care (Stahlschmidt et al., 2018).
Comprehensive substance treatment programmes work across systems to co-ordinate health treatment and social care support for  EMP mothers completed all drug court requirements compared to those receiving standard treatment. However, a subsequent RCT (n = 62) showed no benefit of EMP (Dakof et al., 2010). A very small RCT (n = 17) of a comprehensive treatment programme for fathers with substance abuse and domestic violence found no improvement in engagement (Stover, 2015). Interviews with parents from the Strengthening Families programme reported the benefits of overcoming transport barriers and the ability to have separate parent and child groups as well as groups together (Akin et al., 2018). Focus groups with mothers mandated to attend treatment programmes identified strategies supporting engagement including mother-child residential programmes, specific parent-child therapies, material support and treatment workers acting as advocates (Seay et al., 2017).
Giving parents more control. The Nurse-Family Partnership (NFP) (Olds, 2006) is an evidence-based home visiting programme for lowincome, first-time parents consisting of 64 scheduled visits following manualized guidelines. Postulating that high drop-out rates may be partly due to the programme's inflexibility, Ingoldsby et al. (2013) conducted a quasi-experimental study (n = 2419) to give parents more control over programme delivery and frequency. They found a lower risk of drop-out for mothers compared to control, but the effect size was small with an increase of only 1.4 visits on average (95% CI, 0.58-2.2), p < 0.001.

Parental motivation
Parent advocates/mentors. Three papers reported on using parent advocates to support families with child protection proceedings to improve parental engagement. An evaluation of an advocacy scheme for English families (n = 52) subject to child protection proceedings reported that 13/18 families giving feedback found advocacy helpful; 66% of social workers and 79% of child protection conference chairs felt advocates increased parental engagement (Featherstone & Fraser, 2012). A small-scale study reviewed parents' perspectives on parent advocates within the Child Welfare Organizing Project (CWOP) in New York City (Lalayants, 2013). It reported that the  (Forrester et al., 2018).
Two studies combined MI with SafeCare, a modular hometreatment programme for child abuse and neglect (Edwards & Lutzker, 2008). A RCT (n = 398) compared SafeCare with MI (SC+) to standard services for families with children considered to be at high risk of maltreatment using data from service records. There was a highly significant benefit from SC+ with 80% enrolment compared with 49% enrolment in standard services (OR: 4.3, 2.6-7.0, p < 0.001) and 50% completing the programme compared with 21% receiving standard care (OR: 8.5, 3.3-22.1, p < 0.001) (Damashek et al., 2011).
A similar RCT (n = 105) compared SC+ with community-based mental health services and found SC+ families attended more treatment sessions with a mean of 36 h total compared with 8 h for control (p < 0.001) but less than half of SC+ participants completed the programme .

| Interventions with limited impact
A RCT of SafeCare alone (n = 1305) compared with usual services showed a small improvement in parental engagement (Damashek et al., 2012), although another study reported low completion rates (Bolt, 2015). There were two RCTs of technology assisted interventions, neither of which found any significant benefit, one (n = 19) of a home-visiting parenting programme with text reminders (Bigelow et al., 2008) and one (n = 159) of online parenting training (Baggett et al., 2017).

| Barriers to engagement
All the qualitative studies noted barriers to parental engagement in programmes including lack of motivation or recognition of need to change, limited time with children allowed for in programmes, programmes that were not perceived to be culturally applicable beyond white families, feeling pre-judged and blamed. Being mandated to attend programmes was a significant negative factor for many families (Akin et al., 2018;Seay et al., 2017), but despite this in a study of the Positive Parenting Programme (Triple P), parents still reported gaining parenting skills (Garcia et al., 2018). Parents in Lalayants (2013) study spoke of stigma, negative attitudes of child welfare workers and the mistrust and prior negative experiences of child protection services as key barriers to engagement. For some parents, their children became inhibitors to their programme engagement as they did not want to be away from their child in order to attend, could not manage parenting and programme requirements, and suffered overwhelming guilt about their child once they commenced programmes (Seay et al., 2017).

| DISCUSSION
This review found limited evidence for interventions to improve engagement in families with children considered to be at risk of significant harm. Of the interventions that showed some benefit, these were all face-to-face programmes with high intensity family contact and close working and co-ordination between agencies and community support. The quality of the relationship between a skilled professional and family is key to engagement for meaningful change; but this is not something that can be achieved in the short term. The wider engagement literature acknowledges that engagement is much more than just engaging in information giving or ensuring compliance of parents (Platt, 2012), and this was very much evident in the studies reviewed where it is important to note engagement as a concept was not consistently defined. Services needed to be easy for families to access, with practical support such as programmes that combined sub- and only two of these RCTs with a clinically significant increase in engagement. Improving the engagement of vulnerable families is challenging and resource intensive. The most effective practices will involve professionals working with families regularly, over long periods of time, starting in the neonatal period, to build trusted relationships; and for professionals and families to be linked with community-based support services. An approach that focuses on the wider needs of the family including housing and mental health needs is also important. Our review highlights the importance of relationship-based practice (Ruch, 2005) and the characteristics of these relationships reported to be important: trust based, nonstigmatizing, and non-judgemental whilst bringing an 'integrated understanding of the individual-structural causes of social distress and dysfunction' (Ruch, 2005, p. 114) and a focus on reducing anxiety (for the worker and the parent). This demands a high level of reflective practice and a much more complex perspective on engagement and intervention. Practicing reflective relationship-based practice with clients who may be experiencing extremely challenging life circumstances is not always straightforward (Ruch, 2005) and a model of parental engagement that recognizes the relational, attitudinal and behavioural components in parents and in workers is essential (Platt, 2012). Applying this to SUDI risk practice may help with understanding non-engagement even when parents have access to clear information. For example, nearly half of the families in the SUDI Serious Case Review study had received safe-sleep advice from Health Visitors, but this had not been acted on by parents (Garstang & Sidebotham, 2019), so regular contact with trusted professionals who are able to practice in relationally warm and supportive ways could improve the uptake of safe-sleep messages. Indeed, a recent Canadian study of parental engagement in child protection services (Charest-Belzile et al., 2020) has demonstrated that high-quality relationships between parents and worker strongly predicts for a more positive attitude during intervention. Future interventions to improve the uptake of safer sleep advice in families with infants at risk of SUDI should consider how to maximize on the benefits of these positive relationships between families and health professionals.
Two technology assisted interventions were not found to be effective, however it is worth noting that these types of 'digital health' options are becoming more popular, particularly within the UK response to the SARS-COV-2 global pandemic. There may be a longterm impact from the pandemic on how personal services are provided to families with young children in future-services may be less likely to be face to face, more linked to the use of technology such as mobile phones or phone apps, and to be more targeted to families with particular needs or risks. Studies showing the relative ineffectiveness of technology-based interventions may need to be repeated using better population selection and better technology. The rapid changes in communications seen during the pandemic may mean that some of the features of digital health interventions that made them less effective in the past may be less of an impediment in the future.
Conversely, over-reliance on technology at the expense of trusted relationships with professionals may lead to less engagement by vulnerable families and greater risk of SUDI as parents either cannot access, do not trust or fail to implement safer sleep advice.

| Conclusions
This review forms part of a wider review and project to review practices and make recommendations about how best to increase uptake of safer sleep advice in families with children at risk of significant harm (Child Safeguarding Practice Review Panel, 2020). We found no engagement studies that specifically included this population, although several used overlapping risk factors. We conclude that effective engagement is facilitated by experienced professionals given time to

AUTHOR CONTRIBUTIONS
All authors contributed to the study design. The literature searches, screening and data extraction were conducted by J.G., D.W., A.P. and CE. The narrative synthesis was carried out by J.G. and D.W. The first draft was written by J.G.; all authors commented and contributed to subsequent revisions. All authors have reviewed and approved the final manuscript.

ETHICAL APPROVAL
This systematic review involved secondary analysis of existing research so no ethical approval was required.

DATA AVAILABILITY STATEMENT
There are no further data available concerning this systematic review.