Young carers, mental health and psychosocial wellbeing: A realist synthesis

Growing evidence demonstrates that the mental and psychosocial health impacts of caring vary significantly for individual children, depending on who they are, the person that they care for, their responsibilities and the wider family situation. Although individual studies have made progress in identifying the range of impacts, there is a lack of clarity around which impacts affect who and in what circumstances. This synthesis, based on RAMESES realist protocols, aims to increase clarity concerning how and why the mental and psychosocial health impacts of caring for a family member vary for different children. There were 12 391 unique search results screened at title, abstract and full-paper levels. Forty-four retained studies were analysed, resulting in the development of a model with 17 context-mechanism-outcome configurations. The model divides the configurations into three interlinking domains. The caregiving responsibilities domain considers how the impacts of caring vary with the circum-stances of the individual young carer, the person they care for and their family. The identity domain details the development of a caring identity that potentially mitigates the negative effects of caring and enables positive benefits. The support domain concerns the support provided from family, community and services that, depending on quality, can mitigate or exacerbate the impacts of caring. Support also moderates the care identity by affecting self-perception of the caring role. The model has the potential to inform the development of interventions that target particular mechanisms to enable positive change for young carers. This potential can be enhanced by further research to test the model, with a focus on refining configurations where less evidence is available. There is a particular need to focus on identification which is under-represented in the model as both a mechanism and a contextual factor due to unidentified young carers being largely absent from past research.

S e e h t t p://o r c a .cf. a c. u k/ p olici e s. h t ml fo r u s a g e p olici e s.Co py ri g h t a n d m o r al ri g h t s fo r p u blic a tio n s m a d e a v ail a bl e in ORCA a r e r e t ai n e d by t h e c o py ri g h t h ol d e r s .

| INTRODUCTION
Early young carer research focused on the lives, experiences and needs of children undertaking primary care for adults in the home (Aldridge & Becker, 1993).Following that, research prioritized how to best identify those most in need of support, resulting in different approaches to measuring level of care including a focus on 'substantial' care (Department of Health, 1995), assessing the appropriateness of particular caring responsibilities (Aldridge, 2006;Andersen, 2012) and looking at impact irrespective of responsibilities (UK Government, 2014;Warren, 2007).There has been an increasing focus on caring for people with particular illnesses and disabilities, including substance misuse (Moore et al., 2011) and mental health issues (Aldridge, 2006), with an internationalization of research to include different countries and cultures.It has become increasingly clear that young carers are a non-homogenous group of children taking on a range of caring responsibilities for different family members, while receiving varying levels of support from individuals and services.
The 1990s research impacted UK policy and led to the development of young carers provision offering support and respite.Although there is evidence that these services are valued and beneficial (Aldridge, 2006), there has been little diversification of services to support the specific needs of different young carers.Consequently, there is a need to incorporate the research evidence into a structured understanding of why the positive and negative impacts vary for different young carers, to inform what young carers' differing support needs are.

| Objectives and focus of review
The primary research objective is to clarify how the experiences of caring impact individual children differently depending on their responsibilities but also their wider lives and family circumstances.
This will be achieved by identifying and incorporating evidence from previous studies to develop a model relating to young carers and mental health.Additional objectives concern whether there is sufficient evidence underpinning the different components of the model to ensure reliability, and therefore have confidence in the model's potential to inform service development.

| Rationale for using realist synthesis
Realist syntheses utilize previous research to conceptualize a social programme through constructing a series of contextmechanism-outcome (CMO) configurations (Pawson & Tilley, 1997).
Each configuration enables the witnessing of transformation in the A traditional realist synthesis of young carers support would need to incorporate research concerning both mainstream services (schools, health and social services) and specialist projects.Although young carer projects have been studied in the past (Coles et al., 2007;Cunningham et al., 2017;Fraser & Pakenham, 2009;Richardson et al., 2009;Schlarmann et al., 2011), neither intervention evaluations or wider research have considered provision by mainstream services in terms of what works and for whom in any detail.
There is however a growing focus not on interventions but on evaluations of the pre-existing social context (de Souza, 2013).
Through developing a better understanding of this pre-existing social context, the accompanying mechanisms and the resulting outcomes, we can better understand the change that is needed and develop suitable interventions.Therefore, this synthesis will focus on the pre-existing context and mechanisms relevant to young carers, and the resulting negative impacts and benefits.
The procedure used to identify, screen and analyse papers to develop the model follows the RAMESES realist protocols (Wong et al., 2014).

| Scoping the literature
Prior to the realist synthesis, a concept analysis had been developed to define young carers as follows: Children and young persons under 18 who provide, or intend to provide, care, assistance or support to another family member (or members) due to an illness or disability, mental health or substance misuse issue.
They assume a level of responsibility which may, depending on the support that is in place both within

Key messages
• Realist synthesis clarifying how and why the mental and psychosocial health impacts of caring vary for different children.
• Resulting model has three domains relating to the child's caregiving responsibilities, support received from different levels of society and the development of a carer identity.
• A positive carer identity mitigates the negative impacts of caring and enables benefits.
• Positive support mitigates the impacts of caring responsibilities and aids development of a positive carer identity.
and from outside the family, have an impact on their lives.
The concept analysis had also identified antecedents and consequences for young carers, with these used to initiate theory-building.
The concept analysis and definition informed the development of inclusion criteria (Table 1) to support development of the search strategy and subsequent screening of search results.

| Search process
A search strategy was developed with two strands.The main search strand contained two fields relating to young carers and mental and psychosocial health impact.
It was also important to identify research from disability studies research where terminology is significantly different and includes 'children of disabled people' in place of 'young carers'.In using these terms, it was recognized that not all children of disabled people are young carers and that young carers may be carers for other family members as well as parents.This increased the potential for the inclusion of irrelevant studies and omission of key papers and led to a decision to develop a separate strand for 'children of disabled people'.
The results of the two strands were combined in the final search stage.Surrogate terms were also identified for all fields.
The strategy guided the search, which was amended to suit each of the seven electronic databases (PsycINFO, MEDLINE, EMBASE, ASSIA, Sociological Abstracts, IBSS and JSTOR), and the Web of Science citation index.The search was carried out in June 2018.
Search results are in Table 2 with the full details included in Data S1.

| Selection and appraisal of documents
The different stages of the search and screening process are included in the realist flow diagram (Figure 1).Following deduplication, studies were screened against the inclusion criteria at title (n = 12 391), abstract (n = 4351) and full paper (n = 208) levels.Due to the large number of results, titles were single screened to remove studies that clearly did not concern young carers.Abstracts were double screened against the inclusion criteria by two independent researchers using Rayyan QCRI software.This was a blind procedure with the screeners unaware of each other's decisions and meeting periodically to discuss and resolve conflicts.
Of the 208 papers that were progressed to full-paper screening 10 could not be obtained.The first author screened the remaining 198 studies and identified studies to include in the synthesis.
Snowballing enabled the checking of reference lists in included papers for any unidentified studies (n = 3).The full papers were double screened with conflicting decisions again resolved through discussion.
Seventy-two papers were progressed for inclusion in the review.

Young carer status
•Child is taking on extra responsibility due to caring for someone.This can include sibling care (caring for an able-bodied sibling due to this person's disability).
•The caring relationship may be fully inverted where the child cares for the adult, or partially inverted meaning there is a two-way caring relationship.
•In addition, papers concerning the physical abuse, emotional abuse, neglect or parentification of children, due to adult being unable to care or child having inappropriate caring tasks

Relationship
•Care receiver is a member of the family, and usually living in the same home

Care receiver illness
•The care receiver has a chronic illness or disability.This can include mental health issues or a substance misuse problem

Outcome
•The mental health of the young person.This can include both potential impacts (e.g., depressions, anxiety and conduct disorder) and benefits (improved relationships).
•Their psychosocial health including impacts (e.g., sleep and eating disorders, risky behaviour and adjustment and personality disorders including ADHD) and benefits (life skills and resilience).
•Outcomes should be a result of their environment (i.e., their family situation and caring role).
•Inclusion of education due to clear links between education success and mental health.

Research design
•Empirical studies including qualitative, quantitative and mixed methods will be included.

•Articles and book chapters
Article language •All articles in the English language Due to the large number of studies included, the first author assessed and prioritized the studies on the basis of relevance and the potential to inform the model, and whether the methods that generated the data were rigorous.Studies were divided into primary (n = 16), secondary (n = 28) and surplus (n = 30) sets.Surplus papers with lower model-building potential were removed from the set, with the potential to utilize them to resolve uncertainties in the model.

| Analysis and synthesis processes
The first author was responsible for the analysis of data and the development of the model.Analysis and synthesis were completed in Nvivo 11.In line with the realist approach that all information has the potential to inform the model, all studies were assessed using lineby-line coding.The process was incremental with primary papers analysed in turn to identify key evidence and develop concepts.As the analysis continued, these concepts became increasingly detailed configuration components, resulting in the individual configurations.
Analysis of the secondary paper analysis enabled further strengthening and modification of the configurations and consideration of how the different configurations linked together.This resulted in significant 'embedding' (Jagosh et al., 2014) where evidence used to inform the mechanism in one configuration was also the basis of context in a subsequent configuration.
Tables 3 and 4 include full details of the primary and secondary studies, respectively.

| Main findings
The resultant model contains 17 CMO configurations in three domains concerning caregiving responsibilities, identity and support (Figure 2).The impacts of caring itself were captured in the caregiving

| Caregiving responsibilities domain
The caregiving responsibilities domain (Figure 3) contains six CMO configurations relating to the responsibilities of the young carer and the person that they care for.
Level of caregiving (CMO 5 ) is the penultimate configuration in the domain.It is informed by six studies and is key to considering the different ways of assessing young carers.Both contextual factors are embedded mechanisms that have been already detailed in previous configurations; appropriateness of responsibilities (Lane et al., 2015;   C 5a ) and time spent caring (Aldridge & Becker, 1993; Champion   et al., 2009; Joseph et al., 2009; Kallander et al., 2018; Lane   et al., 2015; Moore et al., 2011; C 5b ).
'Caregiving responsibilities' is the final configuration (CMO 6 ) and considers why caregiving responsibilities impact on children differently.The configuration is sourced from 11 studies and includes the previous mechanism, level of caregiving role (Lane et al., 2015; C 6a )a s one of three contextual factors.Young carer identity (C 6b ) is key with evidence that female young carers are more likely to have emotional health impacts including depression (Sharer, 2015), risky behaviour, stress and self-harm (Cree, 2003) and to develop negative coping strategies (Fraser & Pakenham, 2009).In comparison, males develop more positive family relationships but have higher emotional stress (McMahon & Luthar, 2007) though this opposes findings by (Stein et al., 1999).Considering age, there is evidence that older young carers have increased problems with sleep, self-harm, stress (Cree, 2003) and risky behaviour (Stein et al., 1999).Minority ethnic young carers experience greater stress (Cree, 2003), but also improved family relationships compared to other young carers (Khafi et al., 2014).The final contextual factor, duration of time as a young carer (C 6c ), can result in negative impacts (Doutre et al., 2013), but there was conflicting evidence over whether impacts grow with duration (Aldridge, 2006) or if some peak for middle range young carers (Cree, 2003).Young carers have poor mental health in comparison with peers (Lloyd, 2013;Tseliou et al., 2018) and are more likely involved in risky behaviour (Cree, 2003).

| Identity domain
The identity domain is a series of four configurations with each mechanism embedded in the next configuration as a contextual factor (Figure 4).The first, assignment configuration (CMO 7. ) concerns how children can potentially be selected to be a carer by their family, and is evidenced by nine studies.A single contextual factor, duty (C 7a ), reflects the cultures of African countries (Andersen, 2012;Lane et al., 2015;Olang'o et al., 2012;Robson et al., 2006;Skovdal, 2011;Skovdal et al., 2009;Skovdal & Ogutu, 2009) and South Korea (Hwang & Charnley, 2010) where children are raised to have a duty (C 7a ) towards elderly or unwell relatives.In place of cultural duty, Aldridge and Becker (1993) evidenced familial duty as a reason for caring in the UK.There is significant evidence that gender is an important part of this duty, with females more likely selected to be young carers (Olang'o et al., 2012;Robson et al., 2006;Skovdal, 2011;Skovdal et al., 2009), though Hwang and Charnley (2010) found that FIGURE 4 Identify domain male young carers in Korea are common due to the honour attached to the role.Evidence was also found of age as important with older children becoming carers (Andersen, 2012;Hwang & Charnley, 2010;Skovdal et al., 2009).Evidence in the synthesis indicated that assigned young carers can feel victimized in the role (Skovdal et al., 2009;Skovdal & Ogutu, 2009).

| Support domain
The support domain includes seven configurations (Figure 5).The first, supportive community configuration, refers to the support made available to young carers and their families in the local area and FIGURE 5 Support domain contains data from 12 studies.The first of three contextual factors, stigma (C 11a ) is due to particular illnesses being viewed as undesirable, for example AIDS (Bolas et al., 2007;Martin, 2006;Olang'o et al., 2012;Skovdal & Ogutu, 2009) or substance misuse issues (Moore et al., 2011).This can result in significant isolation from the community.In contrast social recognition (C 11b ) of the young carer can increase community support and reduce this isolation (Andersen, 2012;Skovdal et al., 2009;Skovdal & Ogutu, 2009).The presence of understanding friends (C 11c ) is the final factor with positive friendships potentially reducing the stress, anxiety and depression caused by the caregiving itself (Kain, 2009;Kavanaugh, 2013;Nichols et al., 2013;Sharer, 2015;Skovdal et al., 2009;Williams et al., 2009).
The professional carer support configuration (CMO 12 ) concerns the provision of formal support for the care receiver from outside the family.It is informed by four studies and has one contextual factor with social services capacity (C 12a ) limiting the number of families who receive professional support (Aldridge, 2006;Aldridge & Becker, 1993;Andersen, 2012;Olang'o et al., 2012).
The young carer informed configuration (CMO 13 ) relates to the provision of information about the care receiver's illness from social care and health professionals.The configuration is evidenced by four sources and has a single contextual factor relating to young carer identity (CMO 13a ) and the appropriate age to receive particular information to support them as a carer (Aldridge & Becker, 1993;Martin, 2006;Nichols et al., 2013;Trondsen, 2012).Trondsen (2012) found that a lack of information or training can cause stress and increase frustration.
Both the professional carer support (Aldridge, 2006; Aldridge &   Becker, 1993; C 14a ) and informed young carer (Aldridge &   Becker, 1993; Martin, 2006; Nichols et al., 2013; C 14b ) configurations are embedded as contextual factors in the supportive services configuration (CMO 14 ).The configuration concerns whether education, health and social services meet the needs of young carers and is informed by 11 sources.The configuration includes three further contextual factors.There is evidence that a lack of identification (C 14c )b y services can cause isolation (Aldridge, 2006;Aldridge & Becker, 1993;Thomas et al., 2003), with negative experience of school increasing conflict.In contrast positive experiences increase hope (Andersen, 2012;Skovdal et al., 2009).There was also evidence that young carers are not involved in decision making (Andersen, 2012;Martin, 2006;Robson et al., 2006;Thomas et al., 2003).The provision of young carers projects (C 14e ) are valued by as a place to meet other young carers and receive support (Aldridge, 2006;Kavanaugh, 2013;Moore et al., 2011;Thomas et al., 2003), with benefits including isolation and stress reduction, increased knowledge of the illness and family cohesion (Coles et al., 2007).
The young carer-care receiver relationship configuration (CMO 15 ) is informed by eight studies.Symptoms relating to the care receiver's illness (C 15a ) have the potential to affect their behaviour and the relationship with the young carer (Aldridge & Becker, 1993;Bolas et al., 2007;Martin, 2006;Nichols et al., 2013), potentially resulting in conflict (Kavanaugh, 2014).The relationship is also affected by the appreciation of the young carer (Kain, 2009; Martin, 2006; C 15b ), with the potential for depression (Bauman et al., 2007) if the young carer does not feel fully appreciated (Kavanaugh, 2013(Kavanaugh, , 2014)).
The supportive family configuration (CMO 16 ) is based on nine studies and has three contextual factors including the previous young carer-care receiver relationship configuration as an embedded mechanism (Kavanaugh, 2013; Martin, 2006; Nichols et al., 2013;   Trondsen, 2012; C 16a ).Being informed about the illness (C 16b ) by family is important (Kain, 2009;Nichols et al., 2013;Robson et al., 2006;Trondsen, 2012), but evidence suggests that families are reluctant to tell the young person (Martin, 2006) despite evidence that this reduces frustration and enables the development of coping strategies.
Each factor is an embedded configuration concerning support from different levels of society, namely, community (C 17a ), services (C 17b ) and family (C 17c ).informed by two and four sources.When evidence is sparse, there is an increased risk that additional contextual factors or outcomes have not been identified or that those identified are incorrect.In addition, evidence from a larger number of sources could help resolve uncertainties that could indicate potential reliability issues.For example, it is unclear whether the impacts of caregiving responsibilities (CMO 6 ) grow over time (Aldridge, 2006) or peak for middle range carers (Cree, 2003), and further research would resolve this conflict.
There is a larger limitation to the realist synthesis in terms of Pl e a s e n o t e: C h a n g e s m a d e a s a r e s ul t of p u blis hi n g p r o c e s s e s s u c h a s c o py-e di ti n g, fo r m a t ti n g a n d p a g e n u m b e r s m a y n o t b e r efl e c t e d in t his ve r sio n.Fo r t h e d efi nitiv e ve r sio n of t hi s p u blic a tio n, pl e a s e r ef e r t o t h e p u blis h e d s o u r c e.You a r e a d vis e d t o c o n s ul t t h e p u blis h e r's v e r sio n if yo u wi s h t o cit e t hi s p a p er. Thi s v e r sio n is b ei n g m a d e a v ail a bl e in a c c o r d a n c e wit h p u blis h e r p olici e s.
service and participants.Within a configuration, a mechanism (M) can produce a change in the system, with contextual (C) factors having the potential to trigger the mechanism.Outcomes (O) are the product of the context and mechanism together.Modelling the programme in full provides a complex picture of its success in triggering desired mechanisms while also considering the reasons why other mechanisms remain unchanged.
total, 44 of the 15 518 unduplicated search results were included in the synthesis.The dataset had 40 different lead authors and spanned from 1993 (Aldridge and Becker) to 2018 (Kallander et al; Tseliou et al.).Seventeen of the studies were based in Europe including the UK (n = 15), with 11 American and 10 African studies.The method of recruiting young carers varied and included care receiver services (n = 12), mainstream (education, health and social care) services (n = 12) and carer projects (n = 11).

FIGURE 1
FIGURE 1 Realist flow diagram of the realist search and screening process responsibilities domain and varied significantly due to the circumstances of individual young carers, the person that they care for and their family.The development of a caring identity within the young person's wider identity was detailed in the Identity domain, and the support and recognition provided to young carers at different levels of society was encapsulated in the support domain.Data S2 contains a guide detailing every evidence component (contextual factors, mechanisms and outcomes), with excerpts demonstrating links between components.The organizing construct demonstrates how the three domains interact with each other.The identity domain moderates the caregiving responsibilities domain through the potential development of a caregiving identity that mitigates the negative effects of caring and enables positive benefits.The support domain moderates the caregiving responsibilities domain through positive support.This mitigates the negative effects and enables the development of positive benefits, though a lack of support exacerbates the negative impacts of caregiving.The support domain moderates the identity domain as support improves self-perception of caring and aids the development of the caregiving identity, but a lack of support leads to young carers devaluing their role.

FIGURE 2 FIGURE 3
FIGURE 2Organizing construct of the model and how the three domains relate to each other

|
Summary of findingsThis realist synthesis of previous research has resulted in a model of the relationship between young carer status and mental and psychosocial health, with three domains concerning caregiving responsibilities, identity and support.The realist approach of utilizing CMO configurations provides a framework by which to structure previous findings and gives clarity to the question of the ways in which young carer status impacts children differently depending on their lives.Despite the synthesis focusing on the outcomes of caring as a result of the social system rather than interventions, CMO configurations are still central.They are however framed differently with the mechanisms not targeted by a social programme but mapped in situ, based on pre-existing societal contextual factors.The next stage for a traditional realist synthesis would be to utilize the model to evaluate a type of intervention by assessing whether it is able to trigger mechanisms and enable positive change.The results would then inform recommendations for the future of the interventions.As this is a realist synthesis of pre-existing social context the model can instead be used to inform intervention development rather than evaluation.4.2 | LimitationsDespite including 44 studies, there is a need to refine and test the resultant model further with the reliability of the individual configurations, and therefore the wider model, dependent on the diversity of the dataset evidence.The most reliable configurations concern topics regularly researched in the past with the time spent caring configuration (CMO 4 ) and the caregiving responsibilities configuration (CMO 6 ) informed by 16 and 11 studies, respectively.In contrast the other family members in employment configuration (CMO 2 ) and the professional carer support configuration (CMO 12 ) are young carers who are unidentified, and therefore unsupported, by key services including social services and schools.Although past research has highlighted the importance of identifying these young carers, little progress has been made as to how this should happen with the majority of research recruiting participants through young carer services where they are already receiving support.With the exception of the population-wide quantitative studies, unidentified young carers are not represented in the evidence.As a result of this they are also under-represented in the model, with identification a contextual factor for the supportive services configuration (CMO 14 ) but otherwise absent.This implies that, with the exception of identification by services, the lives and experiences of unidentified young carers are largely similar to as identified young carers when their unidentified status will potentially be a contextual factor for multiple configurations, and disclosure a key mechanism.Research with this population in particular will help strengthen the model.5 | CONCLUSIONThis study adapted the traditional realist synthesis methodology to focus on the pre-existing social context of young carers with a focus on caring responsibilities but also support and the development of a caring identity.The study clarifies why the mental and psychosocial health benefits and impacts vary for different young carers and identifies a number of possible mechanisms, which could be targeted by future interventions.It also reinforces the need to involve all young carers in research, including those not currently known to services.

TABLE 1
Inclusion criteria developed for screening of papers

TABLE 2
Breakdown of the complex literature search results from each database, prior to deduplication

TABLE 3
Primary papers included in the realist review

TABLE 4
Secondary papers included in the realist review