Remember me
ARTICLE
Year : 2022 | Volume
: 20
| Issue : 2 | Page : 161-169
Mechanisms of Change for Child Mental Health and Psychosocial Support in Conflict Settings: A Systematic Review
Tania Bosqui1, Anas Mayya2, Aws Al-Kadasi3
1 DClinPsych, Assistant Professor, Department of Psychology, American University of Beirut, Beirut, Lebanon
2 Graduate Student, American University of Beirut, Beirut, Lebanon
3 Undergraduate Student, American University of Beirut, Beirut, Lebanon
Correspondence Address:
Dr. Tania Bosqui
Assistant Professor, American University of Beirut, Jesup Hall, Beirut
Lebanon
Source of Support: None, Conflict of Interest: None
DOI: 10.4103/intv.intv_25_21
There is a growing evidence-base for the effectiveness of mental health and psychosocial support (MHPSS) programmes for children affected by conflict for a range of mental health and wellbeing outcomes, but with limited evidence for how these interventions produce change. This study aimed to review the evidence for mechanisms of change of MHPSS interventions for children affected by conflict. Systematic review methodology was used to screen the PubMed, Published International Literature on Traumatic Stress (PILOTS) and PsycInfo databases for primary quantitative studies. Out of 3,903 records, seven studies were included, all evaluations of school-based preventative programmes. Results showed a mediation effect of peer relations on treatment change for wellbeing outcomes, while a predictable and cooperative learning environment partially mediated treatment change for academic outcomes, victimisation and mental health. One study found a small negative mediation effect of play-based social support for posttraumatic stress disorder. No evidence was found for any other mechanisms. The findings of this review highlight the role of social relationships as mechanisms of change in MHPSS programmes, but with limited studies to draw from. Future programme evaluations should include measures of proposed mechanisms to further our understanding of how MHPSS programmes work to continue to improve their relevance and scope and to do no harm.
Keywords: children and adolescents, conflict, mechanisms, mental health, psychosocial support
Few studies have tested the mechanisms of change for child mental health and psychosocial support programmes.There is evidence that for preventative school-based interventions, peer relations and a predictable and cooperative learning environment are mechanisms of change for improved mental health and wellbeing.Play-based social support for children with posttraumatic symptoms may worsen symptoms. Introduction Mental health and psychosocial support (MHPSS) programmes for children and adolescents affected by conflict have growing evidence of effectiveness for a range of mental health difficulties. However, there are remaining questions about the mechanisms of change underlying them. Children and adolescents in conflict settings experience a high incidence of mental health difficulties, including depression, anxiety and posttraumatic stress (Bendavid et al., 2021; Fasfous et al., 2013), internalising and externalising symptoms (Peltonen & Punamäki, 2010), as well as low levels of wellbeing and hope (Tol et al., 2013) and a range of local idioms of distress (Hassan et al., 2016) − with severity and impairment impacted differentially dependent on types of conflict experiences, contexts (Tol et al., 2013) and daily stressors (Miller & Rasmussen, 2010). In response to this need, interventions to promote wellbeing and prevent or treat mental health difficulties have continually been developed and adapted for low-resource and conflict-affected populations. These interventions are largely informed by the international Inter-Agency Standing Committee guidelines (IASC, 2007) which recommend a multilevel approach that builds on existing resources, grounded in principles of human rights, participation and to do no harm.
There is a growing evidence-base for the effectiveness of these interventions for a range of clinical and nonclinical outcomes and in multiple different countries and contexts (O’Sullivan et al., 2016), though with some mixed findings, and gaps in the evidence for more community-focused programmes and the mechanisms of change (Jordans et al., 2016; Betancourt et al., 2013; Brown et al., 2017). Mechanisms of change refer to the core processes by which intervention aims and therapeutic change are achieved. Mechanisms can be tested in trial research through mediational analyses that identify temporal and modal relationships between intervention and outcome. Mechanisms of change, therefore, aim to answer the question of how an intervention produces change rather than simply whether it does (Kazdin, 2007, Kazdin & Nock, 2003. An understanding of the mechanisms of change is vital to improve the evidence-base of MHPSS because these mechanisms can be more easily incorporated into existing interventions in low-resource settings, which rely on training non-specialists to provide treatments. For example, the Common Elements Treatment Approach draws on this kind of evidence to inform a nonspecialist, transdiagnostic, mental health treatment for low-resource and emergency settings (Murray et al., 2018). Intervention content may differ in format, cultural acceptability and materials, but can be linked to the evidence-base through these core elements. Interventions based on high-quality evidence of mechanisms of change in different contexts and populations may therefore be able to offer a low-cost solution to poor access to evidence-based MHPSS, optimising the change process of the programme. This is particularly important in many conflict-affected settings where there is often a high need but few specialised providers.
Part of the challenge in identifying mechanisms of change, in high- and low-resource settings alike, is how to measure the complexity of how change occurs during interventions, for different people and in different settings. Researchers have tested mechanisms of change in a range of ways: Mediational analyses can help to identify variables statistically, directly and indirectly, related to outcome change; qualitative feedback can provide insight into the lived experiences of those using the intervention; moderation analyses may identify individual or treatment characteristics that influence the strength of intervention outcomes (such as age or gender); practice element coding can systematically identify the common intervention components across interventions that have existing evidence of effectiveness for mental health outcomes; and testing the addition of a proposed active ingredient through multiple treatment arms in a clinical trial or through mixed methods multiple n = 1 designs that link qualitative session by session data and external life events to profiles of symptom score change for individual participants (Kazdin, 2007).
Based on the existing evidence in high-income countries and non-conflict settings, specialist interventions have good evidence for a range of “active ingredients” associated with change, tested through multiple arms of clinical trials. The most consistent evidence exists for cognitive restructuring and behavioural activation for depression, gradual exposure for anxiety, prolonged imaginal and in vivo exposure for traumatic stress and positive parenting for children’s externalising behaviours (Janssen et al., 2021; Kaczkurkin & Foa, 2015; Weber et al., 2019). In addition, there is strong and consistent evidence for the role of a therapeutic alliance or rapport as a mediator of change (Baiar et al., 2020). The current evidence-base for the mechanisms of change in MHPSS for conflict-affected children and adolescents, however, is limited. A review by Brown et al. (2017) included an analysis of common practice elements in effective MHPSS interventions for war-affected children and adolescents, highlighting treatment components like psychoeducation, prolonged exposure and cognitive restructuring. However, included studies did not test whether these components were directly related to changes in mental health outcomes. A qualitative review of reviews of mechanisms of change by Bosqui and Marshoud (2018) also found limited evidence; out of 13 included reviews, only six referred to mechanisms of change, and only one of these reviews drew from studies using mediational analysis. Finally, a meta-analysis of focused psychosocial support interventions for children in low-resource humanitarian settings found no evidence to support a mediation effect of proposed mechanisms of coping, hope or social support (Purgato et al., 2020). Given the implications for improved service provision for children and adolescent mental health, this study aimed to systematically review quantitative primary studies to identify the current state of the art for the mechanisms of change of MHPSS interventions for children, adolescents and young adults affected by conflict.
MethodA systematic review was conducted drawing on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Shamseer et al., 2015) and the Cochrane Handbook for Systematic Reviews (Version 5.1.0). The protocol for the review can be provided upon request.
Search Strategy
The databases Published International Literature on Traumatic Stress (PILOTS), PubMed (including MEDLINE) and PsycInfo were searched on 11 October 2019 for studies on MHPSS interventions for children, adolescents and young people affected by armed conflict, with no restriction on publication language, country or region. A pilot search using these databases was conducted beforehand to ensure that all known published studies were picked up by the search terms. Terms related to the study setting were set for “all fields,” while other search terms were restricted to “title/abstract” based on the pilot search. The final search terms included words for the population (e.g. child, adolescent and youth), the setting (e.g. armed conflict, armed violence and low-income), intervention type (e.g. psychotherapy, psychosocial and mental health) and mechanism of change (e.g. mechanism, process and mediation; see [Table 2] for the full list of search terms).
Inclusion and Exclusion Criteria
Inclusion criteria for reviewed studies were children, adolescents and young people aged ≤25 years, who were affected by war, armed conflict or political violence at the time of data collection. To determine whether the population was conflict-affected we relied on the setting description provided by the authors of the study. Studies including children affected by single terrorism-related events were excluded, as well as displaced populations in non-war-affected countries who were granted leave to remain (such as refugees in high-income countries). This was due to contextual differences in experience and threat, which relate to different service models, accessibility and interventions. Included interventions were any mental health or psychosocial intervention that aimed to improve the wellbeing, mental health or resilience of children, adolescents and young adults, in any format, and at every level of the IASC pyramid model for MHPSS in emergencies (IASC, 2007). Health or social care interventions unrelated to wellbeing, mental health or resilience, were excluded, such as medical interventions for health problems or risk-focused child protection interventions that did not measure mental health, wellbeing or resilience.
All outcomes were included as long as they related to child or adolescent wellbeing, mental health or resilience, and were measured quantitatively. No restriction was made for the type of mechanism or process of change as long as they were measured quantitatively. Qualitative and mixed methods studies were excluded. However, relevant qualitative and mixed methods studies were flagged to inform the discussion of findings.
For the purposes of this review, “mechanisms of change” were defined as the process or steps responsible for a therapeutic outcome, and were used interchangeably with mediators of change, therapeutic processes, mechanisms of counselling, treatment processes and factors that produce change or mediators of change. This definition, therefore, excludes moderators of change (characteristics that influence the magnitude of the relationship, such as age or gender) and protective factors, as well as practice elements, clinical techniques, strategies or intervention aims. This was to distinguish between processes of change and treatment content, with this review focusing on the former.
Procedure
Records were first screened by title by one research team member (AM) to exclude irrelevant studies. Abstracts and full texts were then independently screened by two members of the research team (AM and AAK). Reference lists of included studies were also screened. Discrepancies in the final included studies were resolved through discussion with a third party (TB) and a consensus agreement. The inter-rater agreement rate was high (Cohen kappa statistic = 0.83). Data were split and extracted by two research team members (AM and AAK), which were then quality checked by the third member (TB). Extraction included data on 1) study characteristics, 2) population and setting, 3) intervention and theory of change, 4) outcomes and mechanisms and 5) main findings for mechanisms of change.
The quality of each included study was assessed using the Standard Quality Assessment Criteria (Kmet et al., 2004), with an additional item developed for this study that assessed the quality of analysis for the mechanism or mechanisms of change. This item “Is the analysis of mechanisms of change defined and appropriate?” was scored as “yes” when the analysis was conducted using a mediational analysis or other relevant statistical analysis, “partial” when there was insufficient reporting, limited sample size or unclear analysis and “no” when there was no description or the method was inappropriate.
ResultsOut of 3,903 records identified in the databases search, 364 were selected by title, 41 by abstract, six by full text and one was added through the reference list screening (see flowchart in Figure 1). In total, seven papers on four trials were included. The main reasons for exclusion were that the study had no analysis of mechanisms of change or was not set in a conflict-affected setting at the time of data collection.
Study Characteristics
Included studies, shown in [Table 1], were conducted in Palestine, Indonesia, Sri Lanka, Burundi and the Democratic Republic of the Congo, with children and adolescents aged between 7 and 17 years old. Sample sizes ranged from 329 to 4,142 children. All five interventions were school-based preventative programmes, three of which drew on cognitive behavioural techniques, one on an ecological resilience theoretical framework and one on a bioecological framework that focused on social processes. All programmes were found to be effective for at least one child’s mental health outcome, except for Tol et al. (2014) who found no main effects. Tol et al. (2014) were, therefore, not included in the synthesis of mechanisms. All but one study used a cluster randomised control trial design, and all used a waitlist comparison group. A range of mental health disorders was included as outcomes, most commonly post-traumatic stress disorder (PTSD). Mechanisms measured were peer and sibling relations, posttraumatic cognitions, emotional regulation, hope, coping, social support, supportive schools and teachers and predictable and cooperative learning environments.
Quality
Quality assessments indicated a high methodological quality of all included studies. The most common bias was related to a lack of blinding of investigators and the local criterion validity of measures. All studies conducted statistical mediation analysis using structural equation modelling.
Intervention Effectiveness
A significant direct effect of Teaching Recovery Techniques (TRT) for adolescents was found for improvement in mental health (Diab et al., 2014; Punamäki et al., 2014), but not for posttraumatic stress (Kangaslampi et al., 2016) or emotional regulation (Punamäki et al., 2014). Significant direct effects of the school-based psychosocial intervention were also found for better outcomes on hope, positive coping, peer and play social support, but not for negative coping and emotional social support (Tol et al., 2010); and the nonspecialised mental health intervention for conduct problems but not for PTSD, depression and anxiety symptoms (Tol et al., 2012).
No significant direct effects of Learning to Read in a Healing Classroom (LRHC) were found for victimisation or mental health (Aber et al., 2016), nor the classroom-based intervention for posttraumatic symptoms, depression, anxiety or hope (Tol et al., 2014).
Mechanisms of Change
Three studies found evidence of a mediation effect between treatment and outcome. For the TRT for adolescents, treatment effects for mental health were mediated by improvement in peer relations (β = 0.21, t = 1.96, Standard Error = 0.07, P = 0.05; Diab et al., 2014). The LRHC programme for children found a significant partial mediational effect of predictable and cooperative learning environments for the outcomes of improvements in victimisation (△x2 = −46, df = 1, P < 0.05) and mental health (△x2 = −63, df = 1, p < 0.05; Aber et al., 2016). For the school-based psychosocial programme for children and adolescents, Tol et al. (2010) found a small mediation effect of play social support for PTSD, but the effect was the opposite of the predicted direction (Coefficient = 0.004, 95% Confidence Intervals: 0.00–0.07, P = 0.03). Therefore, a reduction in play social support was marginally associated with better treatment outcomes for PTSD.
No evidence for a mediational effect was found for sibling relations (Diab et al., 2014), posttraumatic cognitions (Kangaslampi et al., 2016), emotional regulation (Punamäki et al., 2014), hope (Tol et al., 2010), coping (Tol et al., 2010, 2012) and supportive schools and teachers (Aber et al., 2016) on any of the included outcomes.
DiscussionThis systematic review aimed to identify the mechanisms of change of MHPSS interventions for children and adolescents affected by armed conflict, focusing on quantitative evidence using mediational analyses. Only seven studies of four trials were found, all preventative school-based programmes. The studies identified significant mediators of peer relations for wellbeing, predictable and cooperative learning environments for reduced victimisation and improved mental health, and a negative mediational effect of play social support for PTSD. None of the other included mechanisms were found to mediate the relationship between the programme and outcome.
The low number of included studies reflects a continuing research gap in understanding how MHPSS interventions produce therapeutic change, supporting the call for improved research on mechanisms of change and the inclusion of expected mediational variables in study designs (Bosqui & Marshoud, 2018). From the limited existing evidence, this review indicates that improved peer relations and predictive and cooperative learning environments are significant mechanisms for interventions to improve mental health and wellbeing. This is in line with research that has highlighted the importance of the social ecology of children in protecting and supporting their mental wellbeing during major adversities (Betancourt et al., 2010), as well as the importance of social relationships (Maercker and Horn, 2012). The programmes with these identified mechanisms include the TRT programme (Diab et al., 2014; Kangaslampi et al., 2016, Punamäki et al., 2014) and the LRHC programme (Aber et al., 2016).
The TRT programme draws on cognitive behavioural techniques and focuses on trauma processing, which for this study was in the context of the aftermath of the 2009 Gaza war. However, only improvement in peer relationships mediated the treatment effect, not sibling relationships, posttraumatic cognitions or emotional regulation. This highlights the importance of drawing on the social-ecological model for children affected by conflict, where the wider social environment may be crucial for recovery and resilience to ongoing adversity. Kangaslampi et al. (2016) and Punamäki et al. (2014) point out the discrepancy between the null findings for posttraumatic cognitions and emotional regulation, compared to the significant mediating effects of the same mechanisms in non-conflict-affected settings and populations, further supporting the potential need for programmes to shift focus from individual and internal psychological processes to the wider social world of children and adolescents, in the context of conflict and collective trauma. This is in line with findings from a related study on the TRT programme in Palestine, in which the programme was more effective for children with an insecure attachment style (Eloranta et al., 2017). The authors suggest that children with fewer social resources and poorer family functioning benefit more from the programme. Kangaslampi et al. (2016) also highlighted the importance of matching the intervention with the context and level of exposure to the current threat. Assumptions about common adaptive mechanisms in non-conflict-affected populations may be unhelpful in conflict settings, whilst some risk factors may actually have adaptive functions. For example, emotional ventilation (Punamäki et al., 2014) and emotional processing (Bosqui & Marshoud, 2018) have been associated with poorer psychosocial outcomes during acute exposure, while high appetitive aggression has been associated with lower trauma symptomatology (Hecker et al., 2013). These differences demonstrate that misplaced assumptions about what an effective programme should entail may unintentionally interfere with adaptive survival instincts.
The LRHC programme in the eastern Democratic Republic of the Congo was based on a systemic theoretical model focused on supporting social relationships and fostering change through bidirectional processes and mutual peer support. The importance of predictive and cooperative environments, rather than just supportive teachers, in improving mental health and academic outcomes in children again highlights the importance of social processes in supporting change.
The reasons for the null findings for the mechanisms of change of hope (Tol et al., 2010), coping (Tol et al., 2010, 2012) and the inverse relationship for play social support (Tol et al., 2010), for school-based programmes in Indonesia, Sri Lanka and Burundi are less clear. Though the effect size is small, the authors (Tol et al., 2010) suggest that universal school-based programmes that include play activities may unintentionally slow down trauma recovery for children who are experiencing specific posttraumatic stress symptoms, because the intervention is indirect and not trauma-focused. Following this explanation, and similar to the harmful effects of debriefing for adults after a traumatic event (Rose et al., 2002), play-based activities for children may interfere with natural recovery and trauma processing, or enable avoidance. This is potentially an important finding that needs further exploration. Given the heavy use of play-based MHPSS in humanitarian settings, any negative impacts and the factors associated with this (type of play, context and population characteristics) would need to be communicated to MHPSS providers to respect the do no harm principle of humanitarian intervention. The school-based programmes were found to directly improve hope and coping, which may be considered positive outcomes independently of their mediating effect on mental health (Tol et al., 2010, 2012) and this may be particularly relevant for universal programmes and non-clinical settings where presenting issues extend beyond traditional mental health symptom profiles (O’Sullivan et al., 2016).
The evidence identified in this review, therefore, indicates that social and interpersonal mechanisms (peer relations, predictable and cooperative learning environments) are important in improving mental health and wellbeing during school-based MHPSS interventions. Evidence for mechanisms like cognitive restructuring, identified in the context of high-income, non-conflict settings, was not supported by this review in the context of conflict-affected settings. Three programmes were cognitive behavioural but found no mediational effect of posttraumatic cognitions and emotional regulation on treatment outcomes. These interventions were also found not to have a direct treatment effect on posttraumatic stress, anxiety or depression symptoms. We argue that in many conflict-affected settings and for conflict-affected populations social and interpersonal factors are of greater importance in the treatment process, perhaps given the nature of the collective in conflict and community violence or the additional complexities of social determinants of mental health. However, many existing active ingredients and mechanisms were not tested in the included studies, namely, trauma processing, exposure, behavioural activation, positive parenting and therapeutic rapport. Further evidence is needed to identify their role in MHPSS in conflict settings.
A previous qualitative review of the mechanisms of change in MHPSS programmes for children and adolescents affected by conflict (Bosqui & Marshoud, 2018) also found several proposed mechanisms which were not tested by any of the studies identified in this quantitative review. This indicates a theory–practice gap in which clinicians’ hypotheses of intervention mechanisms have not been sufficiently translated into research questions and designs. Comparing the findings of the qualitative review with the findings of this review, the following mechanisms remain to be tested: a sense of safety, community relations, engagement with values and beliefs, mental health literacy, problem-solving, altruism and therapeutic relationships. Additionally, the trajectories of change in a multiple n = 1 study conducted in Burundi (Jordans et al., 2012) identified mechanisms of client-centeredness, therapeutic alliance, active problem solving, trauma-focused exposure and family involvement. Building evidence from observations and other evidence generated in conflict-affected settings − rather than assuming the same mechanisms and programme content in non-conflict-affected settings applies to conflict settings − is essential to better inform our understanding of how MHPSS programmes work. A contextualised and evidence-based understanding could help to improve the relevance of interventions and better inform the content and scale-up of programmes to support child and adolescent mental health and wellbeing.
Strengths and Limitations
This review aimed to identify the current state of the art on how mental health and psychosocial programmes work, benefiting from a robust systematic review methodology and a research team with significant experience of working in conflict-affected settings. However, the review was limited due to the restricted capacity of the researchers to complete a comprehensive grey literature search, which could have led to the identification of relevant unpublished programme reviews. The capacity of the researchers was affected by the onset of the economic crisis and the 2019 October revolution in Lebanon. For this reason, the research team could not update the search before publication. Furthermore, as the review aimed to identify only quantitative studies that statistically tested mechanisms, the inclusion criteria were quite stringent. This may have meant excluding studies that may have given an insight into other potential mechanisms, such as more qualitative mechanisms related to the experience of implementation and acceptability. Finally, as the review relies on published data in international academic databases, included data are likely to miss the diversity of settings, programmes and community initiatives modelled on localised and contextually grounded understandings of therapeutic change.
ConclusionsThe number of studies found in this review testing pathways of change for MHPSS programmes was low. However, the findings of included studies highlight the importance of the social world of children and adolescents affected by conflict in producing treatment effects for mental health and wellbeing, in particular in peer and classroom environments. The review also highlights several noneffects for expected mechanisms, such as cognitions, coping and hope, which further highlights the role of social, rather than internal, protective mechanisms. Further research that tests these proposed pathways is needed to inform our understanding of how programmes work to improve their contextual relevance and utility, and to do no harm.
Financial support and sponsorship
This study was funded by the American University of Beirut’s University Research Board (URB grant number: 103782).
Conflicts of interest
There are no conflicts of interest.
References
Comments (0)