Evaluating a Movement-Based Mental Health Promotion Intervention for Refugee Children in Uganda: A Quasi-Experimental Study

ABSTRACT

Objective

Mental health promotion interventions are widely implemented in humanitarian settings and low- and middle-income contexts (LMICs), yet evidence on effectiveness is scarce and mixed. This study evaluated the movement-based mental health promotion intervention “TeamUp” in Bidibidi refugee settlement, in Northern Uganda.

Method

A quasi-experimental study including four schools (two per arm) assessed the outcomes of 10- to 15-year-old South Sudanese and Ugandan children (n = 549). Randomly allocated, they either participated in up to 11 TeamUp sessions (n = 265) provided by trained facilitators; or belonged to a control group, which continued care as usual (n = 284). Primary outcomes measured psychosocial wellbeing, friendships and attitude to school; secondary outcomes included traumatic distress, depressive symptoms, quality of life, physical health, bullying, interoceptive awareness, and irritability. Data were collected at baseline and endline.

Results

Children joining TeamUp, showed significantly more improvements on primary outcomes: emotional and psychosocial wellbeing (M_diff = −1.49, SE = 0.6, p = .01), satisfaction with and attitude toward school (−0.57, SE = 0.2, p = .004); and secondary outcomes: traumatic stress (2.64, SE = 0.8, p < .001), health-related quality of life (−1.56, SE = 0.4, p = .001), physical health (−0.78, SE = 0.3, p = .014) and the TeamUp mechanisms of action scale (−3.34, SE = 0.9, p < .001), specifically the subscales social connectedness (−0.74, SE = 0.3, p = .007) and sense of agency (−0.91, SE = 0.3, p = .005), compared to the control group. No significant differences were found on bullying, interoceptive awareness, irritability and depressive symptoms.

Conclusion

The results are promising for TeamUp as a mental health promotion intervention for children affected by armed-conflict, displacement and on-going adversity. Further research will need to assess the intervention’s effectiveness.

Introduction

Psychosocial interventions are widely implemented in humanitarian settings, aiming to strengthen individuals’ psychosocial wellbeing and resources. There has been an increase in evaluations of mental health and psychosocial (treatment) interventions by nonprofessionals in recent years. Psychosocial interventions have shown to effectively reduce disruptive behavior problems (Burkey et al., 2018), and anxiety, depression and post-traumatic stress disorder (PTSD) (Barbui et al., 2020; Klasen & Crombag, 2013; Trimboli et al., 2021; Uppendahl et al., 2020). However, evidence is still scarce and heterogenous for mental health promotion and prevention interventions (Barry et al., 2013; O’Reilly et al., 2018; Patel et al., 2018; Uppendahl et al., 2020) and high-quality studies, which also assess quality of implementation and process outcomes (Rose et al., 2022), conducted in LMICs are needed (Uppendahl et al., 2020; Venturo-Conerly et al., 2022). The Lancet Commission for Global Mental Health and Sustainable Development has called for an increased attention to mental health promotion and prevention, as treatments alone cannot reduce the overall burden of mental health problems (Patel et al., 2018; Tol, 2015).

Improving the mental health and wellbeing of young people is especially important, given that the majority of adult mental illness has its onset in childhood and adolescence (Kessler et al., 2007, 2009). The neuroplasticity in the first two decades of life provides for an important sensitive period for promoting mental health and development (Patel et al., 2018). Children and adolescents affected by armed-conflict, violence and displacement, are particularly impacted, as experiences of traumatic events, loss of or separation from loved ones, unsafety and uncertainty, have social and psychosocial consequences (Adaku et al., 2016). While most people are incredibly resilient in the face of adversity, adequate interventions can promote social interaction as well as finding ways to distract and reduce “thinking too much” (Adaku et al., 2016; Kaiser et al., 2015; Ventevogel et al., 2013). Services for children at community and school-level are commonly implemented by humanitarian actors, yet most often lack rigorous evaluation or research (Barry et al., 2013; Jordans et al., 2016; Tol et al., 2011; Uppendahl et al., 2020; Venturo-Conerly et al., 2022).

Existing mental health and psychosocial interventions tend to focus on rather verbal and cognitive processing of experiences, feelings and behaviors, applying a cognitive-behavioral perspective (Jordans et al., 2016; Purgato et al., 2018). These, however, often miss an integration of the body-mind (inter)connection which is emphasized by movement-based and non-verbal approaches (Bareka et al., 2019; Dieterich-Hartwell & Koch, 2017; Dieterich-Hartwell et al., 2020; Homann, 2010). Using movement and the body cannot only promote embodiment but also support a more holistic approach to wellbeing (Bareka et al., 2019; O’Brien & Charura, 2022; Panhofer et al., 2012), while assuring trauma-informed care principles (Hobfoll et al., 2007; Ley & Rato Barrio, 2019). Especially for multi-cultural and multi-lingual groups, applying a non-verbal, movement- and play-based approach that goes beyond language and cognitive processing may be most relevant and appropriate (Ager et al., 2011; Bareka et al., 2019; Dieterich-Hartwell & Koch, 2017; O’Brien & Charura, 2022; Panhofer et al., 2012).

The movement-based mental health promotion intervention TeamUp was developed to support children between the ages of 6 and 17 who have been affected by armed-conflict and displacement (Bleile et al., 2021; TeamUp, 2020b). The intervention consists of diverse movement-based activities, that combine group games, sports-based and physical activities, dance and creative movement, routines, songs, settling moments, body awareness, and breathing exercises. Based on mental health and psychosocial literature and practices (Inter-Agency Standing Committee [IASC], 2007), including trauma-informed care principles (Hobfoll et al., 2007), physical movement (Bergholz et al., 2016; Purgato et al., 2018), embodiment and mind-body interconnection (Dieterich-Hartwell & Koch, 2017; Ley & Rato Barrio, 2019), a Theory of Change was developed (see online supplement). Sessions follow a clear structure and work on predetermined psychosocial themes with the ultimate aim to improve children’s psychosocial wellbeing (TeamUp, 2020b). A process evaluation showed TeamUp to be implemented with adequate quality and perceived positively by children, facilitators and reception center staff in the Netherlands (Bleile et al., 2021). Stakeholders regarded the intervention as providing children with the opportunity to; (i) experience positivity, normalcy and a sense of safety, (ii) outlet energy and strong emotions, and (iii) strengthen peer-relationships (Bleile et al., 2021).

The current quasi-experimental study aimed to evaluate the psychosocial outcomes of TeamUp for children, compared to a control group, living in a refugee settlement in Northern Uganda. We hypothesized that children participated in TeamUp will have improved psychosocial outcomes compared to children in the control group.

Methods

Setting

Uganda is the largest refugee-hosting nation in Africa and among the largest globally, hosting around 1.5 million refugees, primarily from South Sudan and the Democratic Republic of Congo (UNHCR, 2022b). Children make up over 60% of this population. Most experienced multiple traumatic events, prior or during their flight and displacement and may continue to face ongoing stressors and adversity (Adaku et al., 2016; Fazel et al., 2012; Reed et al., 2012). While Uganda’s government and multiple international agencies provide basic services such as shelter, sanitation, food and education, children’s psychosocial needs are often overlooked, unidentified and remain unaddressed (Purgato et al., 2018; Tol et al., 2015).

The current study was conducted in Bidibidi refugee settlement, in Yumbe district, in the very north-western part of Uganda – commonly called the “West-Nile” area. The settlement provides refuge to about 228,000 refugees, mainly from South Sudan (UNHCR, 2022a). The region is rural with most people engaging in agriculture and receiving additional support from humanitarian and non-governmental organizations (NGOs).

Study Design and Randomization

A quasi-experimental study design with four schools – two schools in either study condition – was conducted. The schools were randomly allocated to either experimental (TeamUp intervention) or control condition by flipping a coin in presence of two members of the implementation team and two of the research team. (The study is however not a Randomized Controlled Trial (RCT), as the sample size and power analysis were not adjusted for the effect of clustering.) TeamUp was provided twice per week for a period of six weeks in two schools (10 or 11 sessions, reduced from 12 sessions, due to public holidays during the implementation period). The other two schools served as the control group, where children continued with care as usual, thus education and physical education (PE) activities as usual. However, due to COVID-19 and prolonged school closure prior to the study, and reopening of schools for a short two-month window, schools gave prioritization to academic learning. The usual PE lessons were either irregularly or not at all offered, and children were primarily engaged in self-initiated, non-guided/unsupervised free play. Yet, this was the same across study arms. The intervention schools integrated TeamUp into their weekly curriculum.

From each school, two classes within grade 4 were randomly selected following coin flips, in order to achieve the needed sample size (≥250 children per study arm and approximately even numbers across schools). Figure 1 depicts the flowchart of the sample, from enrollment to analysis.

Figure 1. CONSORT flow chart.

Research assistants, statistician (GKG) and principal investigator (MJ) were masked to the group allocation, while research coordinator (BO), lead researcher (AB), TeamUp facilitators, and TeamUp coordinator, were not due to their involvement in the coordination of the intervention and/or study. Research assistant masking was assessed with a question in KoboCollect at the end of each endline assessment.

Sample

Four public primary schools were identified meeting the selection criteria: (i) primary schools with predominantly refugee children (ii) no psychosocial interventions having been provided prior nor at the time of selection, (iii) sufficient distance of at least 4 km between study schools, and schools or child-friendly spaces (CFSs) where TeamUp was provided, to avoid spillover.

The selected participants were aged 10–15 years, given the primary target population of TeamUp (6–17 year-olds), adequate developmental/cognitive capacity to respond to questions (child-reported outcomes) on wellbeing (≥10 years) and the reality of the age range present in grade 4 (ranging from 9–16 years). Children enrolled had sufficient understanding of one of the three primary local languages of Bidibidi refugee settlement (Bari, Juba Arabic and Aringa). Children with severe hearing, vision, and speech disabilities that significantly impaired participation in data collection (questionnaire administration) would have been excluded from data collection. However, no child met the these criteria. All children within the selected classes – in the intervention schools – were welcomed to participate in TeamUp, in line with the inclusive and open-structured nature of the intervention, while only those whose caregivers provided consent and who met the inclusion criteria were enrolled in the study. Of the TeamUp participants 3% (n = 8) had some kind of (mild) physical, visual or hearing impairment, but were not excluded.

Procedures

Twelve research assistants were selected based on prior experience and fluency in one to three local languages (Bari, Juba Arabic, Aringa) and trained for 5 days on research methods, ethics, procedures, data collection tools and electronic tablet manipulation, using scenarios and interactive role plays. Prior to data collection written caregiver consent (≥18 years and primary caregiver for child) and child assent was obtained. School directors and teachers supported the process by asking children of the selected classes to request their caregivers to come the following days to receive information about the study. About 78% of caregivers came (data collection was happening for 4–5 days per school) and all of those present provided consent. Children whose caregivers provided consent and who themselves provided assent, were invited to join the study (merely <2% of children did not assent or participate, thus enroll, in baseline data collection). All quantitative data was collected using electronic tablets equipped with the KoboCollect App (http://hhi.harvard.edu/research/kobotoolbox).

Data collection took place within the schools at baseline (for 3–4 weeks in April–May 2021) and endline was gathered in central places within the community (for 4 weeks in June–July 2021) due to COVID-19-related school closures at the time. Caregivers received a bar of soap to compensate for the time not spent on possible income-generating activities and children were offered a small snack (biscuit and juice) to recognize their time commitment.

Intervention

TeamUp is as a movement-based mental health promotion intervention developed for children affected by armed-conflict, violence, displacement and ongoing adversity (https://www.warchild.net/intervention-teamup/). The intervention consists of movement-based activities aiming to improve children’s psychosocial wellbeing by strengthening social connectedness, reducing stress and tension, as well as facilitating self-regulation and a positive outlook and, through creating positive experiences, developing their playing resources and offering a safe space where children are protected, heard and respected (see Theory of Change; Dieterich-Hartwell & Koch, 2017; Hobfoll et al., 2007; IASC, 2007; Ley & Rato Barrio, 2019; TeamUp, 2020b).

The movement-based activities combine group games, sports-based and physical activities, dance and creative movement, routines, songs, body awareness, and breathing practices (see Bleile et al., 2021; TeamUp, 2020a, 2020b). TeamUp’s theoretical underpinnings include trauma-informed care principles (Bergholz et al., 2016; Hobfoll et al., 2007), and the value of the body, movement and play to support socialization, attunement, mind-body connection and self-regulation abilities (Dieterich-Hartwell et al., 2020; Ley & Rato Barrio, 2019; Verreault, 2017). The intervention is in line with international guidelines on mental health and psychosocial support (MHPSS) (IASC, 2007) as well as creative and movement-based approaches used when engaging with refugee populations (Dieterich-Hartwell & Koch, 2017; Dieterich-Hartwell et al., 2020; Verreault, 2017).

Trained facilitators offer children a weekly 45- to 60-minute TeamUp session following a clear session structure with an opening, a check-in, a warm-up, main activities, a cooling down and a check-out (see HandBook: TeamUp, 2020b; and GameBook – TeamUp, 2020a; Dieterich-Hartwell et al., 2020). Each component in this structure comprises an activity, which facilitators can select from the GameBook. For example, warm-up activities would involve slower movements, stretching or breathing exercises; main activities would include those involving running, jumping, and possibly competing group/team games; and a cooling down would encompass calmer, breathing or other settling practices. This allows facilitators to include both energizing and more relaxing activities to assure settling and grounding opportunities while, i.e. integrating various types of games based on eight psychosocial themes (i.e. fear, conflict, respect, assertiveness, anger, bullying, friendship, stress; TeamUp, 2020b).

Children are offered to experience a safe space with trustworthy adults, engaging in social interaction with peers, strengthening playing resources while releasing stress in their bodies and experiencing settling and grounding (Bergholz et al., 2016; Dieterich-Hartwell et al., 2020; Ley & Rato Barrio, 2019; O’Brien & Charura, 2022; Porges, 2015). The trauma-informed framework provides children choices through the opportunity of trying new activities and movements, making adaptations or deciding to take a “time out” when desired. The sessions also encourage children to reconnect with themselves, their peers and supportive adults, by strengthening their experience of being in the present moment as well as their sense of safety, belonging and agency (Bareka et al., 2019; Hobfoll et al., 2007; Ley & Rato Barrio, 2019; Malchiodi, 2008; Van der Kolk, 1994, 2014).

A Theory of Change (see online supplement), a HandBook (TeamUp, 2020b) and GameBook (TeamUp, 2020a) were developed and provide facilitators with guidance as well as specific games for specific age ranges and psychosocial themes. TeamUp was originally developed for children in asylum seeker centers in the Netherlands in 2016; implemented, evaluated (Bleile et al., 2021) and strengthened. Since 2017, TeamUp has been adapted to various contexts, including Uganda, Sri Lanka, Palestine, and Colombia in collaboration with local teams and implementing staff, encompassing local songs, suitable games and routines.

The movement-based and non-verbal nature of the methodology, using demonstration and gestures, allows for an inclusive engagement with children and facilitators of multi-cultural and multi-lingual groups. Cultural adaptation of games, songs and routines according to the context is important and possible (Hassan et al., 2016; Ley & Rato Barrio, 2019; Shah, 2012; TeamUp, 2020b). TeamUp can be integrated into diverse settings, such as child-friendly spaces, schools or community services, and is implemented by trained community facilitators.

For the study, we recruited and trained 10 facilitators for 5 days in the TeamUp methodology. The training took an embodied, experiential learning approach, including session preparation, implementation, and evaluation practice, coupled with understanding the intervention’s theoretical underpinnings and child safeguarding. Facilitators were 18 years of age or older, from Bidibidi settlement or the surrounding host community area (30% were South Sudanese, 70% Ugandan; on average 27.7 years of age, 40% female). All had at least some secondary school education and prior experience providing psychosocial support or other recreational activities for children in their context (2.6 years on average). Given the peculiar COVID-19 situation and delayed timeline, which did not allow new facilitators to obtain practice experience ahead of the study (original plan), the implementing NGO War Child decided to re-hire facilitators who had previously provided TeamUp at home (COVID-19 adapted sessions whereby individual facilitators went from house/compound to house for 15–30 minute sessions with the child, siblings and family).

Sessions were originally integrated into the school curriculum by the school administration as an alternative to physical education (PE), yet affected by COVID-19. Sessions took place on the school’s playground/field. Each two facilitators (dyads) provided TeamUp sessions for 30–40 children (group) at a specific time during the school week in the TeamUp schools. Concurrently, yet physically separated, another two facilitators provided a session to another group of 30–40 children. Due to differing classroom sizes, one school had two groups (4 facilitators) and the other had three groups (6 facilitators) receiving sessions simultaneously. Each group received a 45-minute TeamUp session twice per week for a duration of six weeks with a maximum of 10 or 11 sessions, depending on the weekday of their sessions. Weekly in-person supervision sessions were offered in addition to monitoring using the mentioned checklists.

Instruments

The instruments were chosen to map closely onto the primary and intermediate outcomes stipulated in the Theory of Change (see TeamUp, 2020b; and online supplement). The primary outcomes were; (a) emotional and psychosocial wellbeing, measured using the Stirling Children’s Wellbeing Scale (SCWBS) (12 items, Cronbach’s α = .81 and 3 additional items measuring social desirability, 5-point scale, total score 12 to 60 points) with subscales positive outlook (6 items, α = .70) and emotional state (6 items, α = .70) (Liddle & Carter, 2015); (b) satisfaction with friendships, assessed with the Multidimensional Students Life Satisfaction Scale (MSLSS) Friends subscale (9 items, 5-point scale, total score 0 to 36, α = .73), and; (c) satisfaction with and attitude toward school, measured with the MSLSS School subscale (5 items, only the positively worded, 5-point scale, total score from 0 to 20, α = .64) (Huebner, 1994; Huebner & Gilman, 2002; Park et al., 2004).

Secondary outcomes included; (a) traumatic stress measured using the Children’s Revised Impact of Event Scale (CRIES-8) (8 items, 4-point scale, total score 0 to 40, α = .80) with subscales intrusion (4 items, α = .81) and avoidance (4 items, α = .74) (Giannopoulou et al., 2006; Perrin et al., 2005); (b) health-related quality of life assessing subjective wellbeing, health and functioning, measured using the KIDSCREEN-10 (10 items, 5-point scale, total score 10 to 50, α = .70) (The KIDSCREEN Group Europe, 2006; Ravens-Sieberer et al., 2010); (c) physical health using the subscales of the KIDSCREEN-52 of physical activities and health (5 items, α = .73); (d) bullying using the subscales of the KIDSCREEN-52 bullying (3 items, α = .63) (The KIDSCREEN Group Europe, 2006; Ravens-Sieberer et al., 2005) both on 5-point scales, with total scores 5 to 25 and, 3 to 15 respectively; interoceptive awareness, capturing the ability of noticing sensations in the body, assessed with the Multidimensional Assessment of Interoceptive Awareness (MAIA) – version 2, subscale Noticing (4 items, 5-point scale, total scores 4 to 20 α = .70) (Mehling et al., 2012); (f) irritability, encompassing the frequency, duration and threshold of angry reaction, measured using the Affective Reactive Index (ARI) (7 items, 3-point scale, total score 0 to 14 α = .75) (Stringaris et al., 2012), and; (g) depressive symptoms measured using the Short Moods and Feelings Questionnaire (SMFQ, 13 items, 3-point scale, total score 0 to 26, α = .84) (Angold et al., 1995; Costello & Angold, 1988; Eyre et al., 2021; Thabrew et al., 2018).

In addition, we developed a TeamUp-specific mechanisms of action scale, consisting of 28 items (α = .75, excluding 4 items, thus for 24 items, 5-point score, total score 24 to 150, α = .78), asking about children’s sense of safety, social connectedness, sense of agency, emotional regulation and interoception (i.e. the ability to notice ones bodily sensations). This tool, developed for the purpose of this study, assessed the use of coping strategies and socio-emotional skills that map onto the content of TeamUp and its Theory of Change, yet were formulated so that it can be scored independent of knowledge or experience of TeamUp, in order to gauge the degree by which participants adopt coping strategies underlying the TeamUp intervention (see Jordans et al., under review). Visual pictograms (images of glasses with different amounts of water) were used to aid children in providing their answer. Measures were all child-reported outcomes in order to directly assess the target group.

Most measures had been used with different populations and languages, yet none of them previously with our study population. All measures were systematically translated, back-translated and harmonized, and adapted to the linguistic and cultural context using cognitive interviewing (see Van Ommeren, 2003; Van Ommeren et al., 1999). The cognitive interviewing exercise was conducted with eight bilingual adults, speaking English and one of the three local languages (Bari, Aringa and Juba Arabic), all who had children within the age range of our study. Interacting with children for cognitive interviewing was not possible at the time due to COVID-19 restrictions. Given the dearth of validation studies in LMICs, we assessed as a minimum the internal consistency, i.e. Cronbach’s alpha, prior to the main analyses.

Demographic characteristics of participants and their caregivers, including age, educational level, nationality, gender of household-head, number of children in the household, duration of displacement and duration of separation from their child, were recorded at baseline. During implementation, implementation indicators were assessed, including (i) children’s TeamUp session attendance, assessed with attendance registers tracking individual children’s presence or absence per session; (ii) facilitators’ fidelity to the TeamUp methodology, measured at the individual facilitator level as the mean percentage of TeamUp elements “done,” “partly done,” and “not done” as well as at the facilitator team (dyad) level as the mean percentage of TeamUp elements using the same three ratings, and; (iii) individual facilitators’ TeamUp-specific competencies, again using the same rating. An observation checklist with 33 items, including 12, 15 and 6 items (3-point scale) were used to assess facilitators’ individual fidelity, team (dyad) fidelity and facilitators’ individual TeamUp-specific competencies, respectively. The observation checklist was previously developed and piloted in asylum seeker centers in the Netherlands (Bleile et al., 2021) and revised for the current study. The supervisor observed, rated and mentored the facilitator dyads 2–5 times within the six-week period. In total 16% of sessions were observed; the majority by one observer (TeamUp coordinator), yet two sessions (12%) were rated by two other trained mentors. Attendance was reported as the mean percentage of sessions attended by children as well as the percentage of children attending each of the sessions.

Statistical Analysis

Prior to the main analyses, assumption checks were conducted on the distribution of the scores, outliers and missing values. The assumption of normality held for all the outcomes. There were only a few outliers and items with a missing value; the outliers were not removed and the missing values were replaced by the median of that item. Additionally, the internal consistency of each outcome and corresponding subscales was calculated with Cronbach’s alpha (reported above). Because there were four schools, two in each arm, and the data of groups of children were collected by a research assistant, the intra-class correlation (ICC) of each outcome was checked to detect clustering at the school and research assistant level. The ICCs had values ranging from almost zero to .05, with most around .02. Therefore, the main analysis was a linear regression with endline scores as the response variable, group (TeamUp vs control) as the predictor and baseline score as covariate. All the analyses were conducted using an intention-to-treat (ITT) approach, dealing with missing values at the endline by multiple imputations (5 datasets).

The additional analyses concerned per protocol analyses on all the outcomes, including only those children who had both baseline and endline data (n = 510), and subgroup analyses on the primary outcomes; (a) excluding children with high social desirability scores (see SCWBS), (b) split by gender, (c) split by nationality, and; (d) only including children with high scores (≥12) on the SMFQ, indicating depressive symptoms (Eyre et al., 2021; Thabrew et al., 2018). All the analyses were done in SPSS version 26.

Ethics

The study protocol was submitted to and approved by the Makerere University School of Health Sciences Institutional Ethics Review Board (MAKSHS-REC) and the Uganda National Council for Science and Technology (UNCST) in Kampala, Uganda (reference number HS941ES). Caregiver consent and child assent was obtained prior to data collection.

Results

Demographics

The participants were 549 children aged 10–15 years (M_age = 13.4, SD = 1.2), 261 boys (47%), 286 girls (52%), and 2 children (1%) with a missing value on gender. The majority of children were South Sudanese (87%), some Ugandan nationals (11%) or other (2%). A total of 505 caregivers (M_age = 37.2, SD = 12.1; 124 (25%) males, 347 (69%) females, 34 (6%) missing value) were asked for demographic characteristics of themselves and their child(ren) participating in the study. Almost all (99%) had an adult caregiver as head of the household with only 1% with child-headed households. Based on child-report, households contained on average 4.4 children (SD = 2.7). Of the 114 cases where the child had been separated from their caregiver(s), this was on average for a duration of 18.6 months (SD = 23.5). Table 1 gives a summary of the child and caregiver demographics for the total sample and split into TeamUp and control groups, showing that only caregiver nationality was not equally distributed across TeamUp and control groups. This was not expected to influence the child outcomes and was not included in subsequent analyses.

Table 1. Baseline Demographics of Caregivers and Children, total sample and TeamUp vs control.

Demographic	Total sample	TeamUp	Control	χ^2(df)
Caregiver education				7.3 (3)
no schooling	149 (29%)	84 (33%)	65 (26%)
primary (completed or not)	263 (52%)	129 (51%)	134 (53%)
secondary (completed or not)	75 (15%)	29 (12%)	46 (18%)
certificate, degree or other	18 (4%)	11 (4%)	7 (3%)
Caregiver nationality				4.1 (1)^*
Ugandan	31 (6%)	21 (8%)	10 (4%)
South Sudanese	474 (94%)	232 (92%)	242 (96%)
Relationship with child				6.7 (5)
mother	233 (43%)	119 (45%)	114 (40%)
father	84 (15%)	43 (16%)	41 (14%)
sibling	64 (12%)	33 (13%)	31 (11%)
other family	150 (27%)	63 (24%)	87 (31%)
other type	12 (2%)	4 (1%)	8 (3%)
missing	6 (1%)	3 (1%)	3 (1%)
Child nationality				4.7 (2)
Ugandan	59 (11%)	36 (14%)	23 (8%)
South Sudanese	479 (87%)	223 (84%)	256 (90%)
other or missing	11 (2%)	6 (2%)	5 (2%)
Child separated from caregiver				1.1 (1)
No	428 (78%)	207 (78%)	221 (78%)
Yes	114 (21%)	54 (20%)	60 (21%)
missing	7 (1%)	4(2%)	3 (1%)

*p < .05.

Outcome Measures

The results of the main analysis following an ITT approach (see Table 2) showed a significant effect of TeamUp on emotional and psychosocial wellbeing (M_diff = −1.49, SE = 0.6, p = .01) and satisfaction with and attitude toward school (M_diff = −0.57, SE = 0.2, p = .004). Of the emotional and psychosocial well-being subscales, only emotional state showed a significant effect of TeamUp (M_diff = −1.02, SE = 0.3, p < .001), while positive outlook was not. The differences in friendships were similar for the TeamUp and control groups. Effect sizes (Cohen’s d) for primary outcomes ranged from d = 0.22 to 0.29. Regarding the secondary outcomes, TeamUp showed a significant effect on reduction in traumatic stress (M_diff  =  2.64, SE = 0.8, p < .001), and on both the subscales intrusion (M_diff = 1.29, SE = 0.5, p = .005) and avoidance (M_diff = 1.22, SE = 0.5, p = .01). Similarly, a significant between-group difference was seen in health-related quality of life (M_diff = −1.56, SE = 0.4, p=.001) as well as physical health (M_diff = −0.78, SE = 0.3, p = .014). Furthermore, the effect of TeamUp on the total score on TeamUp mechanisms of action was significant (M_diff = −3.34, SE = 0.9, p < .001), specifically on both the subscales social connectedness (M_diff = −0.74, SE = 0.3, p = .007) and sense of agency (M_diff = −0.91, SE = 0.3, p = .005). Effect sizes for secondary outcomes ranged from d = 0.21 to 0.31. The TeamUp and control groups did not show significant differences on bullying, interoceptive awareness, irritability and depressive symptoms.

Table 2. Descriptive statistics and estimated treatment effects of the outcome measures per group and timepoint, following intention-to-treat (ITT) analyses.

Outcome measures	construct	Time point	Mean (SD)			Estimated Treatment Effect
			Total (n = 549)	Control (n = 284)	TeamUp (n = 265)	Mean Difference* (95% CI)	Cohen’s d	significance
Primary outcomes
SCWBS	Emotional and psychological wellbeing	T0 T1	30.50 (7.7) 30.49 (6.7)	30.89 (7.4) 29.83 (6.5)	30.09 (8.1) 31.18 (6.9)	−1.49 (−2.65;-0.33)	0.22	.012
Positive outlook	Positive outlook	T0 T1	14.88 (4.5) 14.84 (4.1)	15.25 (4.4) 14.69 (3.9)	14.48 (4.7) 15.01 (4.2)	−0.44(−1.14;0.26)	0.10	.222
Emotional state	Emotional state	T0 T1	15.62 (4.1) 15.64 (3.5)	15.64 (3.8) 15.15 (3.4)	15.60 (4.3) 16.17 (3.5)	−1.02 (−1.63;-0.42)	0.29	.001
MSLSS Friends	Friendships	T0 T1	25.12 (5.9) 24.98 (5.6)	24.88 (6.0) 24.57 (5.6)	25.39 (5.9) 25.42 (5.5)	−0.74(−1.70;0.22)	0.13	.085
MSLSS School – positive items only	Satisfaction with and attitude toward school	T0 T1	17.15 (2.6) 17.21 (2.4)	17.18 (2.4) 16.94 (2.4)	17.12 (2.8) 17.51 (2.3)	−0.57(−0.97;-0.18)	0.24	.004
Secondary outcomes
CRIES-8	Traumatic stress	T0 T1	14.03 (9.6) 13.41 (9.1)	14.38 (9.4) 14.74 (8.8)	13.66 (9.7) 11.99 (9.2)	2.64(1.13;4.15)	0.29	.001
Intrusion	Intrusion	T0 T1	5.99 (5.6) 5.55 (5.4)	6.44 (5.9) 6.29 (5.3)	5.50 (5.2)4.75 (5.5)	1.29(0.40;2.18)	0.24	.005
Avoidance	Avoidance	T0T1	8.05 (5.9)7.86 (5.7)	7.94 (5.6)8.45 (5.6)	8.16 (6.2)7.24 (5.8)	1.22(0.27;2.18)	0.21	.012
KIDSCREEN – Health	Health-related quality of life	T0 T1	27.91 (5.6)28.12 (5.1)	27.86 (5.3)27.35 (4.9)	27.97 (5.9)28.94 (5.2)	−1.56(−2.44;-0.68)	0.30	.001
KIDSCREEN – Phys.Act. Health	Physical health and activity	T0 T1	13.72 (3.7) 13.97 (3.6)	13.98 (3.7) 13.66 (3.6)	13.43 (3.7)14.30 (3.7)	−0.78(−1.41;-0.16)	0.21	.014
Bullying	Bullying	T0 T1	3.93 (2.7) 3.76 (2.6)	4.08 (2.8) 3.95 (2.6)	3.79 (2.6) 3.57 (2.8)	0.34 (−0.13;0.80)	0.12	.153
MAIA – noticing	Interoception	T0 T1	10.39 (3.5) 10.61 (3.0)	10.42 (3.5) 10.51 (3.0)	10.35 (3.5) 10.70 (3.0)	−0.20 (−0.72;0.31)	0.07	.437
ARI	Irritability	T0 T1	3.94 (3.1) 4.41 (3.2)	4.67 (3.1) 4.81 (3.1)	3.16 (2.9) 4.00 (3.2)	0.44 (−0.12;1.00)	0.13	.124
SMFQ	Depression symptoms	T0 T1	7.62 (5.4) 7.86 (5.6)	8.36 (5.5) 8.52 (5.6)	6.84 (5.2) 7.16 (5.4)	0.92 (−0.07;1.92)	0.15	.069
TeamUp mechanisms of action scale	TeamUp mechanisms of action scale	T0 T1	60.52 (12.0) 61.23 (10.7)	59.86 (11.4) 59.42 (10.2)	61.22 (12.6) 63.16 (11.0)	−3.34 (−5.19;-1.48)	0.31	.001
Subscale social connectedness	Social connectedness	T0 T1	9.36 (3.8) 10.04 (3.2)	9.19 (3.6) 9.65 (3.0)	9.53 (3.9) 10.46 (3.3)	−0.74 (−1.27;-0.21)	0.23	.007
Subscale sense of agency	Sense of agency	T0 T1	10.92 (4.0) 10.75 (3.6)	10.89 (4.0) 10.31 (3.5)	10.95 (4.1) 11.22 (3.6)	−0.91 (−1.54;-0.27)	0.24	.005

T0 = baseline, T1 = endline; *Mean difference at endline is corrected for baseline scores.

SCWBS = Stirling Children’s Wellbeing Scale; MSLSS = Multidimensional Students Life Satisfaction Scale; CRIES-8 = Children’s Revised Impact of Event Scale; MAIA = Multidimensional Assessment of Interoceptive Awareness; ARI = Affective Reactive Index; SMFQ = Short Moods and Feelings Questionnaire.

The per protocol analysis including only the n = 510 children with both baseline and endline data (n_control = 260 and n_TeamUp = 250) showed very similar results on all the outcomes (see supplemental Table A) as the ITT results above. Thus, they also demonstrated significant improvements on emotional and psychological wellbeing (including emotional state), satisfaction with and attitude toward school, traumatic stress (including intrusion and avoidance), health-related quality of life, physical health and activity, TeamUp mechanisms of action scale (including social connectedness and sense of agency) with p-values between <.001 and .015. No significant differences were found for positive outlook, friendships, bullying, interoception, irritability and depressive symptoms with all p values > .069).

When splitting the sample into boys (238 in total; 120 control and 118 TeamUp) and girls (270 in total; 140 control and 130 TeamUp) the effect of TeamUp was stronger for boys compared to the total sample and significant on all primary outcomes (all p-values between <.001 and .030), while for the girls none of these outcomes were significant (all p-values between .324 and .810). The 125 children (22.8%) with high levels of depressive symptoms (SMFQ scores ≥12; 82 control and 43 TeamUp) did not seem to benefit from TeamUp in terms of emotional and psychosocial wellbeing (M_diff = −0.55, SE = 1.3, p = .667) and satisfaction with and attitude toward school (M_diff = −0.22, SE = 0.5, p = .689), but they did appear to benefit from TeamUp in terms of friendships (M_diff = −2.19, SE = 1.0, p = .038), showing the opposite pattern than the total sample. Both subgroup analyses were underpowered, thus should be interpreted with caution and no firm conclusions can be drawn. All the descriptive and test statistics are given in supplemental Table B.

Implementation Indicators

Children’s attendance, facilitator’s adherence and competencies were overall high. Children’s session attendance was on average, 91.3% (n = 250; n = 133 girls, n = 117 boys; n = 234 refugee children, n = 16 host community children). The large majority (91.6%) attended 80–100% of the sessions (i.e. 8, 9 or 10 out of 10, or 9, 10 or 11 out of 11 sessions) across the six week period, and attendance percentage varied between 73.63% and 96.8% per session, with no large differences between subgroup (gender or refugee/host community). There was a clear dip in the attendance for the 5^th session, likely due to multiple public holidays within that week.

Facilitator fidelity was on average 97.1% across individual facilitators and 96.2% across teams (dyads), and TeamUp-specific competencies on average 97.2%. Specifically, the average adherence percentages of individual facilitators for “partly done” was 6% and “well done” was 94%, very similar to the teams with the average percentages for “partly done” of 8% and “well done” of 92%. For the TeamUp-specific competencies, the average “partly done” was 3% and “well done” was 97%.

Masking, Adverse Events and Referrals

Research assistant masking was maintained. At endline 47.2% and 43.1% of the RAs guessed the TeamUp group and control group allocation correctly, respectively, which is below 50% pure chance. Also, 24.3% of RAs reported “I do not know” when asked about the child’s group allocation. Twelve adverse events were reported, using standard operating procedures and relevant forms, explained during training days. In total, 9 child protection concerns and 3 grief cases were reported to the study team, and referred to other medical, social and psychosocial services. None of the cases were perceived as linked to the study or intervention by the core study team.

Discussion

This quasi-experimental study demonstrates that children participating in TeamUp sessions showed significant improvements on; (i) emotional and psychosocial wellbeing, (ii) satisfaction with and attitude toward school, (iii) health-related quality of life, (iv) physical health and (v) reduction in traumatic stress, in comparison to the control group. We also found a significant improvement on the TeamUp mechanisms of action scale and its subscales (vi) social connectedness and (vii) sense of agency. No differences were shown at endline for friendships, bullying, interoceptive awareness, irritability and depressive symptoms. This discussion will address some of the main findings.

These are promising results for the potential effectiveness of TeamUp. The findings largely align and support the intervention’s Theory of Change and fit the conceptual model TeamUp as a mental health promotion intervention. Supporting that argument, the control group actually showed a reduction in emotional and psychosocial wellbeing from baseline to endline.

These findings are particularly relevant given the absence of strong evidence for mental health promotion (Barbui et al., 2020; Barry et al., 2013; O’Reilly et al., 2018; Purgato et al., 2018; Tol, 2015; Uppendahl et al., 2020) and high demand of evidence-based interventions in the field of MHPSS in humanitarian settings and global mental health work more broadly (Jordans et al., 2016; Patel et al., 2018; Tol et al., 2011; Trimboli et al., 2021; Venturo-Conerly et al., 2022). The effect sizes are small to moderate, in line with mental health promotion interventions (Barbui et al., 2020; Burkey et al., 2018; Klasen & Crombag, 2013; Purgato et al., 2018), and are meaningful improvements from a public health perspective wherein a small enhancement following a universal intervention can shift the dial at a population level. Baseline wellbeing scores of the entire sample were lower than the mean average indicated in the original tool (Liddle & Carter, 2015), yet not assessed nor validated for our study population.

The TeamUp is among the few evaluated interventions taking a non-verbal and movement-based approach, that may serve as an important addition to interventions that rely on verbal and cognitive processing (Jordans et al., 2016; Purgato et al., 2018). Our results correspond and build onto existing literature in the field (Levine & Kline, 2006; Ley & Rato Barrio, 2019; Marley & Mauki, 2019; Porges, 2015; Van der Kolk, 1994, 2014; Verreault, 2017). Especially, on to the importance and value of play for the reduction of negative affect and stress, while stimulating the development of empathy and coping with adversity (Nijhof et al., 2018) as well as the benefit of body-mind (re)connection (Adaku et al., 2016; Coker, 2004), movement and sports-based group/peer interventions for global mental health settings that are multicultural/-lingual (Allen & Kern, 2017; Cleary et al., 2018; Dieterich-Hartwell et al., 2020; Hassan et al., 2016; Ley & Rato Barrio, 2019; O’Brien & Charura, 2022; Shah, 2012).

The socialization opportunities and connection with peers, provided through TeamUp, have shown to alleviate stress, promote wellbeing in children and adults (Bleile et al., 2021; Ley & Rato Barrio, 2019; Porges, 2005, 2015; Verreault, 2017) by fostering a sense of belonging and social cohesion (Allen & Kern, 2017; Cleary et al., 2018). The relative safe environment for playing and moving, guided by supportive, caring and trustworthy adults offers children to explore and experience predictability and safety (Levine & Kline, 2006; Ley & Rato Barrio, 2019; Marley & Mauki, 2019). Thereby, TeamUp’s components and practices such as following a clear and welcoming structure, and the encouragement of settling and grounding moments in a safe environment (see TeamUp, 2020a, 2020b), may not only explain the improvement in wellbeing and health-related quality of life (Bareka et al., 2019; Hobfoll et al., 2007; Levine & Kline, 2006; O’Brien & Charura, 2022), but also children’s increase in satisfaction with and attitude toward school.

The clear benefits in reduction of traumatic stress for children participating in TeamUp may be explained by the intervention taking a trauma-informed movement- and body-based approach (Bergholz et al., 2016; Dieterich-Hartwell et al., 2020; Hobfoll et al., 2007; Ley & Rato Barrio, 2019; Porges, 2015; Van der Kolk, 1994, 2014). Processing sensations end experiences somatically (through the body) has shown to support children in modulating their energy, space and body awareness (Bergholz et al., 2016; Ley & Rato Barrio, 2019; Rothschild, 2000) and to foster children’s abilities to learn, connect and thrive (Levine & Kline, 2006; Malchiodi, 2008). TeamUp offers children the opportunity to be distracted, be in their bodies and alleviate mental and physical tension through engaging their bodies (Adaku et al., 2016; Coker, 2004). This may allow them to strengthen and sustain their social engagement capacities (connecting with others), encouraging them through playful ways to regulate and move away from automatic traumatic responses, such as “flight, fight and freeze” (Levine & Kline, 2006; Porges, 2015; Van der Kolk, 1994, 2014). By supporting the children to move their bodies in different ways, offering various activities (e.g. dynamic game, slow-motion movement, relaxation), children may experience more integrated ways of being, regulating the autonomic nervous system (ANS) and diminishing traumatic symptoms while developing new movement patterns (Dieterich-Hartwell et al., 2020; Homann, 2010; Ley & Rato Barrio, 2019; Porges, 2015; Van der Kolk, 1994). With the study population evidently having been and currently being faced with high levels of adversity and distress (the sample’s CRIES-8 mean = 14 points being close to the proposed cutoff ≥17; Perrin et al., 2005) as well as limited resources and services, a reduction in traumatic symptoms appears likely with merely a low-intensity mental health promotion intervention.

The lack of between-group differences on depression symptoms and irritability as well as the lack of improvement for children with high levels of depression symptoms (SMFQ ≥ 12; Eyre et al., 2021; Thabrew et al., 2018) confirms that such symptom reduction falls outside the scope of a mental health promotion intervention (Barbui et al., 2020; Purgato et al., 2018). Mean depressive symptoms at baseline and the percentage of children with high depressive symptoms (22.8% showing SMFQ ≥ 12) were in line with recent prevalence estimations of common mental health disorders of youth in LMICs (Jörns-Presentati et al., 2021; Yatham et al., 2018). This also highlights TeamUp to be implemented in conjunction with interventions that cater for children with more severe symptomatology and within a stepped care model (Dawson et al., 2019; Jordans & Kohrt, 2020; Jordans, Tol et al., 2010; Jordans et al., 2018), including strong identification with tools such as ReachNow (previously known as the community case detection tool (CCDT); see Van den Broek, Hegazi, et al. (2023); Van den Broek, Ponniah, et al. (2021)) and a functioning referral mechanisms and support systems (Esponda et al., 2020; IASC, 2007; Tol et al., 2015). The subgroup analyses of children with high levels of depressive symptoms (SMFQ ≥ 12; Eyre et al., 2021; Thabrew et al., 2018) showing no effects on primary outcomes except friendships, further confirms this interpretation. Further research may investigate whether a higher dosage (increased number of sessions) and/or prolonged session implementation may be of (more) benefit to these children.

Not finding between-group differences on friendships and bullying was unexpected, yet might be explained by children already building (or having built) friendships while being at school, rather than during TeamUp per se. This might be different from implementation of TeamUp in asylum seeker centers or other child-friendly spaces where rotation/turnover of children is more common. Bullying entails a complex phenomenon that may require a whole-school approach and extended time of up to one to five years (Nickerson et al., 2013; Smith et al., 2005).

Subgroup analyses demonstrated that TeamUp primarily improved outcomes for boys, unlike other MHPSS interventions showing more benefits for girls (Ager et al., 2011; Bolton et al., 2007) or differences for different outcomes, i.e. aggression reduction for boys and promotion of prosocial behaviors for girls (Jordans, Komproe, et al., 2010). This might be due to the more active, dynamic approach of the intervention, particularly supporting boys who more commonly externalize distress (Jordans, Komproe, et al., 2010), while girls might require more verbal, small-group support such as interpersonal therapy (Bolton et al., 2007) or potentially with more cognitive-behavioral techniques (Tol et al., 2008). Consequently, this highlights the need for increased attention to girls and their needs during implementation, post-session debriefing and monitoring, in order to also effectively promote their wellbeing.

TeamUp session attendance was high. This may be explained by the “catch-up term” nature during which our study was conducted in, resulting from a re-opening of schools after a prolonged closure due to the COVID-19 pandemic outbreak and in turn increased commitment to learning and school from caregivers and children. This was similar across all four schools. Facilitators’ fidelity and competencies were also high, and so were retention rates. These positive implementation indicators – as called for by the literature on task-sharing MHPSS (Rose et al., 2022) – are noteworthy as the study was conducted amidst the pandemic. Certainly, for reasons of sustainability, scalability and system integration, having teachers deliver TeamUp rather than community facilitators may be further explored and studied in the future.

The strength of this study is the overall rigor of this quasi-experimental study into a mental health promotion intervention. The main difference with a cluster RCT is that we did not power and adjust for clustering in this study. Yet low ICCs at schools and research assistant level indicate that the found effects between the study arms were not biased by school nor interviewer selection. Therefore, it is fair to conclude that the found effects are likely caused by the only difference between the two study arms, i.e. the implementation of TeamUp. The study also involved a non-active control condition – although consisting of irregular or no formal PE lessons, but rather children’s self-initiated, free play, for both study arms. It thereby allowed evaluating TeamUp as an added value to the status quo in refugee primary schools in West-Nile, Uganda. Still, a fully powered cluster RCT is now needed to determine TeamUp’s effectiveness, taking into account clustering at school and group level. Additional research may examine longer term effects and underlying mechanisms (Jordans et al., under review), and cost-effectiveness to evaluate the potential of TeamUp for the humanitarian sector. A study in another setting is needed to evaluate the generalizability of the study findings for other contexts and populations.

The limitations of our study may include that none of our instruments’ construct validity was assessed for the specific study population, as this was outside the scope of our study. However, psychometric properties were assessed for all of our instruments and only those with strong Cronbach alpha values were included in the final analyses. Besides, the cutoff (≥12) on the SMFQ, indicating depression symptoms (Eyre et al., 2021; Thabrew et al., 2018) was not validated for our study population, yet merely used as an indication for our subgroup analyses. The inter-rater reliability was not assessed for observations of facilitator fidelity and competencies. Furthermore, we did not exactly monitor the activities of the control group, thus are unsure about the activities or “dosage” of physical/movement-based activities received by them. Moreover, we cannot be sure what the impact of COVID-19 might have had on our results, as the study was implemented in a very tight window of re-opening of schools in Uganda. Schools were closed for over one year (from March 2020), re-opened (April 2021), and closed again following a sudden surge in cases after two months (June 2021). This unprecedented situation may have influenced children’s attendance, motivation and interest in school and TeamUp, and possibly their scoring on outcome measures. As COVID-19 affected the entire nation and globe, any impact was likely the same for all schools and both study arms.

Conclusion

A six-week TeamUp movement-based mental health promotion intervention demonstrated positive and promising results in improving children’s emotional and psychosocial wellbeing, satisfaction with and attitude toward school, health-related quality of life, physical health, traumatic stress as well as social connectedness and sense of agency – compared to a control group. Children (re)created positive experiences, release stress and tension and strengthened social interaction, while developing play and movement-based capabilities. The results strengthen a much needed evidence base of mental health promotion interventions in humanitarian contexts. More specifically, our study highlights the potential for interventions that include play, the body and movement for strengthening, activating and restoring children’s resources, based on psychosocial and trauma-informed frameworks. The detected effects are small, as can be expected for mental health promotion interventions and are meaningful from a public health perspective. Further research will need to assess the effectiveness, sustainability of effects and cost-effectiveness of the intervention to inform scaling and assess TeamUp’s potential for promoting the psychosocial wellbeing of children within humanitarian settings.

Author Contributors

MJ and AB designed the study. BO and AB were responsible for the overall coordination and oversaw the data collection. MJ, GKG and AB developed the data analysis plan, GKG and AB analyzed the data. AB drafted the manuscript, MJ and KV critically revised the manuscript. All authors commented on and approved the draft and final manuscript.

Acknowledgments

We thank all children participating in our study, school teachers, head-teachers and community leaders (mobilisers) and our dedicated research assistants (ordered alphabetically): Bakhit Bejamin, Caroline Dawa, Emmanuel Ngota, Felix Limo, Gire Nenisa, Hashim Muzamil Bruce, Henry Puni, Jamal Rajab, Kevin Amviko, Oliver Longa Lodu, Shamimu Aligah, Simon Wayi John. We highly appreciate insights we received from Rosco Kasujja and Samuel Ouma (Makerere University), Emmanuel Ngabirano (TPO Uganda), Kevin Aciro (Save the Children Uganda) and Mathew Atibuni (Lutheran World Federation). We also thank the TeamUp facilitators and supervisors, the TeamUp Global team, and War Child staff in Uganda for all their support.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary data

Supplemental material for this article can be accessed online at https://doi.org/10.1080/15374416.2024.2330073

Data availability statement

Data sets are available upon request via the corresponding author.

Additional information

Funding

This project was made possible through the support of a grant from Templeton World Charity Foundation, Inc. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of Templeton World Charity Foundation, Inc.