The first aim of research project is to test whether two school-based psychosocial intervention arms, (a) the combined In-Service Teacher Training (INSETT) and Teaching Recovery Techniques (TRT) and/or (b) the Peer Integration and Enhancement Resource (PIER), are effective in improving the mental health among refugee, migrant, and Finnish native adolescents. The criteria for intervention effectiveness are reducing psychological distress (internalizing and externalizing symptoms) and increasing prosocial behavior and resilience compared to adolescents in control schools. The second aim is to examine which adolescent- and/or school-related factors could explain the effectiveness of these two arms of school-based psychosocial interventions (Mediation analysis). Third, we analyze whether the adolescent- and family-related preconditions differ in the effectiveness of these two intervention arms (Moderation analysis).
The research hypotheses are:
1. Do the psychosocial school interventions have a positive effect on the mental health of immigrant adolescents? In more detail, the aim is to compare the impact of INSETT+TRT and PIER to waiting-list control group on internalizing and externalizing symptoms and prosocial behavior and resilience.
1.1 We hypothesize that internalizing and externalizing symptoms will decrease statistically significantly only among adolescents participating in the INSETT+TRT and PIER interventions and not in the control group from baseline (T1) through six-month (T2) and twelve-month (T3) follow-ups.
1.2 We hypothesize that prosocial behavior and resilience will increase only among adolescents participating in the two arms of interventions and not in the control group.
2. Do different adolescent- and school-related factors explain (mediate) the effects of the INSETT and PIER interventions on adolescent mental health?
2.1. Concerning INSETT, we hypothesize that participation of teachers increases their multicultural awareness, sense of self-efficacy and work engagement, and decreases their work stress, which in turn is associated with the decreased internalizing and externalizing symptoms and increased prosocial behavior and resilience among the adolescents.
2.2. We hypothesize that adolescents’ participation in PIER intervention is associated with their increased social support, feeling of belongingness, number of inter-ethnic friendships, and satisfaction on friendships, which in turn is associated with the decrease of internalizing and externalizing symptoms and the increase of prosocial behavior and resilience.
2.3. We hypothesize that adolescents’ participation in TRT intervention is associated with decreased PTSD symptoms of intrusion, avoidance and hyperarousal, which in turn is associated with decreased internalizing and externalizing symptoms and increased prosocial behavior and resilience.
3. How do parent- and adolescent-related factors moderate the effectiveness of the two arms of school-based psychosocial interventions of INSET+TRT and PIER?
3.1 We hypothesize that good parental mental health and parents’ high sense of competence and confidence in parenting are associated with statistically significant positive intervention-induced change in adolescents’ mental health (i.e., decrease in internalizing and externalizing symptoms and increase in prosocial behavior and resilience).
3.2 We hypothesize that adolescents who report low severity of daily stressors and low perceived discrimination show more statistically significant positive intervention-induced change in their mental health (i.e., decreased internalizing and externalizing symptoms, and increased prosocial behavior and resilience).
Trial design
The study is a three-arm clustered parallel assignment, multi-center, exploratory randomized, controlled trial (RCT) comparing INSETT+TRT and PIER to waiting list conditions located in 16 schools in Finland. Primary unit of analysis is the school. The schools were block-randomized(44) in a parallel-group condition with varying block sizes and an allocation ratio of 1:1. The voluntary schools with (intensive) immigrant student population were randomly allocated to different intervention and control conditions. Blinding of participants, intervention providers, outcome assessors or data analysts was not possible due to explicit nature of interventions and small number of research staff. The control schools will receive the INSETT intervention toolkit after the research part of the project has ended. Trial aims to follow SPIRIT guidelines in reporting. The table 1 below is designed as comparable to the SPIRIT figure and the checklist can be found as an attachment.
Study setting and participants
Our aim was to have 15 schools that have altogether 250 immigrant and 250 native secondary school pupils (age 13─15 years) to participate in each of the two intervention conditions. We focused on immigrant youth who have resided in Finland for less than 6 years. In addition, a control group of 250 immigrant and 250 native youths would be gathered from 15 comparable schools. The EU-research group has made the power analyses that serve as motivation for the sample sizes. Figure 1. presents the flowchart of recruitment. 16 schools expressed their interest in participating the study. According to the randomization eight schools were allocated into INSETT+TRT intervention, three schools were allocated into PIER and five schools serve as waiting-list controls. Also, Figure 1. reports the numbers of students, teachers and parents who answered the survey in the first round of data collection.
Recruitment and Sampling
The recruitment of schools was based on three aspects. First, voluntariness i.e., their (headmasters and teachers’) self-defined need for intervention, knowledge and training on psychosocial skills implemented in teaching refugee and immigrant adolescent students. For that purpose, the first recruitment was through advertisement in the national teachers’ magazine (Opettaja) and through email lists to school administration in every municipality in Finland. Second, the recruited and voluntary schools should have a considerable share of refugee and immigrant background pupils in secondary school classes, estimated as 30─50 percent (%) in each class. Third, the schools accepted the possibility of serving as a waiting-list school until the next school year when they would be provided the INSETT intervention in electronic form. This primary recruitment was complemented with about 100 phone calls to headmasters of schools residing close to the asylum seekers’ centers in numerous towns and municipalities in Finland. In addition, phone calls and e-mail communication were established to invite schools from metropolitan area due to their large share of adolescents with refugee and immigrant background. As a whole, 16 schools participated, and most extra work to reach schools near asylum-seeker centers and metropolitan area was not successful as schools informed that they already had a number of general interventions (such as anti-bullied whole-school programs) or trauma and refugee focusing programs (such as cultural couching programs).
Table 1. presents the timeline for implementing interventions, collecting data (T1, T2, T3) and providing the intervention for waiting-list control schools. Baseline data was collected before the interventions, INSETT+TRT and PIER interventions were implemented between August 2019 and January 2020, post-intervention data collection happened around February 2020 and the follow-up data will be collected August 2020. INSETT will be provided to waiting-list control schools October 2020.
Table 1
Timeline of Data Collection and Intervention Participation according to School Randomization (comparable to SPIRIT Figure 1)
Eligibility criteria
The unit of recruitment was schools. As both the INSETT+TRT and PIER interventions aim at enhancing school belonging, high peer quality, interethnic friendships and understanding and protecting cultural diversity, we included both native-Finnish and refugee-immigrant adolescents to the interventions and effectiveness study. The only exclusion criterion for schools is not to have other similar psychosocial interventions running at the same time. For students there were not exclusion criteria and all signing the informed consent could participate. The 16 schools locate across the country of Finland and represent mainly refugee and immigrant dense urban areas. One school is from the South, (the metropolitan area), two locate in the East, three in the Middle, two in the Western and three in the Northern Finland. Whether a participant would not want to continue participating the study, s/he may do so simply informing the researchers and the given participant data will be deleted from the system. If the participation in an intervention causes harm to a participant, then s/he will be advised to appropriate services. Reported experiences of caused harm will be stored anonymously and described in results.
The Interventions
In-service Teacher Training (INSETT) is delivered by lower secondary school teachers and counsellors in introductory-, preparatory-, or ordinary classes. The INSETT aims at enhancing teachers’ competence and self-efficacy in three key areas: (1) Promoting and supporting mental health and well-being among trauma-affected students with refugee and migration backgrounds. (2) Encouraging positive interethnic relationships in linguistically and culturally diverse school settings. (3) Fostering relationships with parents/caregivers to promote co-operation and involvement with schoolwork. The teachers’ training and awareness building is expected to benefit wellbeing and sense of belonging of refugee- and immigrant students directly and their caregivers indirectly.
Practically the intervention involves a combination of two seminar days with lectures, discussions, exercises, and exchange of ideas and experiences between participating teachers, and an online training course to be completed by the teachers individually. The INSETT manual provides seminar topics, materials and literature to prepare thematic sessions, and slides(45) and the online training course “Providing support to refugee minors” by the Augeo Foundation in the Netherlands(46) extensive local-language material to schools in Finland, Norway and Sweden.
Peer Integration and Enhancement Resource (PIER) is delivered by lower and upper secondary school teachers, counsellors, and special teachers in school classes including both native and refugee- and immigrant students. The PIER aims at supporting safe and positive peer interactions and social relationships in multi-ethnic schools by strengthening sense of belonging, empathy, role-taking and learning from each other and get and give more social support through group exercises. The manualized intervention consists of eight sessions ranging between 45 to 90 minutes, with structured welcoming and ending-rituals, multimodal group activities such as cartoon drawing, role-play, movies, drama, and ways of reflecting various identities, migration and racism.
The school staff delivering the PIER participated in a two-day training. The first included practicing each intervention session and going through the resource material and the second sharing of experiences and providing the RWS trainers (researchers) with valuable ideas to improve the intervention.
Teaching Recovery Techniques (TRT) is a manualized group treatment developed by the Children and War foundation(47), based on TF-CBT. The aims are to reduce PTSD symptoms of intrusion, avoidance and hypervigilance, increase resilience, stabilize trauma reactions, and provide practical techniques and strategies to deal with traumatic memories, physical and behavioral arousal and withdrawal. Treatment elements include psychoeducation (in playful and multimodal ways), normalizing reactions to trauma, working with nightmares, framing techniques to master intrusive memories and trauma reminders and scaling techniques to deal with avoidance and arousals. Homework such as sleep hygiene or reflective observation of trauma reminders are an essential part of the TRT.
The TRT consists of five sessions of 90 to 120 minutes including skills training, rehearsal and homework, and the handbook provides each session with several tools, techniques and procedures. The core aim is to create a sense of safety, increase feelings of competence and shared hope. The two parent/caregiver sessions include information about various responses to traumatic events and effective coping strategies. School personnel delivering the TRT participate in two-day intensive training provided by a licensed trainer from Children and War Foundation or from Finland. Qualified trauma psychologists offer TRT providers work-counselling during the intervention.
Data collection and management
Survey data is collected fully online from students and teachers and informed consent is collected both on paper and online in the beginning of the survey. Foreign-speaking parents receive both online and paper survey access together with the informed consent form both on paper and online. Online survey is managed using a secure online survey tool(48) and all identifiable information will be stored separately from the research data. Researcher team will collect informed consents first in schools and later via mail. In order to avoid attrition participants will receive a personal invitation via e-mail and two reminders to answer the post-intervention (T2) and follow-up (T3) surveys. RWS-consortium will manage and store the research data and will have a common data management plan available online.
Measures
Table 2 presents the measurements taken at each three assessment points before, during and after the interventions. Model form of the informed consent and detailed description of questionnaires is available from the first author upon request.
Table 2 Study Constructs, Measures and Questionnaires of Students, Parents and Teachers
|
Construct
|
Measure
|
Assessment tool
|
STUDENTS
|
Youth mental health
|
PTSD symptoms
|
Children's Revised Impact of Events Scale (CRIES-8)
|
|
|
Internalizing and externalizing problems and prosocial behaviour
|
Strengths and Difficulties Questionnaire (SDQ)
|
|
|
Positive development and resilience
|
Child and Youth Resilience measure (CYRM-12)
|
|
Associated factors of Youth Mental Wellbeing
|
Academic achievement / Executive functions
|
The Amsterdam Executive Function Inventory (AEFI)
|
|
|
Wellbeing
|
Item developed for this study
|
|
|
Experience of the number of stressors in daily life
|
Daily Stressors questionnaire (DSSYR; Vervliet, Derluyn, & Broekaert, unpublished)
|
|
|
Social Support
|
Multidimensional Scale of Perceived Social Support (MSPSS)
|
|
|
Existence of interethnic friendships and Satisfaction on friendships
|
Items developed for this study
|
|
|
Discrimination
|
The Perceived Ethnic Discrimination Questionnaire Community Version (PEDQ-CV)
|
|
|
School Belonging
|
The psychological sense of school membership among adolescents (PSSM)
|
|
Sociodemographic factors
|
gender, age, country of birth, migration status, time in host country and family composition
|
PARENTS
|
Parental health
|
Self-reported health
|
One item of the SF-36
|
|
|
Mental health
|
General health questionnaire (GHQ-12)
|
|
|
PTSD symptoms
|
PTSD-8 questionnaire
|
|
Home-school relations
|
Teacher-Parent collaboration
|
Trust Scale
|
|
Parent experiences of inclusion
|
Experience of discrimination
|
Brief Perceived Ethnic Discrimination Questionnaire – Community version (PEDQ-CV)
|
|
|
Social support
|
Enriched Social Support Instrument (ESSI)
|
|
Family relations
|
Parenting
|
Me as a Parent (MaaP)
|
|
|
Parent-rated strengths and difficulties of the child
|
Strengths and Difficulties Questionnaire - parent rated (SDQ-25 with impact supplement)
|
|
Sociodemographic factors
|
sex, age, marital status, number of children in a household, employment situation, education, income, time in host country, migration status, reason for migration, separation of family members during migration
|
TEACHERS
|
Cultural competence
|
Multicultural awareness and understanding
|
Teacher Multicultural Attitude Scale (TMAS)
|
|
Self-efficacy
|
Teachers' self-efficacy
|
Teachers’ Sense of Efficacy Scale (TSES)
|
|
Stress and work engagement
|
Stress symptoms
|
Single item stress index (SISI)
|
|
|
Work exhaustion/burnout
|
Bergen burnout inventory (BBI)
|
|
|
Work engagement (vigor, dedication, absorption)
|
Utrecht work engagement scale (UWES)
|
|
Home-school interrelations
|
Teacher-parent collaboration
|
Trust Scale
|
|
Intervention
|
Classroom atmosphere
|
Items from a school intervention study
|
|
Sociodemographic factors
|
sex, age, teaching background, number of students
|
|
Primary measures
Psychological distress, i.e., internalizing and externalizing problems are measured with the self-report version of the Strengths and Difficulties Questionnaire (SDQ)(49) for 11–17-year-olds. SDQ is a screening questionnaire that measures 25 attributes divided into either two sub-scales of internalizing and externalizing problems, or five sub-scales: emotional symptoms, conduct problems, hyperactivity-inattention, peer problems, and prosocial behavior. All questions are asked on a scale ‘not true’, ‘somewhat true’, to ‘certainly true’. ‘Somewhat true’ is always scored as 1, but the scoring of ‘not true’ and ‘certainly true’ varies with the item. For each of the five scales the score can range from 0 to 10 if all items are completed. Goodman(50) reports SDQ have an acceptable validity and reliability (lowest in peer problems α: .41 and highest in total difficulties α: .80). Although, research shows good validity for internalizing scales indicated by correspondence between children, parents and teacher reports, and clinical review(51).
Resilience is measured with an adapted, focus group-based Child and Youth Resilience Measure (CYRM-12)(52). CYRM was originally designed as a 28-items measure for youth aged 9 to 23 years old. A reduced resilience measure (CYRM-12) consists of 12 items, which are scaled as 5-point (1 ‘not at all’ to 5 ‘a lot’). A total score is created by summing up the score of each item. CYRM is a questionnaire exploring the individual, relational, communal and cultural resources that may bolster the resilience of 9-23-year-old youth. The participant reports on a 5-point scale as to what extent he/she feels he/she has certain resources. Preliminary results show sufficient content validity of the CYRM-12 to merit its use a screener for resilience processes in the lives of adolescents, and reliability was α: .84.
Executive functions are measured with The Amsterdam Executive Function Inventory (AEFI)(53) subscale called “planning and initiative”. AEFI subscale has five items that are scored on a three-point scale (1 ‘not true’, 2 ‘somewhat true’, 3 ‘true’). Authors report AEFI has adequate construct validity, but moderate reliability (α: .60).
Overall wellbeing is measured with one question developed for this study. It asks person’s overall wellbeing on a five-point scale (1 ‘very bad’, 2 ‘bad’, 3 ‘not good or bad’, 4 ‘good’, 5 ‘very good’). So far there is no evidence on the reliability or validity of this measure.
Post-traumatic stress symptoms (PTSD) are screened with Children’s Impact of Events Scale (CRIES-8)(54-56). CRIES-8 has two sub-scales, named intrusion and avoidance, and it is designed to be used in children aged 8 and above. The eight items are scored on a four-point scale (0 ‘Not at all’, 1 ‘Rarely’, 3 ‘Sometimes’, 5 ‘Often’). Total score is the sum of scores from the two sub-scales. The screening cut-off is on 17 ≤ points. CRIES-8 has been applied and its factors proved as robust in a variety of cultures. It has good construct validity and quite stable factor structure, it correlates well with other indices of distress, and it has been used to screen very large samples of at-risk-children following a wide range of traumatic events(57). Another study(58) reported CRIES-8 to have good internal consistency (α: .86).
Other measures
School belonging is measured with the Psychological sense of school membership among adolescents (PSSM)(59). Social support is measured with Multidimensional scale of perceived social support (MSPSS)(60). Interethnic friendships are measured as the number of interethnic friendships and satisfaction with friendships, measures were developed for this study. Daily Stressors are measured with the Daily Stressors questionnaire (DSSYR; Verlievt, Derluyn, & Broekaert, unpublished). Discrimination is measured with the Perceived Ethnic Discrimination Questionnaire Community Version (PEDQ-CV)(61).
Teachers’ multicultural skills are measured with the Teacher Multicultural Attitude Scale (TMAS)(62). Teachers' self-efficacy are measured with the Teachers’ Sense of Efficacy Scale (TSES)(63). Teachers’ work engagement and stress are measured with three measures: Utrecht work engagement scale (UWES)(64), single-item stress index (SISI)(65), and Bergen burnout inventory (BBI)(66).
Parents’ general and mental health are measured with one item of the SF-36(67), General health questionnaire (GHQ-12)(68), and PTSD-8 questionnaire(69). Parenting is measured with the Me as a Parent (MaaP)(70), parents’ social support is measured with the Enriched Social Support Instrument (ESSI)(71), and parents’ perceived discrimination is measured with the same as with adolescents(61).
Background variables
Adolescents report gender, age, country of birth, and migrant students also migration status, time in host country and family composition. Teachers report sex, age, teaching background, and number of students. Parents report sex, age, marital status, number of children in a household, employment situation, education, income, and immigrant parents also time in host country, migration status, reason for migration, and separation of family members during migration.
Analytic Strategy
Due to the sampling procedure, our data will be clustered in two levels; pupils are nested in 134 classes, and classes in 16 schools. The data contains variables from individual, class and school levels, and the interventions are given at school level. All the indicators of outcome variables are measured at three time-points in individual level, T1 immediately before the intervention, T2 20 weeks, and T3 as a follow-up 52 weeks after the intervention. We expect the intervention to change the levels in outcomes from T1 to T2, and the new levels to remain the same through T3 follow-up.
The structure of the data requires the use of structural equation (SEM) methods combining longitudinal, multilevel, path analytic models with confirmatory factor analysis. The psychometric properties of the outcome variables, including measurement invariance between timepoints T1, T2, and T3, are tested before constructing summary scales using most reliable items. The scales are then specified as manifest variables in the models. Alternatively, latent constructs are used, depending on the complexity of the final model specifications. The core model in all our analyses will be a version of latent growth curve model(72, 73), that is also known as intervention model(74) due to the coding of the fixed parameters in the model, which are responsible for the shape of the constant growth trajectories at the individual level (the 1’s in Figure 2). The latent growth parameters will be estimated with Mplus 8.0 software using robust maximum likelihood estimator (MLR), which allows non-normal continuous indicators to be reliably analyzed. Additionally, MLR is a full-information estimator allowing the use of missing data without any separate imputations. The validity of MAR assumption is scrutinized and, if needed, auxiliary variables predicting missingness will be added into the models. Demographic and baseline characteristics will be summarized using means and standard deviations for quantitative variables and percentages for categorical variables.
Explanations: Primary outcomes: Y1= SDQ (difficulties); Y2= SDQ (strengths); Y3 = CYRM; Y4 = AEFI (Executive Function); Y5 = Overall wellbeing; Intervention status: I0= waiting-list control group; I1= INSETT+TRT; I2= PIER; Possible moderators: Z1= SF-36; Z2= GHQ-12; Z3= PTSD-8; Z4= MaaP; Z5= PEDQ-CV (parent); Z6= PEDQ-CV (adolescent); Z7= DSSYR; Z8 = (Daily Stressors); Z8…n= sociodemographics; Possible mediators: INSETT Teachers (competence); Mb1= TMAS; Mb2= TSES; Mb3= SISI; Mb4= BBI; Mb5= UWES; Mb6= Trust; Mb7= Classroom atmosphere; TRT Adolescents (trauma) Ma1= CRIES-8; PIER Adolescents (social) Mc1= MSPSS; Mc2= PSSM; Mc3= PEDQ; Mc4= Interethnic friendships.
In the core model (Figure 2), there are two latent state variables, measuring true scores of PRE- and POST-intervention levels of the outcomes. The PRE-intervention state is indicated by the measured T1 variable without time-related measurement error because there is only one measurement point. In the case of POST-intervention state, there are two indicators, enabling the modeling of measurement error. The loadings of the indicators in the measurement part of the model are fixed to 1’s, meaning that the trajectory from T1 to T3 has the same shape for every individual. Only the level of PRE-intervention state and the difference between PRE- and POST-intervention states are random, so that the means and variances of these variables within and between the clusters can be estimated. As predictors of the POST-intervention state, our model specifies the PRE-intervention state (path coefficient a, which is anticipated to be negative due to floor and ceiling effects), a dummy for intervention/control group membership (direct effect c), and potential mediating variables (path coefficient e, implying indirect effect d*e for intervention-control-group difference). The covariance between PRE-intervention state and the dummy for intervention/control group membership (parameter b) is supposed to be zero, indicating that the randomization was successful. Moderation mechanisms are investigated by estimating the model simultaneously in groups created according to categorized moderator variables (multi-group SEM). Finally, the simple estimates of total intervention effects are calculated from a model without mediators and moderators (parameter c’ divided by the variance of POST-intervention latent variable, giving an equivalent to Cohen’s d).