Introduction
Mental health problems in children are often divided into internalizing and externalizing disorders. Internalizing symptoms include anxiety, sadness, low self-esteem, and somatic discomfort, while externalizing symptoms are related to behavioral problems such as disobedience, impulsiveness, overactivity, or aggressive behavior. In children, comorbidity between disorders within each category and between internalizing and externalizing problems is common. This comorbidity leads to greater distress and disability, more severe symptoms, an increased risk of physical problems and poorer treatment outcomes (Dol et al., Reference Dol, Reed and Ferro2022; Melton et al., Reference Melton, Croarkin, Strawn and Mcclintock2016).
Cognitive behavioral interventions are the recommended treatment for internalizing and externalizing symptoms in childhood. However, its effectiveness is limited as it is effective in 50% of children with anxiety (Lebowitz et al., Reference Lebowitz, Marin, Martino, Shimshoni and Silverman2020) and is associated with moderate to low effects on depressive symptoms (Oud et al., Reference Oud, de Winter, Vermeulen-Smit, Bodden, Nauta, Stone, van den Heuvel, Taher, de Graaf, Kendall, Engels and Stikkelbroek2019; Sun et al., Reference Sun, Rith-Najarian, Williamson and Chorpita2019). Therefore, despite treatment, almost half of the children seeking help would still have significant emotional symptoms and life limitations. Several explanations may be related to these results. First, comorbidity between anxiety and behavioral problems is common in children, but there are not enough treatments that adequately address anxiety symptoms and behavioral problems at the same time (Byrne et al., Reference Byrne, Connon, Martin, McHugh and Power2021). Next, disorder-specific interventions have positive effects on the disorder for which they are designed, but effects on other comorbid symptoms are limited or absent (Garber et al., Reference Garber, Brunwasser, Zerr, Schwartz, Sova and Weersing2016). Furthermore, CBT efficacy studies are typically conducted in research settings, which can compromise the generalization of the results (Orgilés et al., Reference Orgilés, Garrigós, Espada and Morales2020). In clinical settings, mental health professionals have to grapple with multiple dilemmas related to these questions, deciding whether to use the gold standard treatment for each disorder at the same time, one specific protocol and the other later, or select components from both protocols but not the whole validated program. All these questions can compromise the effectiveness of the treatment and the limited results.
These limitations, together with data from investigative studies reporting shared features and mechanisms in emotional disorders and benefits in comorbidity symptoms when treating a main disorder (Priede et al., Reference Priede, Cano-Vindel, González-Blanch, Ruíz, Moriana and Fonseca2021); led to the development of transdiagnostic protocols. Transdiagnostic CBT interventions for example are now recognized as an effective treatment for comorbid symptoms of anxiety and depression in children (García-Escalera et al., Reference García-Escalera, Chorot, Valiente, Reales and Sandín2016; Sandín et al., Reference Sandín, Valiente, García-Escalera, Pineda, Espinosa, Magaz and Chorot2019).
A proposed transdiagnostic construct is self-esteem (García-Escalera et al., Reference García-Escalera, Chorot, Valiente, Reales and Sandín2016). Self-concept and self-esteem are related to psychosocial well-being (Daemen et al., Reference Daemen, Postma, Lindauer, der Meulen, Nieman, Delespaul, Breedvelt, van der Gaag, Viechtbauer, Schruers, van den Berg, Bockting, van Amelsvoort and Reininghaus2021). High self-esteem appears to play a protective role in the development of mental health problems (Martinsen et al., Reference Martinsen, Rasmussen, Wentzel-Larsen, Holen, Sund, Pedersen, Løvaas, Patras, Adolfsen and Neumer2021). On the contrary, low self-esteem is associated with the development and perpetuation of mental health problems, with approximately 45% of youth seeking help reporting low self-esteem (Daemen et al., Reference Daemen, Postma, Lindauer, der Meulen, Nieman, Delespaul, Breedvelt, van der Gaag, Viechtbauer, Schruers, van den Berg, Bockting, van Amelsvoort and Reininghaus2021). The role of self-esteem in the ethology of depressive symptoms and their severity has been widely reported (Stadelmann et al., Reference Stadelmann, Grunewald, Gibbels, Jaeger, Matuschek, Weis, Klein, Hiemisch, von Klitzing and Döhnert2017), but is also related to other internalizing disorders such as anxiety disorders (Martinsen et al., Reference Martinsen, Rasmussen, Wentzel-Larsen, Holen, Sund, Pedersen, Løvaas, Patras, Adolfsen and Neumer2021; Stadelmann et al., Reference Stadelmann, Grunewald, Gibbels, Jaeger, Matuschek, Weis, Klein, Hiemisch, von Klitzing and Döhnert2017). The literature indicates that children with mental health problems have lower self-esteem than children without these problems (Martinsen et al., Reference Martinsen, Rasmussen, Wentzel-Larsen, Holen, Sund, Pedersen, Løvaas, Patras, Adolfsen and Neumer2021).
Children are frequently not the ones seeking help, while parents and teachers are often the ones who see them as needing help. These circumstances can result in less motivation for the treatment and a handicap for the professional to overwhelm. For this reason, group treatments can be a helpful resource for clinicians, as they reduce the stigma of treatment and increase motivation. Additionally, group interventions have unique components: peer modeling, peer reinforcement, and group feedback to identify and offer alternatives to thoughts and behaviors (Silverman et al., Reference Silverman, Marin, Rey, Kurtines, Jaccard and Pettit2019; Wolgensinger, Reference Wolgensinger2015). This is particularly important for treatment effectiveness, as children have more credibility and confidence in peer support and advice (Santesteban-Echarri et al., Reference Santesteban-Echarri, Hernández-Arroyo, Rice, Güerre-Lobera, Serrano-Villar, Espín-Jaime and Jiménez-Arriero2018). On top of that, some authors suggest that group interventions in community mental health settings may increase self-esteem (Rippon et al., Reference Rippon, Shepherd, Wakefield, Lee and Pollet2022). In addition, group interventions allow for the management of relationship difficulties and social anxiety as they enable to practice social skills and exposure in a more similar real-world context.
There are several available group transdiagnostic CBT programs for children: EMOTION (Martinsen et al., Reference Martinsen, Rasmussen, Wentzel-Larsen, Holen, Sund, Løvaas, Patras, Kendall, Waaktaar and Neumer2019), Unified Protocol for children (UP-C) (Ehrenreich-May et al., Reference Ehrenreich-May, Kennedy, Sherman, Bilek, Buzzella, Bennett and Barlow2008), and Super Skills for Life (SSL) (Essau & Ollendick, Reference Essau and Ollendick2013). However, only two of them have been translated into Spanish, UP-C and SSL. Furthermore, while there is empirical support for the original English UP-C version (Kennedy et al., Reference Kennedy, Bilek and Ehrenreich-May2018), there are no clinical studies examining the effectiveness of the Spanish UP-C protocol.
SSL is a transdiagnostic intervention that addresses transdiagnostic shared constructs such as low self-concept/esteem and poor social skills. The program includes psychoeducation, behavioral activation, social skills, relaxation strategies, problem-solving techniques and videofeedback with cognitive preparation. The SSL program was originally developed for use in the school context (Essau & Ollendick, Reference Essau and Ollendick2013) but has also been applied in other contexts with positive results (i.e., Ramdhonee-Dowlot et al., Reference Ramdhonee-Dowlot, Balloo and Essau2021). The Spanish SSL program has empirical support as an indicated prevention program for children aged 6–8 years (Fernández-Martínez et al., Reference Fernández-Martínez, Morales, Espada, Essau and Orgilés2019), 8–12 years in a group format (Orgilés et al., Reference Orgilés, Fernández-Martínez, Espada and Morales2019), and 8–12 years in an individual format (Melero et al., Reference Melero, Orgilés, Espada and Morales2021). The effectiveness of the program as a prevention program has been demonstrated in the short and long term (Fernández-Martínez et al., Reference Fernández-Martínez, Morales, Espada, Essau and Orgilés2019, Reference Fernández-Martínez, Orgilés, Morales, Espada and Essau2020; Orgilés et al., Reference Orgilés, Fernández-Martínez, Espada and Morales2019). Based on these results, benefits can also be expected in clinical samples. However, up to date no studies have examined the effectiveness of SSL in children with a diagnosed mental disorder.
This study is part of the SSL randomized controlled trial in a clinical setting (ClinicalTrials.gov NCT05482724), with the aim to evaluate the effectiveness of the program in Spanish children with a clinical diagnosis (based on international manual classifications) in real clinical contexts. A previous related study presented results of SSL in anxiety and depression (Diego et al., Reference Diego, Morales and Orgilés2023). Now we present the efficacy of SSL as a treatment program for common comorbid symptoms. This research is particularly relevant as it demonstrates for the first time the benefits of transdiagnostic SSL intervention for comorbid symptoms in Spanish children with a clinical diagnosis. The limitations and gaps in the presented literature highlighted the need for research that efficiently address comorbid symptoms in children. If not properly managed, they can be associated with poorer current and future well-being (Daemen et al., Reference Daemen, Postma, Lindauer, der Meulen, Nieman, Delespaul, Breedvelt, van der Gaag, Viechtbauer, Schruers, van den Berg, Bockting, van Amelsvoort and Reininghaus2021), limited treatment outcomes (Dol et al., Reference Dol, Reed and Ferro2022), and lack of treatment effects on comorbid symptoms (Garber et al., Reference Garber, Brunwasser, Zerr, Schwartz, Sova and Weersing2016). There is also a need for results from clinical settings that demonstrate benefits to this population as opposed to community sampling or research settings.
A quasi-experimental design with two groups (control and experimental), pretest and posttest, was performed to evaluate the effects of the Spanish version of transdiagnostic SSL in a clinical sample of Spanish children aged 8–12 years with a main emotional disorder. The main objective was to examine changes in internalizing and externalizing symptoms before and after the intervention. Based on previous promising results, we hypothesized: a) reductions in emotional symptoms, behavioral problems, peer difficulties, and hyperactivity symptoms, and b) positive improvements in self-concept.
Methods
Procedure
The data in the present study are part of the SSL randomized controlled study in clinical settings in Valencia, Spain. The study protocols, procedures, and informed consent forms were approved by the Ethics Committee of the Biomedical Research Institute INCLIVA following the Declaration of Helsinki. Clinical trial is registered with ClinicalTrials.gov (NCT05482724). Child and adolescent psychiatric leaders at the Hospital Clínico Universitario in Valencia informed children and their parents who were already attending the Child and Adolescent Mental Health Unit about the program. Interested families were contacted by a study investigator to provide additional information, confidentiality and voluntary consent to participate, as well as full pretest assessment. Sample size was determined in terms of beta strength and clinical need. A total of 100 Spanish-speaking children and their parents were recruited at baseline assessment. Participants with a clinical diagnosis who met the inclusion criteria (N = 88) and agreed to participate (N = 86) were randomly assigned (1:1) to the intervention group (N = 43) and the wait-list control (WLC) group (N = 43). Inclusion and exclusion criteria of the participants can be found in Table 1. The comorbidity in the sample was high: 87.84% (N = 65) of the children met the criteria for two disorders and 45.95% (N = 34) for three or more. The list of all diagnoses is presented in Table 2. Simple randomization was performed by an independent investigator via an Excel file. Neither parents nor children received any compensation for their participation in this study. For a discussion of the participants' CONSORT flowchart, see Diego et al. (Reference Diego, Morales and Orgilés2023). Posttreatment assessments were individual, by appointment, paper and pen, and were conducted in one session simultaneously at the mental health department facilities for the intervention group and the WLC group. These sessions were conducted by an advanced clinical psychology doctoral student who was unrelated to the study and had no knowledge of the allocation of participants.
Table 1. Inclusion and exclusion criteria of the participants
Table 2. Frequency of the principal and comorbid diagnosis in the pretreatment
SSL = super skills for life; WLC = wait-list group.
Wait-list group
Participants in the wait-list group received no psychological (public or private) intervention during the 8-week duration of the SSL program. They were informed that children in this group will receive the intervention once the follow-up visit is completed. The frequencies of all diagnoses in the wait-list group and the intervention group are shown in Table 2.
Intervention group
Participants assigned to the intervention group participated in the Spanish version of the SSL program. For the therapist manual and children’s workbook corresponding to the Spanish version of the program, see Orgilés et al. (Reference Orgilés, Espada, Ollendick and Essau2022). Participant diagnoses are presented in Table 2. The intervention was administered by a clinical psychologist at 3–4 years of residency who received an intensive 1-day training on SSL. Sessions were held after school hours once a week for 8 weeks, with each session lasting one hour and conducted at the Mental Health Service facilities. The contents of the program were conveyed through playful activities in groups of 4–6 children. The program includes a therapist manual with structured steps for conducting the sessions, an 8-session workbook for children, exercises to be completed between sessions, and post-session information for parents (i.e., goals, exercises practiced, guidelines, and assigned homework for the next session). The program teaches children skills and techniques that are highlighted in the literature and play a relevant role in the development and treatment of emotional disorders: social skills, behavioral activation, cognitive restructuring, self-monitoring, relaxation techniques, and problem-solving techniques. The content of the 8 sessions of the SSL program is presented in Table 3. Sessions are structured with adapted explanations and examples, lectures, role plays, group and individual activities and games, video analysis with cognitive preparation and feedback, and easy homework to practice skills learned between sessions. All information about the program, materials, and training can be found on https://superskillsforlife.com and https://superskills.umh.es/.
Table 3. Description of the SSL components in each session
Measures
Sociodemographic variables for parents and children
Sociodemographic variables were collected using an ad hoc questionnaire and prior information. All analyzed variables are listed in Table 4.
Table 4. Sociodemographic characteristics of baseline participating children by intervention condition
Difficulties and strengths
Emotional and behavioral difficulties and positive attributes (Children-report version SDQ and parent-report version SDQ-P, Goodman, Reference Goodman1997). The SDQ measures children’s overall difficulties and positive attributes across five subscales: Emotional symptoms, Conduct problems, Hyperactivity/Inattention, Peer relationships (difficulties) and Pro-social behavior. The psychometric properties of the SDQ-P were confirmed, showing reasonable internal consistency and satisfactory validity (Goodman, Reference Goodman2001). In the current study total reliability was high for the parent-report version (ordinal alpha = .80) and moderate for the child report version (ordinal alpha = .67). Reliability for the subscales was also high in the parent version: Emotional symptoms (ordinal alpha = .73), Conduct problems (ordinal alpha = .83), Hyperactivity (ordinal alpha = .82), Peer relationship problems (ordinal alpha = .77) and Prosocial behavior (ordinal alpha = .77). The ordinal alpha for the child version was moderate for Emotional symptoms (ordinal alpha = .56), Conduct problems (ordinal alpha = .56), Hyperactivity (ordinal alpha = .66), and Prosocial behavior (ordinal alpha = .69); while for Peer relationship problems (ordinal alpha = .71) was high.
Self-concept
Self-Concept Form 5 (AF-5, García & Musitu, Reference García and Musitu2023). The AF-5 measures five dimensions of self-concept: Social; Academic/Professional; Emotional; Family; and Physical. The literature supports the reliability, internal consistency, structural validity, and convergent validity of the AF-5 (García et al., Reference García, Gracia and Zeleznova2013). In the current study, children rated the questionnaire using a 5-point Likert scale, as suggested by Galindo-Domínguez (Reference Galindo-Domínguez2019). Cronbach’s alpha was 0.79 for the AF-5 total score and ranged from 0.62 to 0.80 for the subscales: Academic (Cronbach’s alpha = .80), Social (Cronbach’s alpha = .68), Emotional (Cronbach’s alpha = .54), Familiar (Cronbach’s alpha = .62) and Physical (Cronbach’s alpha = .75).
Statistical analysis
SPSS v.27 software was used to performed statistical analysis. A p-value less than .05 is considered statistical significance. Groups had<50 cases, and variables were not normally distributed (p > .05); therefore, the Mann Whitney U test was performed to determine the equivalence of the two conditions (SSL and WLC). Cross-tabulation (2) was used for categorical variables. The effectiveness of the posttest intervention was assessed using generalized estimating equations (GEE) since they are considered a robust method to check the effectiveness of interventions in randomized controlled trials, even with a reduced sample size (Liang & Zeger, Reference Liang and Zeger1986). Analyses were adjusted for sex, age, and baseline measurements. All variables were assessed separately using participants as the unit of analysis and randomization. Estimates indicate the estimated difference units in the outcome between the SSL group and the WLC in the posttest, compared to the pretest. For example, a value of 3.01 indicates an estimated difference of 3.01 units in the outcome between the treatment group and the WLC in the posttest compared to pretest. This means that, on average, the treatment group is expected to have an outcome value 3.01 units lower than the control group. The associated 95% confidence interval (CI 95% = 0.90, 5.12) provides an estimation of the coefficient’s precision. Cohen’s (Reference Cohen1988) effect size (d) was also calculated to provide an additional effect size indicator. It can be interpreted with Cohen’s (Reference Cohen1988) convention: small effect: d ≈ 0.2, medium effect: d ≈ 0.5, and large effect: d ≈ 0.8. Finally, the reliability for the current sample was calculated according to Cronbach’s internal consistency coefficient and the ordinal alpha for ordinal measures (using R Studio).
Results
Baseline
At baseline, the intervention group and the WLC group did not differ significantly in terms of sociodemographic children’s characteristics (p > .05): age, pharmacotherapy, number of siblings; parents: age, family situation, educational level, nationality, work status, and mental health history. The sociodemographic characteristics of the participants assessed at pretest by group are shown in Table 4.
In addition, the two groups did not differ in any of the study variables at the start of the study. Table 5 shows mean values and standard deviations of outcome measures reported by group condition. Mann–Whitney U test results show that scores were equal (p > .05) for the intervention group and the WLC group for difficulties and strengths: SDQ self-report and parent-report, and self-concept: AF-5 self-report.
Table 5. Preintervention and postintervention means (SD) of the outcomes
Effects of intervention
Table 5 shows the SSL and WLC means and standard deviations before and after the intervention. Significant reductions from pretest to posttest were observed on several variables in children participating in SSL compared to those in the WLC (Table 6).
Table 6. Generalized linear model-based estimates 95% confidence intervals (CI), and significance tests for intervention effect on the outcomes
Global difficulties
At baseline, children from both conditions scored 17-20 on the SDQ total difficulties subscale, which is consistent with Goodman’s proposed cutoff for the abnormal/borderline band (http://www.sdqinfo.org). Children who participated in SSL were placed in the normal range after intervention, while children in WLC were still placed in the abnormal/borderline group at posttest.
Statistically significant decreases from the baseline to follow-up on total difficulties were reported by children (p = .005, Estimates = 3.01, 95% CI: .90, 5.12) and parents (p = .006, Estimates = 3.40, 95% CI: .95, 5.84). The size of the Cohen’s effect for this variable was medium when assessed by the children (Cohen’s d = 0.70) and large when assessed by adults (Cohen’s d = 0.80).
Emotional symptoms
Children of both groups were initially classified by their parents in the abnormal band for emotional problems (mean value close to 6). After the test, the children in the SSL group descend to a lower level of categorization, while the WLC group maintained the emotional problem level.
A significant reduction in emotional symptoms from baseline to follow-up was observed in children from the intervention group when these symptoms were reported by the children (p < .001, Estimates = 1.72, 95% CI: .83, 2.61) and by their parents (p = .005, Estimates = 1.35, 95% CI: .41, 2.28). The effect size of these differences was medium when assessed by parents (Cohen’s d = 0.58) and large when assessed by children (Cohen’s d = 0.82).
Conduct problems
Considering the three bands proposed by Goodman, children belonging to both conditions were themselves classified in the normal band for conduct problems at pre- and posttest (mean value close to 3), while parents’ reports placed them in the borderline band before and after the intervention (SSL pre- and posttest, WLC pre- and posttest) (mean value close to 3).
Differences on the SDQ and SDQ-P behavioral problems subscales did not reach the level of statistical significance (p-values > 0.05).
Hyperactivity and inattention symptoms
No changes in hyperactivity and inattention symptoms reached the alpha threshold when reported by the children or their parents (p-values > 0.05).
However, reductions in the hyperactivity/inattention subscale were associated with changes in the Goodman classification. Children from WLC were classified at the abnormal band (SDQ-P) or borderline band (SDQ) at pre- and posttest (mean values close to 6–7). In contrast, children who participated in SSL were classified in the borderline band at pretest (SDQ and SDQ-P) and close to normal after the intervention (SDQ and SDQ-P mean values close to 5).
Peer relationship problems
A significant reduction in parent-measured peer problems was observed in children who participated in SSL (p = .002, Estimates = 1.55, 95% CI: .58, 2.51). Cohen’s effect size of these differences was medium (Cohen’s d = 0.66).
Benefits were also noted in Goodman classification. Children in the SSL and the WLC groups could initially be placed in the borderline group (mean value approximately 3). The WLC group remained in this category at posttest, while the children in SSL reached the normal range (mean value approximately 2).
Children reported a smaller effect on this variable and did not result in statistically significant changes (p = .184, Estimates = 0.63, 95% CI: −.30, 1.55).
Prosocial behaviors
No significant benefits were found for prosocial behaviors as measured by the SDQ. SDQ and SDQ-P scores at pre- and posttest indicated that children in SSL and WLC groups could be placed within the normal range for prosocial behaviors both before and after the intervention (values 6–10).
No statistically significant benefits for children’s prosocial behaviors were reported by children or parents (p-values > 0.05).
Self-concept
Children in the SSL condition showed significantly greater improvements in global self-concept compared with children in the WLC condition (AF-5 total score p = .002, Estimates = −6.93, 95% CI: −11.41, −2.45). The effect size of these differences was medium (Cohen’s d = 0.48).
On the self-concept subscales, there were marginally significant improvements for academic self-concept (p = .09, Estimates = −1.33, 95% CI: −2.9, 0.22), social self-concept (p = .06, Estimates = −1.64, 95% CI: −3.38, 0.09), and emotional self-concept (p = .05, Estimates = −1.91, 95% CI: −3.84, 0.013).
Gender-based assessment
The GEE analysis was performed to compare postintervention changes between boys and girls (Table 6). After participating in the SSL intervention, boys and girls improved equally in the internalizing and externalizing symptoms as measured by the SDQ/SDP-Q. However, significant differences were found in some self-concept subscales: academic self-concept (p = .05) and social self-concept (p = .04) (Table 6). Girls presented higher academic and social self-concept compared to boys.
Discussion
Overall, the results of the present study suggest that the SSL intervention, used for the first time in a clinical sample of Spanish children with a mental disorder, had a positive short-term impact on internalizing symptoms, externalizing symptoms, and self-concept.
Regarding the first objective of the research, statistically significant reductions in emotional symptoms and total difficulties reported by children and parents, and parent-reported peer difficulties were found. As a novelty in SSL research, in this study we tested the intervention in a clinical sample with a DSM-5 diagnosis and jointly present clinical data from parents and children for the same variables.
The fact that parents, not children, are the best informants about externalizing symptoms (Grills & Ollendick, Reference Grills and Ollendick2003), and the strong component of emotional regulation strategies (psychoeducation, emotion recognition, regulation strategies, and control of subjective well-being) and social skills training (basic skills, role playing, and problem-solving strategies applied to social conflicts) may explain the benefits of the SSL program for emotional symptoms, overall difficulties, and peer relationships in Spanish clinical children.
Both children and parents reported a significant reduction in children’s suffering and disability in a short period of time (8 weeks), making SSL a promising treatment to avoid difficulties and limitations often associated with comorbid conditions, especially in the context of public health (clinical treatment dilemmas, worse treatment outcomes, higher burden). Short-term effects on depressive symptoms, anxiety symptoms, and remission from diagnosis were further examined in a previous related publication (Diego et al., Reference Diego, Morales and Orgilés2023).
In contrast to previous SSL effectiveness studies, statistically significance short-term benefits (rather than long-term benefits) were found in peer difficulties (Essau et al., Reference Essau, Olaya, Sasagawa, Pithia, Bray and Ollendick2014; Fernández-Martínez et al., Reference Fernández-Martínez, Morales, Espada, Essau and Orgilés2019), underlining the importance of our results. Our sample was composed by children with a clinical diagnosis, as contrast to subclinical or community samples from previous research, suggesting that children attending mental health services (with higher levels of stigma frequently associated with low self-concept) may be particularly benefit from the unique advantages of group interventions such as SSL.
Like previous SSL studies, we found no significant short-term reductions in hyperactivity/inattention symptoms and conduct problems. However, statistically significant improvements in these variables are obtained at 6-month/1-year follow-up studies (Fernández-Martínez et al., Reference Fernández-Martínez, Orgilés, Morales, Espada and Essau2020; Orgilés et al., Reference Orgilés, Fernández-Martínez, Espada and Morales2019). These findings suggest the importance of long-term assessments as a better estimate of SSL effectiveness. Long-term positive effects are also expected to be identified in our sample. Nevertheless, these results should be evaluated in further studies as the benefits observed in previous studies with nonclinical samples could not be extrapolated to children with a clinical diagnosis.
It is also important to emphasize that 31.6% of the children in the intervention group (and 36% in the WLC group) had an additional ADHD diagnosis (pharmacotherapy remained unchanged throughout the study). This fact might help explain why changes in the Goodman classification were reported, but the results did not reach statistical significance.
As for conduct problems, a possible floor effect in SDQ measures and SSL not being designed to address behavioral problems may be behind these results. Furthermore, CBT gold standard programs for hyperactivity and behavior problems in children usually incorporate parent sessions. Therefore, adding family sessions to SSL may boost these outcomes and should be tested in future studies.
A second objective was to test the effectiveness of SSL on self-concept. SSL showed a significant improvement in global self-concept compared to the WLC group. Increases in academic, social, and emotional self-concept were also found. Girls showed higher academic (p = 0.05) and social (p = 0.04) self-concept than boys after the test, suggesting that the program may be more effective on these variables for girls. However, long-term data and further research should confirm these results.
Positive effects on emotional symptoms and peer interactions, as well as emotional and social skills learned during the program, may be related with greater children’s self-efficacy and satisfaction. Likewise, changes in hyperactivity/inattention, along with reductions in impairment associated with decreases in internalizing and externalizing symptoms, may be associated with improvements in academic self-concept.
Short-term impact could be enhanced at long-term follow-up with more opportunities to practice the skills learned and a greater quantity of positive experiences. These have been examined in previous work (Orgilés et al., Reference Orgilés, Garrigós, Espada and Morales2020), and it was found that academic, social, emotional, and physical self-concept increased significantly at the 12-month follow-up.
Positive benefits on all dimensions of self-concept may have important implications for internalizing and externalizing symptoms. Recent work in this area (Melero et al., Reference Melero, Orgilés, Espada and Morales2021) has shown that increases in self-concept dimensions mediate the long-term effect of SSL on internalizing and externalizing symptoms. These findings are consistent with previous research that linked self-esteem to mental well-being and viewed the dimension as a transdiagnostic construct that may lie behind a different set of mental health symptoms. Again, these results (Melero et al., Reference Melero, Orgilés, Espada and Morales2021; Orgilés et al., Reference Orgilés, Garrigós, Espada and Morales2020) were not performed in a clinical sample, and there is no previous literature on the short- and long-term effects of SSL on self-esteem in clinical children. It is also important to note that the children enrolled in the program are of pre-adolescent age and are beginning to experience puberty-related changes, which tend to be associated with a decline in self-esteem (Martinsen et al., Reference Martinsen, Rasmussen, Wentzel-Larsen, Holen, Sund, Pedersen, Løvaas, Patras, Adolfsen and Neumer2021; Tirlea et al., Reference Tirlea, Bonham, Dordevic, Bristow, Day, Brennan, Haines and Murray2019).
Taken together, these results suggest that the 8-week SSL intervention could be an effective intervention for comorbidities in daily clinical practice in the Spanish public psychiatric healthcare system, although this result should be confirmed by future studies. Previous studies examining the effectiveness of transdiagnostic treatment in children with emotional problems typically present results on symptoms of anxiety and depression, while information on possible benefits in other comorbidities is lacking (e.g., Bilek & Ehrenreich-May, Reference Bilek and Ehrenreich-May2012; Djurhuus & Bikic, Reference Djurhuus and Bikic2019; Kennedy et al., Reference Kennedy, Bilek and Ehrenreich-May2018; Loevaas et al., Reference Loevaas, Lydersen, Sund, Neumer, Martinsen, Holen, Patras, Adolfsen, Rasmussen and Reinfjell2020; Martinsen et al., Reference Martinsen, Rasmussen, Wentzel-Larsen, Holen, Sund, Løvaas, Patras, Kendall, Waaktaar and Neumer2019). Therefore, the results of this study on the benefit of transdiagnostic SSL on symptom externalization and self-concept are of particular relevance.
Limitations
Research limitations should be considered when interpreting the results. The sample size was small and precedent from a specific area in Spain. Blinding of participants and therapists was not achievable, and possible intervention effects were reported by the unblinded participants. In addition, the control group received no intervention and was not an active control group. The interpretation of the results should take into account the internal consistency of certain SDQ subscales, as well as the potential presence of a floor effect. It would be interesting to measure symptoms on other scales and in other relevant contexts (e.g., school) as it can be difficult for parents to assess changes in contexts where they are not present. This could help us better understand the impact of the intervention on variables that reach statistical significance and on measures that do not reach the threshold of statistical significance. Future research addressing these limitations will lead to clinical and scientific advances.
Conclusion
Despite the limitations mentioned, the present study has some strengths that need to be highlighted. This is one of the first studies to evaluate the short-term effectiveness of SSL in children ages 8 to 12 with a confirmed diagnosis of a mental disorder, in addition, parent- and child-reported measures are included. It provides preliminary support for the short-term effectiveness of SSL group version on emotional symptoms, peer problems, overall difficulties, and self-concept. The research extends the evidence supporting SSL’s effectiveness by examining it with a control condition. Our results suggest that SSL appears to be a promising program that, with further research, can contribute to the transferability of evidence-based interventions to real-world clinical settings.
Acknowledgments
The authors would like to thank everyone who participated in the study, particularly children and parents, as well as the program facilitators.
Funding statement
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
None.
