The Effectiveness of Modular Transdiagnostic Cognitive Behavioral Therapy Versus Treatment as Usual for Youths Displaying School Attendance Problems: A Randomized Controlled Trial

School attendance problems (SAPs) refers to difficulty attending school or absence from school that is problematic because of its frequency and/or duration (Heyne et al., 2019). SAPs are associated with negative outcomes for youths such as poor academic achievement (Gottfried, 2014), school dropout (Schoeneberger, 2012), and later unemployment (Attwood & Croll, 2015). Studies further show that most youths displaying SAPs experience a range of mental health problems including depression, anxiety, and behavioral problems (Askeland et al., 2015; Egger et al., 2003; Finning et al., 2019; Vaughn et al., 2013). In various countries, missing 10% or more of school is regarded as the threshold for the presence of a SAP, and prevalence rates of youths crossing this threshold range from 11 to 25% in Denmark, Australia, the UK, and the USA (Australian Curriculum Assessment and Reporting Authority, 2023; Danish National Agency for IT and Learning, 2019; Department for Education, 2019; U.S. Department of Education, 2019). Rates of SAPs typically increase with age and are more prevalent among older youths (Gubbels et al., 2019; Heyne, 2022). Because SAPs are prevalent and associated with negative short- and long-term outcomes, there is need for effective, scalable, and accessible interventions.

Previous research on SAPs has often focused on one of two types, namely truancy or school refusal. Truancy involves unauthorized non-attendance or ‘skipping’ school, while school refusal refers to non-attendance associated with emotional distress (Heyne et al., 2019). This distinction between truancy and school refusal has impacted the development of SAP interventions, inasmuch as interventions are often designed to address one of these two types (Heyne et al., 2015). Truancy interventions primarily aim to increase school attendance by involving various stakeholders, including the young person (e.g., through mentoring), parents/families (e.g., via parent training), or the school/community (e.g., by fostering school bonding) (DeSocio et al., 2007; Franklin et al., 2007; Mazerolle et al., 2017). School refusal interventions often focus on improving school attendance using cognitive behavioral therapy (CBT), incorporating graded exposure and cognitive restructuring techniques to reduce mental health problems related to symptoms of anxiety and/or depression (Heyne et al., 2011; Heyne & Sauter, 2013). Similar to truancy interventions, school refusal interventions engage not only the youths but also parents and school personnel [e.g., (Heyne et al., 2002; Melvin et al., 2017)]. Promisingly, meta-analyses of truancy interventions (g = 0.46; (Maynard et al., 2013)) and school refusal interventions (g = 0.54; (Maynard et al., 2018)) have demonstrated overall positive and significant effects on attendance.

The fact that truancy and school refusal are commonly perceived as different types of SAPs does not mean that they are mutually exclusive. Both may be associated with mental health problems and disorders, including anxiety, depression, behavioral problems, and other disorders such as autism spectrum disorder and attention deficit/hyperactivity disorder (Egger et al., 2003; Heyne et al., 2019). Therefore, adopting a modular transdiagnostic approach that address different types of SAPs, as well as different types of co-occurring mental health problems appears to be a viable and innovative path forward on conceptual, empirical, and clinical levels (Kearney & Graczyk, 2020). Although evidence-based interventions are available for several mental health problems, many youths with emotional or behavioral problems do not receive them (Costello et al., 2014). Studies show that transdiagnostic approaches have a potential for high scalability, that could make evidence-based practice more readily available (Jeppesen et al., 2021; Weisz et al., 2017).

Evidence supports the effectiveness of modular CBT interventions in addressing anxiety, depression, and behavioral problems (Ginsburg et al., 2020; Jeppesen et al., 2021; Weisz et al., 2012). In a cluster randomized clinical trial, Weisz and colleagues (2012) found that youths receiving modular CBT had fewer diagnoses post-treatment compared to those receiving usual care. However, an individual-level randomized clinical trial of the same modular intervention did not demonstrate its superiority over usual care (Weisz et al., 2020). In a recent individual-level randomized clinical trial, Jeppesen and colleagues (2021) found that a modular transdiagnostic CBT, Mind My Mind (MMM), outperformed management as usual in a community setting across various clinically relevant domains for youths experiencing emotional and behavioral problems. Ginsburg and colleagues (2020) also showed that modular CBT was effective in reducing symptoms of anxiety, in a school-based outpatient setting. These findings collectively suggest that modular transdiagnostic CBT holds promise as a viable approach for addressing the complex needs of youths with multiple co-occurring problems, including SAPs.

No prior study has used a modular transdiagnostic CBT approach specifically tailored to address SAPs. Several studies, however, have used CBT to increase school attendance among youths diagnosed with mental health problems (Hannan et al., 2019; Heyne et al., 2002; Reissner et al., 2015; Walter et al., 2010). In a comprehensive examination, Reissner and colleagues (2015) evaluated the impact of a protocol-guided multimodal treatment (MMT) for youths with confirmed mental disorders and SAPs, comparing it to treatment as usual (TAU). The MMT included modules of CBT, family counseling, school-related counseling, and a psychoeducational exercise program. Interestingly, both arms witnessed significant improvements in attendance, but no difference emerged between the two. Although there is an association between SAPs and mental health problems, youths can also develop SAPs without having a mental health diagnosis. Thus, it becomes imperative to develop programs capable of addressing a broad spectrum of SAPs, ensuring that youths with SAPs, with and without coinciding mental health diagnosis, receive the help they require.

We developed the Back2School (B2S) program, a manualized outpatient CBT intervention for youths displaying SAPs. It was designed to be used alongside the modular MMM-CBT manual. The MMM manual comprises evidence-based CBT methods and techniques organized into disorder-specific modules, targeting both subclinical and clinical levels of anxiety, depression, behavioral disturbances, and trauma-related problems. The B2S manual refers to relevant material from the MMM manual and aims to increase school attendance by using CBT procedures from that manual, including parent management, contingency management, incentives for attendance, cognitive restructuring, and exposure-based practices.

Research consistently indicates a significant increase in self-efficacy for managing situations related to school attendance following CBT interventions for school refusal (Heyne, 2022; Heyne et al., 2011; King et al., 1998). There is also research indicating that truancy interventions can improve school engagement and self-efficacy (DeSocio et al., 2007; Mann et al., 2015). Moreover, it is suggested that intervention that increases self-efficacy may yield reductions in mental health problems such as anxiety and facilitate reengagement with schooling (Heyne et al., 2015). At the core of the B2S intervention lies the strategic use of graded exposure, whereby youths have success experiences as a result of successively manageable exposure-based tasks, helping bolster their self-efficacy.

Our preliminary feasibility trial of the B2S intervention (i.e., B2S implemented concurrently with MMM-CBT) yielded promising results, demonstrating its feasibility and acceptability in addressing SAPs. This trial revealed significant improvements in both youths’ school attendance and mental health, coupled with high levels of treatment satisfaction (Lomholt et al., 2020). Building upon these insights, we used the feasibility trial findings to refine the B2S manual and strengthen the current study protocol. Key modifications to the B2S manual were implemented to improve collaboration with schools, involving revisions in the planning and implementation of school meetings. We identified possible limitations regarding the school attendance data provided by the municipality, and therefore we added parent-reported school attendance data (e.g., hours of attendance) as one of our primary outcomes (Lomholt et al., 2020; Thastum et al., 2019).

The primary aim of the current study was to rigorously evaluate the effectiveness of B2S, comparing it with an active comparator control arm that received TAU. We formulated three primary hypotheses. First, we hypothesized that the B2S intervention would outperform TAU in increasing both the number of hours and days of school attendance. Second, we hypothesized that B2S would outperform TAU in reducing symptoms associated with emotional, behavioral, and social difficulties, ultimately diminishing their interference with daily life. Our final hypothesis was that the B2S intervention would lead to a significantly greater increase in youths’ self-efficacy for managing school situations and parents’ self-efficacy for managing a SAP when compared to TAU. Additionally, we conducted a comparative analysis of participant rating of treatment satisfaction between the two groups.

Participants meeting inclusion criteria received a complete description of the study, gave written informed consent, and were subsequently randomized to one of the two intervention arms: B2S or TAU. Randomization was performed using a computer-generated random digit procedure with two possibilities (B2S and TAU). The primary outcome, increase in school attendance, may be affected by youths age and by the amount of school non-attendance prior to intervention (Heyne et al., 2015). Therefore, to ensure balanced arms, the randomization was stratified on the presence of age (first to fourth grade [younger] or fifth to ninth grade [older]) and amount of school non-attendance (< 50% [low] or > 50% [high]). To maintain similar intervention arms sizes, the randomization was conducted using permuted block randomization. The block size and computer-generated allocation lists were concealed from the research team overseeing the RCT study, throughout the recruitment period. Following randomization to either B2S or TAU, all participants received written and verbal information regarding intervention allocation from the research team. To increase retention of participants in the TAU group, the families were offered a gift card (value 200 DKK/26 EUR) after completing the Post and/or FU assessment. After randomization participants in the B2S group were notified of the time and place of the start of the intervention, and participants in the TAU group were urged to contact their school to get help for the SAP. The participant’s school and the school leader were also notified of the outcome of randomization. The last follow-up assessment (3FU) was in December 2019, for both interventions.

The B2S intervention was delivered by 8 psychologists, and 37 clinical psychology master students serving as co-therapists. The psychologists worked as school psychologists in Aarhus Municipality and were hired part-time to deliver the B2S intervention. Prior to the current study, all psychologists and co-therapists had limited knowledge and experience with CBT and were regarded as novice CBT therapists. Psychologists and co-therapists received a 6-day training course prior to the trial with an introduction to clinical assessment, case formulation, and the B2S and MMM manuals. They also received four 1-day booster training sessions during the study period. In total, psychologists and co-therapists received between 60 and 80 h of training, comparable to previous studies using modular trans-diagnostic CBT (Jeppesen et al., 2021; Weisz et al., 2012). All psychologists and co-therapists received weekly face-to-face group case supervision by specialists in clinical child psychology, for the duration of the trial (i.e., the two years in which interventions were delivered).

The study was powered to provide 80% power at the 5% (two-tailed) significance level to detect a standardized effect size within the range of 0.46–0.54 difference in the primary outcome (change in school attendance), following Maynard and colleagues (Maynard et al., 2013, 2018). To account for the expected attrition rate of 10% (Heyne et al., 2002; King et al., 1998; Last et al., 1998; Wu et al., 2013), the required sample size of 70 participations per intervention group was raised to 77 per group. Mixed linear models (MLMs) were used to determine the time × group interaction effects and the effects of intervention groups over time. MLMs allow for handling of the time variable as a time-varying covariate. Thus, the time variable was calculated from each participant's day of assessment for all time-points. MLMs accommodate missing values in the dependent variable as an integral part of the analysis and do not compromise statistical power. All MLMs were estimated using the maximum likelihood method and based on the intention-to-treat sample. All models included a random intercept, and the slope was specified as random if improving the model fit (evaluated by a significant change in the -2 log-likelihood (-2LL) fit statistics (Heck et al., 2012). To explore the robustness of the primary and secondary outcomes, we estimated unadjusted mean changes and differences in mean changes with corresponding 95% CIs for significant time × group interactions. The mean score from the ESQ measure was compared between intervention groups using an independent t-test. Effect sizes were calculated for all outcome measures, expressed as Cohen’s d, with 0.2, 0.5, and 0.8 considered small, medium, and large effect sizes, respectively (Cohen, 1988). Cohen’s d was derived from the F test calculated as d = 2 × √ (F / df). IBM SPSS Statistics, v.26 was used for all analyses.

In this study, we compared outcomes of interventions for SAPs using two approaches: a manualized modular transdiagnostic CBT program called B2S, and TAU. Contrary to expectations, B2S did not yield a significant advantage in increasing school attendance over TAU. There were substantial within-group improvements in school attendance for both B2S (medium effect sizes: d = 0.73 for hours, d = 0.54 for days) and TAU (medium effect sizes: d = 0.60 for hours, d = 0.68 for days), but no significant differences between the two, refuting our first hypothesis. However, B2S outperformed TAU on most of the emotional, behavioral, and social difficulties measures. Effect sizes between the groups ranged from small to medium (d = 0.29–0.58) and included youth- and parent-reported ratings of total problems, emotional symptoms, and the impact of problems, youth-reported social difficulties related to peers, and parent-reported youths conduct problems, aligning with our second hypothesis. Additionally, B2S yielded significantly higher increases in youths’ self-efficacy (academic/social stress and separation/discipline stress) and parents’ self-efficacy. Effect sizes were significant for both youths (d = 0.29 and 0.47) and parents (d = 0.53), supporting our third hypothesis. Participants in the B2S group also reported significantly higher treatment satisfaction compared to TAU, with a large effect size.

The average intervention time for families in each condition was almost equal. The TAU group received a wide range of interventions from public and private service providers, averaging 13.4 h, closely matching the average 15.0 h for families in the manualized B2S intervention. Thus, the B2S intervention’s length was comparable to TAU, and it could probably be implemented in municipalities without extra costs except for B2S program training and supervision. While we lack specific details on TAU intervention content, Danish law mandates public schools to collaborate with parents to help youths attend school (Danish Ministry of Children and Education, 2019), likely resulting in many youths and families receiving school-initiated interventions to increase school attendance. Notably, 56 out of 60 youths in the TAU group received school-provided interventions (on average 11.0 h).

The TAU group demonstrated a notable increase in school attendance, surpassing the overall effect sizes reported in previous meta-analyses of interventions for truancy (d = 0.46) and school refusal (d = 0.54) (Maynard et al., 2013, 2018). This aligns with findings from Reissner and colleagues (2015), who also observed positive outcomes related to school attendance among participants receiving TAU (Reissner et al., 2015). Given the TAU group’s steadily declining attendance in the year preceding participation in the current study (Online Resource 2), it is likely that participation in TAU contributed to the increase in attendance. The increase in school attendance within the TAU group can be explained in several ways. First, it is possible that both the therapists providing B2S and the service providers in the TAU group were equally effective in improving school attendance. The superior outcomes in B2S, namely improvement in emotional, behavioral, and social difficulties, as well as increased self-efficacy for both youths and parents, and higher treatment satisfaction, may be attributed to the therapeutic components unique to B2S, including graded exposure and disorder-specific modules from the MMM manual. Second, there is a possibility of contamination, where participants in the control arm received active intervention (Magill et al., 2019). In our study, when families were randomly assigned to TAU, schools were informed that the family had been encouraged to contact them for help with their child’s SAP. This proactive approach may have led schools to intensify their support and parents to insist on more help from them. In total, 30 out of 44 schools (68%) had youths participating in both B2S and TAU, potentially influencing the support given to TAU families, because school personnel were familiar with B2S. Additionally, psychologists delivering B2S worked part-time on the project while also serving as school psychologists in municipalities during the study. Although they were instructed not to incorporate B2S elements in their work as school psychologist, their experience with B2S could have unintentionally influenced the intervention for those in TAU. Furthermore, schools may have adopted B2S school meeting practices in their work with TAU youths. This raises the possibility that elements from B2S inspired the intervention received by the TAU group, potentially enhancing its effectiveness. Third, a considerable portion of TAU participants received intervention from private providers, potentially impacting the TAU group’s outcomes. Specifically, nearly one-third of TAU youths (19 out of 60, or 32%) received on average 13.4 h of help from private providers, including private psychologists. Considering these findings, the substantial proportion of the participants in the TAU group receiving support from private providers, suggests that the available public services were not readily available in some instances, due to high demand or were insufficient to meet the needs of the youth and families.

Prior research has revealed discrepancies and weak associations in school attendance data reported by registries and parents (Keppens et al., 2019; Lomholt et al., 2020, Johnsen et al., 2022). This study addresses these issues by using both registry and parent-reported data as primary outcome measures. At baseline we found a large discrepancy between the attendance reported for the last 10 days in the school attendance registries (17.7% had 100% attendance) and what parents reported regarding their children's attendance (1.4% had 100% attendance). In the assessment of the sociodemographic characteristics and previous attendance patterns of the current sample (Johnsen et al., 2022), 22 parents reported that their child had 100% school absence in the three months before recruitment, while registry data showed only two participants had 100% school absence in the same time period (Johnsen et al., 2022). It is unlikely that nearly one-fifth of the youth participating in the current RCT trial had no absence at all from school at inclusion, since the families had sought treatment for SAP, and since the inclusion criteria was 10% absence or more. The results therefore question the reliability of the municipalities register data. Since it is mandatory for schools in Danmark to record absences rather than presence, we can speculate that schools might be omitting to register absences for children who have been absent for extended periods, automatically categorizing them as presents instead.

The study boasts several strengths, including the large sample of youths displaying SAPs, an ecologically valid design with close collaboration with municipalities and intervention conducted by their psychologists, and thus the use of a highly active comparator in TAU. However, there are noteworthy limitations. Firstly, potential contamination from the B2S condition may have influenced TAU effectiveness. Secondly, as the TAU group were free to seek and receive any help and support available (see Online Resource 1), we do not have information on the number of staff involved in the TAU intervention, nor do we know the level of training and expertise of the staff in the TAU intervention. This is a limitation, as the characteristics and training of TAU staff may have affected the outcomes of the TAU intervention. Additionally, we lack information on whether and to what extend B2S-trained staff may have provided intervention to TAU participants (although they were instructed not to). Lastly, concerns arise regarding the validity of the register-based measure of school attendance, as we uncovered discrepancies between parent-reported attendance and registry-based attendance data, with parent reports indicating lower attendance levels (Johnsen et al., 2022).

Our findings have clinical implications. The sample comprised youths with substantial school absence and emotional and behavioral symptoms at a clinical level, impacting their daily lives (Johnsen et al., 2022). B2S delivered in an outpatient, school-based setting yielded benefits in emotional, behavioral, and peer problems, overall life impact, and youths’ and parents’ self-efficacy. Thus, combined with high satisfaction rates reported by both youths and parents, suggest B2S as a viable intervention for SAPs in an outpatient, school-based setting. Notably, the psychologists conducting B2S were previously novices in CBT, but with brief training, manual use, and weekly supervision, they helped achieve positive outcomes in school attendance and mental health. These results, coupled with acceptable competence and adherence measures (Bjaastad et al., 2015) and significantly higher parent-rated satisfaction in the B2S group, suggest that non-expert clinicians can successfully administer B2S with proper training and supervision in an outpatient setting. Moreover, B2S outperforming TAU in reducing youths’ symptoms and enhancing their self-efficacy could have lasting positive effects on their learning and mental well-being. Strengthening self-efficacy may also serve as a protective factor in coping with future life challenges. These findings suggest, that B2S could increase the quality of treatment of SAPs in schools and other outpatient settings, possibly reducing barriers to care (Kearney & Benoit, 2022). Kearney and Graczyk (2014) have proposed a stepped care model featuring a three-tiered service delivery approach: universal, targeted, and intensive interventions for SAPs. Tier 1 interventions are directed toward all students; Tier 2 interventions are directed toward at-risk students who require additional support; Tier 3 interventions are directed toward students with severe or complex problems who require a more individualized and concentrated approach. Following the current evaluation, the B2S intervention seems applicable as an intervention for SAPs (e.g., truancy, school refusal, school withdrawal, and school exclusion; Heyne et al., 2019), in either Tier 2 or Tier 3. Nonetheless, B2S is not suitable for all youth with SAPs, as it requires motivation and commitment from the family to work on increasing school attendance.

To conclude, this study marks the first evaluation of a transdiagnostic CBT outpatient intervention for youths displaying SAPs using a rigorous experimental design. Contrary to our expectations, the B2S intervention did not significantly improve school attendance compared to TAU; both groups showed attendance improvements. However, as hypothesized, youths participating in B2S demonstrated greater improvement in emotional, behavioral, and peer problems, as well as self-efficacy, compared to TAU. Additionally, parents in the B2S group reported greater increase in self-efficacy in helping their child attend school regularly, compared to TAU. The next step in the B2S evaluation involves conducting an economic evaluation using pertinent data up to and including the school year of 2023/24. This evaluation will compare the B2S group with the TAU group, considering cost benefit assessed through obtained grades, youth education, employment, and income (Thastum et al., 2019). Future research should explore the enduring effects of B2S on school attendance and investigate whether changes in emotional, behavioral, and peer problems, and self-efficacy, mediate attendance outcomes. Further, given the non-significant difference between B2S and TAU with respect to school attendance, future research should investigate effect modification to identify subgroups that may benefit more from B2S based on factors such as school attendance levels, sociodemographic factors, family characteristics, the severity of youths’ emotional or behavioral problems, and psychiatric diagnosis (Heyne et al., 2015). Moreover, understanding what aspects of the B2S program work best for specific groups could inform intervention improvements.