Introduction

Worldwide, 25–31% of adolescents suffer from common mental disorders (Silva et al., 2020) including symptoms of depression or anxiety (Patton et al., 2014). In Sweden, the number of children and adolescents with mental health problems is increasing rapidly (The Public Health Agency of Sweden, 2018). Easily accessible interventions may decrease health care costs and prevent mental health problems at a young age, which otherwise may lead to mental disorders in adulthood (Patton et al., 2014). However, before introducing large-scale interventions in school settings, more research is needed on feasible, easily applicable practices that are possible to fit in the school schedule.

Mindfulness practice is gaining popularity worldwide; it originates from meditation (Kabat‐Zinn, 2003) and fosters compassion and non-reaction toward emotions (Bishop et al., 2004). Stress and anxiety could thereby be lessened (Zelazo & Lyons, 2012). In adults, mindfulness has shown positive effects regarding psychological wellbeing (Eberth & Sedlmeier, 2012); (Bohlmeijer et al., 2010), depression, anxiety and stress-related disorders (Blanck et al., 2018; Sundquist et al., 2015).

In youths, mindfulness has been suggested to be beneficial concerning different outcomes, such as stress reduction, resilience, cognitive performance (Dunning et al., 2019; Zenner et al., 2014), emotional wellbeing and self-regulation (Klingbeil et al., 2017). Studying the use and the potential benefit of mindfulness practices in schools have gained popularity globally over the last ten years (Kuyken et al., 2022; Montero-Marin et al., 2022; Tudor et al., 2022; Zenner et al., 2014). Meta-analyses have shown that teachers and students reported a generally positive attitude toward mindfulness. In addition, mindfulness was perceived safe with few adverse events or incidents (Zenner et al., 2014). Compared to other school interventions aiming to prevent mental health problems among school students, a network meta-analysis suggested a reduction of anxiety symptoms after mindfulness and relaxation-based interventions, whereas only weak evidence was shown for more commonly used interventions with cognitive behavioral therapies. None of the interventions showed evidence of reducing symptoms of depression, and the overall evidence was considered weak due to possible bias and small-study effects (Caldwell et al., 2019).

However, even though the body of evidence of mindfulness in schools are growing, the potential benefits of mindfulness are questioned. The heterogeneity among studies is large, and there seems to be a lack of consensus regarding how and for whom mindfulness could potentially be effective (Montero-Marin et al., 2022; Tudor et al., 2022). Even though some studies indicated positive effects, especially in terms of cognitive performance and stress resilience (Zenner et al., 2014), more recent studies have not been able to show superiority of mindfulness in relation to standard teaching in regards of cost-effectiveness and mental health promotion (Kuyken et al., 2022). In one recent study, mindfulness was even found to be worse than standard teaching for risk of depression (Montero-Marin et al., 2022). Small effects on symptoms of depression and anxiety have been described previously (Kannan Kallapiran et al., 2015a, 2015b; Kallapiran et al., 2015a, 2015b; Zenner et al., 2014) including in a meta-analysis based on results from RCTs with active control groups (Dunning et al., 2019). In contrast, no effects have been shown in meta-analyses where only anxiety was studied (Odgers et al., 2020; Ruiz-Íñiguez et al., 2020). Significant effects were observed in three out of 18 studies in Ruiz-Íñiguez’s meta-analysis, but due to large heterogeneity among the studies and small sample sizes, no firm conclusions could be drawn.

Some small RCTs and pilot studies have reported improvements in symptoms of depression and anxiety after mindfulness interventions in adolescents (Johnson & Wade, 2019, Raes et al., 2014). Another recently published meta-analysis on school-based mindfulness interventions’ effect on anxiety and depression found overall significant pre-post effect sizes, but the effects were small, and due to the findings, large-scale implementations were questioned (Phillips & Mychailyszyn, 2022). Different responses to mindfulness have been suggested in previous studies, where mindfulness sessions provided by trained school teachers have been shown to be more effective in reducing mental health symptoms, whereas outside facilitators have been shown to be more effective in increasing mindful awareness (Carsley et al., 2018). So far, most mindfulness interventions have consisted of long and extensive mindfulness interventions, between 45 to 120 min (Kallapiran et al., 2015a, 2015b; Zenner et al., 2014). This might constitute a challenge in feasibility and in terms of implementing the method practically in a school setting. Since a significant correlation between practice time and effect size for anxiety and depression symptoms have not been found in a comprehensive meta-analysis of mindfulness interventions (Phillips & Mychailyszyn, 2022), the effect of briefer sessions are worthy of evaluation. Another large-scale intervention study in a school setting, also investigating a brief and teacher-led method, is currently taking place in England (Hayes et al., 2019), but few large interventions have been performed in other countries, particularly on specific age groups and on both sexes. To better understand for whom mindfulness might be beneficial, it is important to study potential moderating factors. Age, sex and baseline mental health have been suggested to modify the effect of mindfulness, and more studies about effects in different subgroups have been desired (Tudor et al., 2022) to identify if there are certain groups for whom the method is more or less effective or even harmful (Tudor et al., 2022).

Well-designed controlled studies are needed to identify the best timing and delivering of mindfulness and to examine whether mindfulness practices are feasible and have positive effects in school settings. This is to ensure that novel interventions based on certain interests do not surpass evidence; instead, limited resources should be utilized in the most beneficial way in this critical age among students. To make mindfulness training in school easily applicable, and to enable more students to take part of the practice, we have performed an intervention with brief mindfulness-sessions. The sessions were delivered by the schoolteachers after they were educated via an adapted mindfulness instructor course. The intervention will be referred to as brief in regards of the short mindfulness (5–11 min/session) sessions that were provided in the classrooms.

Purpose of the Study

In this controlled study, built on previous research, we hypothesized that brief teacher-led (oral or by audio files) mindfulness sessions in the classroom can have a positive effect on symptoms of anxiety and depression, and that this effect might be different in different subgroups. We investigated the effects of a 10-week school-based mindfulness intervention on symptoms of depression and anxiety among 9–16-year-old school students by examining differences between the intervention and control group in the potential change of depression or anxiety symptoms after 10 weeks. The intervention was integrated in the daily school schedule and included brief sessions delivered by the schoolteachers. Subgroup analyses for age, sex, and levels of symptoms of depression and anxiety at baseline were performed to evaluate whether the effect of mindfulness was modified by certain characteristics. One additional subgroup analysis examined the potential difference in outcome between teacher-led mindfulness practice and mindfulness given by audio-files.

Methods

Setting and Recruitment of School Classes

This 10-week controlled intervention study took place in Sweden’s southernmost county, Skåne, which has approximately 1.4 million inhabitants. School classes were recruited by contacting the principals at primary and middle schools. Between 2016 and 2019, a total of 65 schools were invited via letters and a few additional principals contacted the research group themselves. A total of 12 primary and middle schools with children of ages six to 16 years (corresponding to school grades 0 to 9) were included in the study. Of all invited schools, 18% chose to participate in the study. To increase representativeness, the schools were selected from both high and low socioeconomic areas. The present study includes children in school grade three and upwards since the primary outcomes, symptoms of depression and anxiety, were measured with questionnaires that were validated in the Swedish version from nine years of age (Beck, 2001; Beck et al., 2001). The principals assigned the school classes to either the mindfulness intervention or no intervention (control) group. They were asked to have a mix of mindfulness and control classes in their schools, and to have at least one control class at each school and for each grade. The study went on all year around, except for summer holidays, and schools were recruited continuously. All data collection was performed before the outbreak of the Covid-19 pandemic.

Recruitment of Children

The research team first informed the principals and teachers about the study at workplace meetings. After acceptance from principals and teachers, parents were invited to meetings with the research team where they were provided oral and written study information. If they could not attend the meetings, written study information was sent to their homes. Written informed consent from the parents was a prerequisite for study participation. Parents were encouraged to obtain oral consent from their children by using an age-adjusted research information. Participation in all mindfulness sessions and questionnaire sessions was entirely voluntary and children could refuse participation at any time without giving any further explanation. No exclusion criteria were applied; the study was open to all children. No financial compensation was given to the schools, teachers, parents or children.

Mindfulness Instructor Training, Schoolteachers

The mindfulness sessions in the intervention group were led by the schoolteachers themselves or, in some cases, school nurses (one school) or special educators. Before the start of the intervention, the schoolteachers participated in a mindfulness instructor course led by one of the authors (JS) who is a family physician and mindfulness instructor trained in mindfulness-based cognitive therapy (MBCT) at the University of Oxford Mindfulness Centre and in the mindfulness two step certification program at the Swedish Mindfulness Center (MFC). The schoolteacher mindfulness instructor course was based on contemplative mindfulness traditions (Goldstein & Kornfield, 2001), mindfulness-based stress reduction (MBSR) (Kabat-Zinn, 1990), and MBCT (Segal et al., 2002). The mindfulness instructor course was given at no cost. We did not collect any data regarding the teachers’ motivation for the study. The teachers were not paid and the whole intervention including the training was done during normal school/work hours. The teachers were given a basic education and no continuing education or training. The principals decided which teachers that should participate. After the basic education and some home practice, only a minimal preparation was needed from the teachers. All teachers in the classes assigned to the mindfulness intervention took part in the mindfulness instructor course, which comprised four to six sessions, with a mixture of lectures, meditative exercises and personal mindfulness training for eight weeks. The teachers were trained in how to guide individuals and groups to develop a greater awareness and ability to cope with stress and other difficulties. The teachers received an audio file with recorded instructions and were informed to practice mindfulness 20 min per day at home. They also received a manual with descriptions of the 10-week intervention for different school grades (3–6 and 7–9) and age-appropriate audio files that could be used in the classrooms if the teachers decided to.

Mindfulness Intervention and Control Group

The mindfulness intervention comprised 10 weeks of daily classroom sessions. Mindfulness was encouraged to be practiced at a voluntary time each day when the teachers found it was appropriate. The mindfulness sessions were customized for different grades with shorter sessions for the youngest children (5–9 min in grades 3–6 and 7–11 min in grades 7–9) and comprised different themes such as compassion and body relaxation. All sessions started with a 1.5 min introduction. A short reflection could be added at the end of the sessions. Themes and lengths of all exercises are shown in Supplementary Table 1. The teachers filled in weekly schedules about the number of mindfulness sessions per week, minutes of mindfulness training and mode of the mindfulness sessions (led by teacher or audio file). Children in the control group did not receive any treatment or other intervention and continued their schoolwork as usual, with the exception that they filled in the same study questionnaires as the intervention group before and after the 10-week period.

Data Collection

A couple of days before the start of the 10-week intervention/study period, all children, regardless of group assignment (intervention or control group), filled in an age-appropriate (school grade 3–5 or 6–9) questionnaire about themselves; family situation, school situation, origin, leisure time, perceived health, common diseases, stress, health behavior and the primary outcome symptoms of depression and anxiety. Depression and anxiety symptoms were measured by two of the five Beck Youth Inventories (BYI); Beck Depression (BDI-Y) and Beck Anxiety (BAI-Y). Each inventory consisted of 20 statements, which were rated by the respondents on a four-point scale, from never to always (Beck, 2001; Beck et al., 2001). The BAI-Y includes items that reflect fear, worrying, and other symptoms of anxiety, whereas the BDI-Y covers items associated with depression, such as feelings of sadness, and negative thoughts about oneself, others, and the future. The BYI have been frequently used and have been suggested to be a useful screening instrument in schools (Thastum et al., 2009). Both inventories have shown a high test–retest reliability among adolescents as well as internal consistency (Basker et al., 2007; Thastum et al., 2009). The questionnaire items regarding family and school situation, origin, leisure time, perceived health, common diseases, stress, and health behaviors were adapted from the public health survey of children and adolescents in Scania (Fridh et al., 2018; Lindström & Rosvall, 2018). The questionnaires were tested for readability by children in respective age categories. The same questionnaires were administered after 10 weeks when the intervention had ended. The children filled in the questionnaires in their classrooms with members of the research team present in case any questions arose. Students with language difficulties or dyslexia were provided extra help from the teachers or members of the research team when filling in the questionnaires. Both paper and electronic versions were used. Questionnaires that were filled in on paper were transferred to a database by study personnel via an electronic questionnaire in the research electronic data capture (RedCap) program. Electronically filled-in questionnaires were automatically transferred to the same database. All questionnaires were pseudonymized via a code, thus no individual could be directly identified. The questionnaires were treated as strictly confidential, and the children were informed that neither their parents nor their teachers were given access to their answers. For the safety of the children, when serious symptoms of depression or anxiety were identified, the parents were contacted by one of the authors (JS) who is also a family physician. The children were informed orally when they received the questionnaire that their parents would be informed if necessary. There were only three children where it was considered necessary to contact the parents. This procedure was approved by the ethical authority and all participants were informed about this safety precaution.

Statistical Analysis and Power Calculation

All analyses were repeated for the two Beck inventories: depression and anxiety, each subscale consisting of 20 items. The total score from each subscale was calculated in accordance with the instructions from the Beck manual. Missing values were also handled in accordance with the Beck manual. If more than two items were missing, the subscale was counted as missing. For those who had missing values on one or two items, the missing values were replaced with the mean number of the non-missing items (Beck, 2001). The Beck score was transformed to percentiles using a normative population including approximately 2400 Swedish school children, ages 9 to 19 from five different Swedish cities (Beck, 2001). Subject characteristics and primary outcomes were described with mean values and standard deviations, medians, interquartile ranges, frequencies and percentages.

We aimed to analyze differences between the entire mindfulness group and control group in the potential change of depression or anxiety symptoms after 10 weeks as well as changes within groups. Primary outcomes were analyzed with linear mixed models to account for the nested structure of repeated measurements. Equally correlated observations within individuals were assumed and all available data were used. Mixed models were also used to test the nested structure of students within classes and schools. To examine possible differences in outcome dependent on mode of delivery, the mindfulness group was also dichotomized into two groups, where students in classes who received more than half of the teacher led sessions were classified as the teacher led group and the students where more than half of the sessions were given via audio files were classified as the audio file group.

Assuming a difference in mean score of four of Beck’s subscale for anxiety (BUS-Å) from baseline to a 10-week follow-up (Beck, 2001) (SD = 13 for an effect size of 0.3) between students (grades 3–9) doing group mindfulness and controls, an alpha-value of 0.05 and a power of 80%, each group should include 166 students. Because of the correlation within classes, a cluster effect was taken into account by increasing the sample size with a design effect (deff) (Hemming et al., 2011). We assumed an intraclass correlation of 0.1 and a class size of 25 students and this resulted in a deff value of 3.4. Multiplying the individual group size with the deff value and assuming a loss to follow-up of 10% resulted in a cluster size of 628 (25 classes) per group, i.e., at least 1250 students (50 classes).

Ethical Considerations

The study was approved by the Regional Ethics Review Board in Lund (registration number 2016/299). Written informed consent was secured and the study was registered at ClinicalTrials.gov (identifier: NCT03327714). Individual serial numbers were used to ensure confidentiality.

Results

Of the 2536 students aged 9–16 years who were invited, a total of 1399 students (55%) agreed to be included in the study. More girls than boys consented to participate in the study (60 vs 40%). The students who did not consent to participate were slightly older than the students who consented (data not shown). Drop-out rates were low: seven students in the intervention group and 14 in the control group. See flowchart in Supplementary Fig. 1. Baseline characteristics are presented in Supplementary Table 2. A mean of 38 sessions per school class were given during the 10 weeks of intervention, with a mean of 10 min/session. Other characteristics are presented in Supplementary Table 3 in the Appendix.

For the primary outcomes, no significant change in symptoms of depression was seen in the mindfulness intervention group or the control group after the intervention. Symptoms of anxiety decreased significantly within both groups after the intervention; the mean change in the intervention group was − 2.2 (p = 0.004) and, in the control group, − 2.7 (p = 0.02). However, there was no significant difference in change of symptoms between the intervention group and the control group, 0.5 (p = 0.69) (Table 1). Accounting for correlation between individuals within the same classes and schools negligibly changed the estimates, indicating that clustering of observations was seen primarily within individuals rather than within classes and schools.

Table 1 Primary outcomes after 10 week follow-up
Full size table

Subgroup analyses regarding age, sex as well as Beck scores at baseline were performed and the results are presented in Table 2 and Fig. 1. These subgroup analyses showed that older children and girls scored higher in depression and anxiety scores at baseline compared to younger children and boys, respectively, in both the intervention group and the control group.

Table 2 Stratified analyses on primary outcomes for different covariates
Full size table
Fig. 1
figure 1

Change in mean Beck score from baseline to follow-up. *Much increased was defined as the 10% students with highest score (in depression for ‘Beck score, depression’ and in anxiety for ‘Beck score, anxiety’) at baseline

Full size image

The subgroup of students with high Beck scores (percentile 90 or above) at baseline decreased significantly in both symptoms of anxiety and depression after 10 weeks and in both groups. Symptoms among students in the control group decreased more than in the mindfulness group, especially in symptoms of anxiety where differences between the groups were significant (data not shown). Symptoms of depression increased slightly among the girls, especially in the control group (mean change 3.3, p = 0.01). For the boys in the control group, symptoms of depression decreased (mean change − 4.0, p = 0.009) while no significant change in symptoms of depression was observed among the boys in the intervention group (mean change − 0.5, p = 0.61). The difference in change between the intervention group and the control group was 3.5 (p = 0.06). Symptoms of anxiety were significantly decreased among boys in both the intervention and control group at follow-up compared to baseline, but no difference was observed between the intervention group and the control group, 1.6 (p = 0.43). Compared to the older children, younger children in both groups had lower scores of symptoms regarding both depression and anxiety and showed no change at follow-up. The older children had significant decreases in symptoms of anxiety in both groups (Table 2).

Subgroup Analysis Based on Mode of Delivery

The teacher led intervention was compared with the intervention given via audio files. In total, 16% of the mindfulness sessions were led by teachers, whereas most of the sessions (81%) were given via audio files. Three percent of the sessions were based on other methods such as relaxation or listening to music. As presented in Table 3, stratified analyses were performed on the group with students mainly led by teachers compared to the students mainly led via audio files. With the control group as reference, a significant reduction of symptoms of depression was seen in the teacher led group (mean difference − 7.3, p = 0.009), but not in the group who received the intervention via audio files (mean difference 1.5, p = 0.24). The same trend, albeit non-significant, was observed in symptoms of anxiety, where the scores were reduced -5.5 in the group led by teachers in comparison with the control group (p = 0.07), whereas a smaller difference was observed in the audio file group (mean difference 0.5, p = 0.71). The overall differences between the audio file and the teacher led group were significant for both depression and anxiety; the students’ symptoms of depression and anxiety decreased significantly more if the sessions were led by the teachers rather than given from audio files (mean difference − 8.8, p = 0.001, and − 6.0, p = 0.04, for depression and anxiety, respectively). Symptoms of anxiety and depression at 10 weeks’ follow-up stratified by type of intervention are presented in Table 3.

Table 3 Primary outcomes after 10 week follow-up stratified by type of intervention
Full size table

Discussion

In this study, we investigated the effects of an intervention with brief mindfulness-sessions, integrated in the school schedule, and delivered by the schoolteachers (oral or by audio files). Overall, our findings could not confirm any significant differences regarding symptoms of depression and anxiety between the mindfulness group and the control group. However, subgroup analyses revealed that there was a significant difference in the intervention group based on the mode of delivery of the mindfulness sessions. The group of children and adolescents who received mindfulness sessions led by their teacher (16%) had significantly fewer depression and anxiety symptoms compared to those who received the intervention via audio files.

The potential various beneficial effects of mindfulness among school children have been examined in several studies (Dunning et al., 2019; Hayes et al., 2019; Kallapiran et al., 2015a, 2015b; Kuyken et al., 2017; Perrier et al., 2020) but studies focusing solely on symptoms of depression and anxiety are still sparse and evidence is contradictory.

The prevalence of mental health problems differs by age and sex. Mental health problems are becoming more frequent among adolescents but has also shown to be increasing also in younger children (The Public Health Agency of Sweden, 2018). Due to limited resources, it is of uttermost importance to develop cost-effective preventive interventions to break this trend, but also to investigate when and for whom such interventions are most beneficial. It has previously been described that older school children may benefit more from mindfulness interventions than younger ones (Carsley et al., 2018; Johnson & Wade, 2019). Our research team observed that the youngest children seemed to have difficulties in paying attention during the data collection, which also raised an assumption that the youngest children might be too cognitively immature to benefit from the mindfulness sessions. It is also possible that mental health problems present differently in lower ages. In a relatively small study, Janz et al. found positive effects on attention and focus among younger children where mindfulness activities were integrated into daily school activities; however, mental health symptoms were not measured (Janz et al., 2019). In our study sample, no conclusive differences between younger and older children could be found.

When stratifying for sex, girls scored higher on anxiety at baseline compared to boys. Girls and boys tended to respond slightly differently to the intervention but with no significant differences between mindfulness and the control group. Sex differences after attending mindfulness have been observed before, slightly favoring girls regarding symptoms of depression (Volanen et al., 2020) and emotional wellbeing (Kang et al., 2018). It has previously been shown that the effect of mindfulness was more pronounced among clinical samples than non-clinical samples (Zoogman et al., 2015). In our study, a greater reduction of symptoms was found in the group with high Beck scores at baseline, which could have implied a better outcome among students with more symptoms or reflect a regression toward the mean. However, in our study, symptoms among the students in the control group with high Beck scores at baseline decreased significantly more than among the students in the mindfulness group.

The most interesting finding from this study was the difference in outcome depending on the mode of delivery. The group of children and adolescents who received mindfulness sessions led by their teacher had a significantly better improvement in both symptoms of anxiety and depression compared to those who received the intervention via audio files. However, those who received mindfulness sessions led by their teacher represented only 16% of the whole intervention group. This indicates that, when teachers are allowed to choose the mode of delivery, they prefer to not lead the sessions themselves. In this study, we wanted to investigate the effect of an easily applicable mindfulness intervention, integrated in the school schedule, and delivered by the schoolteachers, as an alternative to lengthy sessions delivered by external instructors. Using the schoolteachers instead of external instructors have previously shown greater effects on mental health problems, probably due to teacher’s familiarity with the students (Carsley 2018). The teachers had the opportunity to lead the sessions themselves or to use prerecorded audio-files. However, as stated before, the engagement and mindfulness skills of the instructor is of vital importance for the effect (Kenny et al., 2020). Even though the use of audio recordings has been proven feasible and well accepted in previous studies (Bakosh et al., 2016), their effect may be insufficient. Differences in the outcome depending on mode of delivery have been indicated in a previous meta-analysis, where a larger effect size was found when mindfulness was provided with face-to-face guidance (Blanck et al., 2018). Unfortunately, most of the teachers used the audio-files instead of guiding the sessions themselves. The reasons why the teachers chose respective method was not investigated, but there are many possible reasons. It could have been due to practical reasons, i.e., that they did not have time enough to prepare mentally for the sessions, or for opening a time window to complete other tasks while the children were engaged in the audio session. It could also be that the training given by the external instructors was too short to give enough comfort and confidence among the teachers to guide the sessions. It should be noted that, for many teachers, this was the first encounter with mindfulness. It is possible that their skills and knowledge would have increased over time if their practice had continued. It is also of importance to address that the use of audio-files could reflect a lack of engagement, which possibly could influence the effects negatively. Thus, it is possible that the better outcomes in the groups that did not use the audio-file was due to better teacher engagement in the orally-led group. From the teacher’s perspective, it might be more practical to use audio-files instead of leading the mindfulness sessions themselves; however, since no effects were seen in this group, this could not be recommended. Although we assumed that the brief intervention would be easily applicable and not take much effort from the teachers, most teachers still chose to use the audio-files. Reasons for choosing one mode of delivery or the other, and potential obstacles to lead the mindfulness orally, need to be investigated further.

It is possible that mindfulness interventions among healthy subjects have more prominent effects regarding other outcomes, such as cognitive performance and stress resilience (Zenner et al., 2014) rather than on symptoms of depression (Bluth & Eisenlohr-Moul, 2017; Volanen et al., 2020). Previous findings underline the important difference between emotional wellbeing and mental health problems, which do not always overlap (Haworth et al., 2017; Patalay & Fitzsimons, 2016). Interventions focusing on reduction of depression and anxiety symptoms could thereby contribute to a greater gain when targeting individuals with psychopathology (Zoogman et al., 2015). Identifying school children with already existing, or with risk of developing mental health problems, is therefore of vital importance, since poor mental health at a young age may increase the risk of developing mental illness later in life (Wolitzky‐Taylor et al., 2014). Recently published findings also indicate that other personality traits and skills such as acceptance, rumination and executive function might have mediating roles in the effects of mindfulness among adolescents and youths. It is still not fully known which individuals benefit most of mindfulness, and it is important to continue research the field not only how, but also for whom, mindfulness should be provided (Saarinen et al., 2022).

Since mental health problems are getting more widespread, it is of importance to develop group interventions that can complement individual cognitive behavioral therapy, which to date is the most used method for treating anxiety and depression among youths. For children with mental health problems, individual CBT is currently the most used method. It is possible that other types of group interventions than mindfulness have beneficial effects on symptoms of depression and anxiety in young individuals as well. For example, significant improvements have been found among girls with depression or anxiety after attending a dance intervention (Duberg et al., 2020).

Strengths and Limitations

The lack of randomization is one limitation that needs to be mentioned in this study. It implies a risk of selection bias that should be considered; however, baseline characteristics were well balanced between the two groups, and the baseline Beck scores were identical between the mindfulness and control group for both anxiety and depression symptoms (Table 1), which suggest that the groups were comparable. In addition, if any large selection bias has occurred, we expect that it would have resulted in overall positive findings, which was not the case. This is because the principals would most likely have selected classes and/or teachers that were more positive toward mindfulness training. We therefore judge that the possible bias due to the lack of randomization is most likely conservative. The results should be interpreted with this limitation in mind though. Despite our intention to provide a feasible and usable intervention, only 18% of all invited schools chose to participate in the study. An overall high workload and other demands, such as national tests in mathematics and languages, could have been reasons for choosing not to participate in the study. In the recruitment process of schools, it was more difficult to recruit participants to the control group than to the intervention group, which resulted in a difference in group size. To include both intervention groups and control groups at the same schools might have contributed to a “rub-off effect”, where the groups may have influenced each other. For example, it is possible that the intervention teachers spoke with the control teachers about the intervention and that the control teachers also implemented mindfulness to a certain extent in their classrooms. On the other hand, it decreased the risk of selection bias on a school level. The use of validated scales for evaluation of symptoms enhances comparability with previous studies. However, the Beck scales were developed to evaluate clinical symptoms of depression and anxiety and may not be useful in universal school settings. A high number of individuals with low levels of depression and anxiety at the outset contributes to a ceiling effect with little room for improvement and thereby a reduced chance to detect significant changes. It is also important to mention that adolescents’ mood fluctuates from week to week. Measurements at one specific timepoint gives us a snapshot, however, tell little about the students’ mental health over time (Alfano et al., 2009; Vandenkerckhove et al., 2021). Another limitation was that few teachers chose to lead the sessions themselves, and that no information was collected about the reasons why the teachers chose one way instead of the other. The difference between the groups in regards of mode of delivery could possibly be related to teacher engagement but this could not be further investigated. In this study, only self-reported data from the children and adolescents were analyzed. Multiple informants (parents and teachers) could possibly have widened the picture of the potential effect of the intervention, especially among the younger age groups. On the other hand, we chose not to involve the parents to protect the children’s integrity with the exception for those cases (n = 3) when parents were contacted for safety reasons.

A key strength of the study is that it investigates the effects of a novel intervention, and a practically appealing brief mindfulness-alternative, using the schoolteachers instead of external facilitators. This could be integrated in the school schedule without compromising with the daily activities. This is the first study of its kind performed in a school setting in Sweden. The study was relatively large, involving participants of varying ages and different socioeconomic backgrounds. In addition, the intervention was easily accessible, suggesting a high applicability, and the drop-out rates were low.

Conclusions

Although we could not detect any significant differences on symptoms of depression and anxiety between the mindfulness group and the control group, we found that the small group of children and adolescents who received mindfulness sessions led by their teacher (16%) had a higher degree of improvement compared to those who received the intervention via audio files. Future studies should examine whether empowering and equipping teachers with skills and knowledge to lead the mindfulness sessions themselves will have an impact on the outcome. Future studies also need to evaluate potential long-term effects and effects on additional outcomes, as well as which children and adolescents that benefit most from mindfulness training in different school settings. The present findings do, however, not support an introduction of large-scale mindfulness interventions in schools.

Author Contributors

All authors contributed to the conception, design, analysis and interpretation of data. Elin Areskoug Sandberg wrote the first draft of the manuscript, Karolina Palmér performed the statistical analysis, and all authors revised and approved the final version of the manuscript. All authors had full access to all the data in the study and takes responsibility for the integrity of the data in the study and the accuracy of the data analysis.