Universal delivery of a dialectical behaviour therapy skills programme (DBT STEPS-A) for adolescents in a mainstream school: feasibility study

The social and emotional wellbeing of young people is an area of increased focus for policy and practice. Schools are required to provide a holistic approach to education that includes teaching and implementing programmes designed to promote resilience and address difficult behaviours. Preliminary studies in the USA have shown promise for DBT STEPS-A (Dialectical Behavior Therapy- Skills for Emotional Problem Solving- Adolescents) but there have been no published UK evaluations. The aim of this study was to adapt DBT STEPS-A for a UK setting and test the feasibility of training teachers to deliver this universal programme as part of a mainstream school curriculum. The programme was delivered by teachers to students aged 13–14 years (n = 183) using a waitlist (non-randomised) controlled design. We investigated potential changes in level of emotional and behavioural problems using the Strength and Difficulties Questionnaire (SDQ) and Difficulties of Emotion Regulation Scale (DERS) and changes in psychological skills using the Child and Adolescent Mindfulness Measure (CAMM) and DBT Ways of coping checklist (DBT-WCCL). The views of students (n = 50) and teachers (n = 4) were explored using a participant satisfaction questionnaire. Feasibility outcomes indicate that it is possible to recruit participants from a school setting and to train teachers in the delivery of DBT STEPS-A. A statistically significant treatment effect was observed on the SDQ prosocial subscale (p = 0.04) with a large effect size (0.92). The intervention and some measures were acceptable to participants and teachers. Areas to improve future implementation are discussed.


Dialectical behaviour therapy
Dialectical Behaviour Therapy (DBT) [39] was originally developed as a treatment for adults with emotion dysregulation and impulsive disorders such as Borderline Personality Disorder (BPD) and for patients who engage in suicidal and selfharm behaviours. It is based on behaviour therapy enhanced with mindfulness practice and acceptance-based philosophy and practice [40]. The standard form of DBT is delivered by a team of mental health professionals and comprises individual therapy, group skills sessions, telephone coaching, and consultation meetings for the DBT team. Research supports the effectiveness of the treatment [e.g., [41][42][43][44][45][46][47][48]. Randomised controlled trials have found a reduction in suicidal behaviours, deliberate self-harm, depression, hopelessness, anger, eating disorders, substance dependence, and impulsiveness respectively. Furthermore, increases in general and social adjustment as well as positive self-esteem have been found [48]. Since its development, the use of DBT has expanded for use with adults presenting with a wide range of difficulties.
Standard DBT for adults has been adapted for adolescents (DBT-A: DBT for Adolescents, Miller et al., 2007) and delivered in inpatient and outpatient settings [46][47][48][49] by the addition of multi-family skills training and optional family sessions. The evidence base for DBT-A is emerging from randomised controlled trials showing reductions in suicide attempts and non-suicidal self-injury [45,49] and depression [46,50,51], and long-term reduction in self-harming behaviour [47] and a meta-analysis which found preliminary evidence for reductions in non-suicidal self-injury [52].
As interest in standard DBT has grown a popular alternative format has been the development of shorter skills-only programmes. Such condensed programmes serve the function of providing skills training to individuals who do not meet criteria for BPD diagnosis but present with emotion dysregulation and may or may not have a history of deliberate selfharm behaviour [52][53][54][55]. As the name suggests, DBT skills only programmes are different to standard DBT programmes in that the skills group is the sole modality of treatment delivered, however, caution is required in drawing comparisons. The term 'skills-only' is often applied to interventions that are based on DBT skills training and then adapted in a variety of ways. For example, they may include one other mode such as individual sessions, telephone support, or consultation meeting, cover fewer modules, vary the number and length of sessions, add disorder specific material, and amend language and examples used. Under this umbrella term DBT skills only has been adapted for a variety of presentations: mood disorders [55,56], binge eating disorder [57], bulimia nervosa [58], anorexia and bulemia 69, intellectual disability [59], Attention Deficit Hyperactivity disorder [60], incarcerated adolescents [61], and carers of adults with dementia [62]. The evidence for DBT skills only adaptations for BPD individuals as well as for individuals with other disorders and diagnoses who also experience emotion dysregulation, has shown a number of positive outcomes including reductions in drop-out rates, suicidal ideation, psychiatric symptoms, emotional symptoms (depression, anxiety, anger, hopelessness), hospital admission and visits, and improvements in quality of life, affective control, and mindfulness [63][64][65]. In a systematic review of DBT skills as a stand-alone treatment, Valentine et al., [66] found preliminary evidence for the feasibility and acceptability of this approach, although the significant variations in implementation of standard DBT (to adapt to different populations) and lack of robust study design mean it is hard to reliably generalise findings. Valentine et al., concluded that DBT skills shows promise as an effective treatment for those who do not meet the criteria for more severe disorders.
The evidence base for adaptations of DBT delivered in schools is emerging and initial studies have shown promising reductions in target problems and encouraging feedback on feasibility and acceptability of the intervention. Nelson-Gray [67] trialled a 16-week skills group with adolescents diagnosed with Oppositional Defiance Disorder (ODD) and found a significant increase in interpersonal strength, and reductions in ODD symptoms, externalizing behaviours, depressive symptoms, and internalizing behaviours. Ricard et al., [68] delivered a 4-week skills group based on DBT, with adolescents attending a disciplinary alternative education programme, reporting reductions in behavioural distress. Zapolski and Smith, [69] delivered a 9-week adapted skills group for at risk youth delivered by school nurses and health care staff, demonstrating feasibility and preliminary efficacy. Scotti [70] conducted a feasibility and pilot trial of a 12-week schoolbased DBT skills groups for adolescents with eating disorders and body image concerns. Improvements were found in behavioural aspects of eating disorder, as well as in externalising and internalising symptoms.
DBT STEPS-A (Skills Training for Emotional Problem Solving for Adolescents: [71]) is an adaptation of DBT specifically designed for delivery in school settings. The programme adapts the skills teaching components into lesson plans for delivery in school settings which can be delivered as a universal, social-emotional skills teaching programme within the school curriculum, or to targeted groups in schools with additional DBT components. The skills taught in Mindfulness (focusing attention/awareness), Interpersonal Effectiveness (managing relationships), Distress Tolerance (coping effectively with crises), and Emotion Regulation (understanding and regulating emotions) are useful skills for adolescents in clinical and universal settings as difficulties in these areas are common amongst adolescents [68]. The standard curriculum of DBT STEPS-A is delivered over 30 weeks to the universal student population (Tier 1) and can be enhanced with supports from Educational Psychology and Child and Adolescent Mental Health Services (CAMHS) as required for Tier 2 (specialist mental health provision) and Tier 3 (specialist mental health provision for complex cases) students. Preliminary results of the application of DBT STEPS-A in school settings have demonstrated its promise for feasibility and acceptability [70,71]. Flynn et al., [72] evaluated the implementation of a 22-week DBT STEPS-A program in Irish secondary schools finding a significant reduction in emotional symptoms but no reduction in dysfunctional coping or increase in skills use. Martinez et al., [73] trained school counsellors in a rural USA high-school to implement a 20-week DBT STEPS-A program and found an increase in understanding and acceptability of DBT skills when compared to business as usual. Chuangi et al. [74] evaluated staff and stakeholder experiences of implementing a 19-lesson program of DBT STEPS-A in a low-income school in the USA demonstrating its acceptability and feasibility for teacher delivery. In Ireland, Flynn et al., found reductions in emotional and internalising problems, reductions in depression anxiety and social stress [72] alongside evidence for the feasibility and acceptability of the intervention [70].
Schools are increasingly facing the challenge of addressing emotional and mental wellbeing as well as concerning behaviours such as self-harm [23,75]. The rationale for selecting DBT STEPS-A was based on the proven contribution that skills training and practice has in the reduction of suicidal and self-harming behaviours [49] and utility of DBT skills training in addressing a range of mental health and behaviour problems [76] making DBT STEPS-A a relevant choice of intervention.

This study
DBT STEPS-A has not been evaluated in a UK school setting and has been the subject of only four previous published evaluations internationally [71][72][73][74]. Feasibility testing covers a number of aspects relevant to study design that need clarification before proceeding to a full trial such as acceptability of the intervention and measures to participants, their willingness to participate, implementation facilitators and barriers, collecting valid and reliable data [77].
The main aim of the current study was to examine the feasibility of delivering DBT STEPS-A to a Welsh mainstream school population. In a non-randomised pilot study design, teachers were trained to delivered DBT STEPS-A as a universal programme taught in 15 lessons as part of the Personal Health and Social Education (PHSE) curriculum. Based on the limited research evidence for this programme and its novel use in a Welsh education system, the study objectives were: 1. To assess the feasibility of data collection as part of the program and their acceptability to students. 2. To evaluate student experiences e.g., relevance of skills, materials used, of the intervention qualitatively through a participant satisfaction questionnaire. 3. To evaluate the experiences of teachers. of the intervention qualitatively through a participant satisfaction questionnaire. 4. To evaluate whether DBT STEPS-A shows promise in improving emotional problem-solving skills in a secondary school population through changes in the outcome measures used.

Study design
The study was a wait-list control design, with two cohorts who received the intervention consecutively. Secondary school education in Wales covers 10-16-year-olds in year groups from 7 to 11. Year 9 was identified by the school as the most appropriate group to receive the intervention as they perceived that emotional and behaviour challenges often increase at this point in adolescent development. In addition, the curriculum structure meant that this year group's timetable was not impacted by exam requirements resulting in greater flexibility for delivery. The students (ages [13][14] were already in seven form (class) groups and were allocated to the intervention (4 forms, n = 113) or wait list group (3 forms, n = 70) based on their school form and the school timetable for PHSE lessons. Measures were administered at three time-points: baseline (Time 1), end of programme for cohort 1 and baseline for cohort 2 (Time 2), and end of programme for cohort 2 (Time 3).

Participants
Students from year 9 (n = 183) were between the ages of 13 to 14, and identified as 71 male, (38.8%) and 112 female (61.2%). Teachers (n = 4) identified as 1 male and 3 females. All students were eligible to be included in the study as DBT STEPS-A lessons were provided as part of the school PHSE curriculum during the academic year 2016-7, and all students were required to attend the lessons but had the choice to opt out of the data collection. Students were only excluded (n = 6, 3.28%) if they or their parents requested to opt their child out of the evaluation. Fifty students were identified by the school as having additional (special educational) needs (categorised by the school as: emotional or behavioural difficulties, visual impairment, general learning difficulties, moderate learning difficulties, hearing impairment, physical/ medical difficulties, speech and language communication, dyscalculia, or dyslexia); these data were not provided to the research team. Chi square analysis of intervention-control group characteristics found no significant group differences on the basis of gender, special needs, or baseline Strengths and Difficulties Questionnaire (SDQ) scores χ 2 (1) = 0.201, p < 0.001 (see Table 1).

Procedure
Participants were recruited from a mainstream secondary school within which four class teachers and all year 9 students (n = 189) were invited to take part in the study. Ethical approval was obtained from Bangor University Research Ethics Committee. Governance approval was gained from the governing body of the school and a whole staff presentation was given outlining the programme and study proposal. Prior to the intervention, four Class Teachers, four staff from the Senior Management team, one administration staff member, and three staff from the pastoral support team attended three days of training in DBT STEPS-A delivered by two clinicians experienced in DBT (A PhD student and accredited DBT Therapist and a Clinical Psychologist); in preparation to deliver the programme as part of the compulsory PHSE curriculum classes over the course of 15 weekly lessons. As there was an existing relationship between the school and local CAMHS, two staff (Mental Health Nurses) from the local Child and Adolescent Mental Health Services (CAMHS) also attended the training to ensure that the local mental health service was aware of the programme being offered to students. An information presentation was given to the whole school staff by the first author at the beginning of the academic year to raise awareness of the programme and the study taking place. Participants were informed about the intervention at the beginning of the academic year in their first STEPS-A class, and were provided with the programme information leaflet, research participant information leaflet, and assent form. Parents were informed about the STEPS-A programme and research element via an information leaflet at the start of the year.

Measures
Feasibility, (particularly teacher and student engagement and acceptability of the measures and curriculum) were assessed using quantitative and qualitative methods. Students also completed the following questionnaires:

Strengths and difficulties questionnaire-self rated (SDQ-S)
The SDQ [78] is a well-validated instrument proven to be effective in identifying clinically significant levels of behavioural and emotional problems in children (3-16 years old) which includes positive and negative items; there is a self-report version for adolescents (11)(12)(13)(14)(15)(16). The 25-item questionnaire has five subscales and measures four problem domains: emotional symptoms, conduct problems, hyperactivity, and peer problems as well as prosocial behaviour. Respondents rate the statements as not true, somewhat true, or certainly true, based on their behaviour over the past 6 months. Example items from each subscale include: ''I am often unhappy, down-hearted or tearful'' (emotional symptoms); ''Generally obedient, I usually do as I am told (conduct problems); ''I am constantly fidgeting or squirming'' (hyperactivity); ''I am usually on my own, I generally play alone or keep to myself'' (peer problems); ''I am helpful if someone is hurt, upset or feeling ill'' (prosocial behaviour). Each item is scored 0-2, with subscale scores across five items ranging from 0 to 10. Higher scores indicate higher levels of problems (except for the Prosocial behaviour scale). The sum of the four problem domains generates a ''total difficulties'' score. Scale scores can be pro-rated if at least three items in each subscale are completed. Screening cut-offs for the SDQ are available from a nationally representative community sample of 10,438 five to 15-yearolds recruited by the Office for National Statistics from a base of 12,529 eligible children, stratified by postal sectors and socio-economic group and weighted to correct for non-response bias by age and sex [78]. The normative school-age data derived allow for identifying children with 'close to average' (80%) 'slightly raised' (10%) and 'high' (5%) and 'very high' (top 5%) levels of difficulties (www. sdqin fo. org). The high (and low for the pro-social scale) cut-offs were used.

Difficulties in emotion regulation scale (DERS)
The DERS [79] is a brief, 36-item, self-report questionnaire designed to assess multiple aspects of emotion dysregulation. There are 6 subscales measuring non-acceptance of emotional experiences, difficulty engaging in goal directed behaviour, impulse control difficulties, lack of emotional awareness, limited access to emotion regulation strategies and lack of emotional clarity and a total score. Each item is rated on a 5-point Likert scale ranging from 1 (almost never) to 5 (almost always), 11 of which are reverse scored. More than one missing item invalidates a subscale score and more than seven invalidates the total score; average scores can be used to compute missing items. Sample items from each scale include: 'When I'm upset, I feel guilty for feeling that way' (non-acceptance), 'When I'm upset, I have difficulty concentrating' (goals), 'I experience my emotions as overwhelming and out of control' (impulse control), 'When I'm upset, I acknowledge my emotions' (emotional awareness), 'When I'm upset, it takes me a long time to feel better' (strategies) and 'I have difficulty making sense out of my feelings' (clarity). The measure yields a total score as well as scores on each subscale. There is no official cut-off, and higher scores indicate greater problems with emotion regulation. The DERS has been shown to have high internal consistency, good test-retest reliability, and adequate construct and predictive validity [79].

Child and adolescent mindfulness measure (CAMM)
The CAMM [80] is a 10-item questionnaire designed to assess mindfulness skills (present-centred awareness and nonjudgemental evaluation of internal experiences) in youth over the age of 9 years. Items are scored on a 5-point Likert scale from 0 (Never true) to 4 (always true). Sample items include: 'At school, I walk from class to class without noticing what I'm doing' and 'I push thoughts away that I don't like' . There is no cut-off score, but higher scores indicate higher levels of mindfulness. The CAMM has been assessed as having adequate internal consistency and being a useful measure of mindfulness skills for school-aged children and adolescents [80].

DBT ways of coping checklist (DBT-WCCL)
The DBT Ways of Coping Checklist [81] is 66 item self-report measure with good to excellent psychometric properties, designed to assess the use of coping skills taught in DBT. Items are scored on a 4-point Likert scale from 0 (never used) to 3 (regularly used). The WWCL has 3 subscales measuring coping by use of DBT skills (DBT skills subscale, DSS) and coping by use of dysfunctional means (Dysfunctional Coping Scale DCS1) and Blaming others (DSC2). Sample items include: 'Talked to someone about how I've been feeling' (DSS), 'Kept feelings to myself' (DCS1) and 'Figured out who to blame' (DCS2). The scoring is the mean of the rated items for each subscale.

Participant satisfaction questionnaire
Informal feedback was obtained by means of a brief participant satisfaction questionnaire completed by a sample of students (n = 47, 25.7%) from both groups and by the teachers (n = 4) at time point 3 to explore views about the intervention. Students were asked nine questions about factors such as how understandable they found the lessons, views on the materials, whether they would recommend the classes to peers, and what they enjoyed about the lessons. Teachers were asked for their views about the support and preparation provided, delivery of the intervention, implementation issues, accessibility of the materials, and how helpful the skills were. Most questions were rated on a Likert scale and some were open ended (Appendix F).

Intervention
DBT STEPS-A, is a DBT based skills training programme for school settings designed to teach skills for managing emotions, behaviours and relationships, within the context of an academic environment [69] (as noted earlier). The curriculum structure is similar to DBT skills-only programmes, with Mindfulness being covered for at least 2 sessions at the start of each of the modules (Distress Tolerance, Emotion Regulation, and Interpersonal Effectiveness). For this study full curriculum was shortened by removing the repeated mindfulness sessions, combining some sessions and removing the module tests to enable it to fit into the school timetable (see Table 2.). Mindfulness was already being taught in years 7 and 8 to all students, so this foundation was already in place. Teachers were supported by the investigator throughout the study at regular face-to-face meetings and/or telephone contact as needed.

Data analysis
The flow of participants in the research evaluation was summarised using a study flow diagram (Fig. 1). Due to school requirements in delivering the curriculum there was a natural opportunity to have a wait list control group for this study. Initial tests for linearity, homoscedasticity, normality and independence for outcomes all met the required assumptions. Attempts to run an analysis including individual classes and taking into account the effect of clustering at class level were unsuccessful, as the model would not converge most likely due to the small number of clusters. A Robust Regression bootstrapped analysis with bias-corrected accelerated confidence intervals was therefore run (in the statistical software package STATA) to explore intervention effects. Changes in SDQ scores in cohort A (intervention) compared to Cohort B (control) at end of intervention were examined with baseline score, SEN status, and cohort (i.e., study group) as covariates. Effect sizes were calculated. Missing data from both cohorts (due to a school data collection error) meant that it was not possible to analyse any other outcomes (CAMM, DERS, DBT-WCCL) at time point 2. However, it was possible to analyse the follow up data for cohort B for pre/post analysis between Time points 2 and 3 using paired sample t-tests to measure changes on the SDQ, DERS, CAMM and DBT-WCCL. Feedback from the participant satisfaction questionnaires was analysed descriptively.

Feasibility and acceptability of the outcome measures
Completion of the measures varied considerably between time points and cohorts; the the full information about % completion is in Fig. 1.

Feedback from students
The satisfaction questionnaire was completed at time point 3 by 47 (25.7%) students taken from both cohorts. Most students (66% of those completing the questionnaire, n = 31) enjoyed the lessons somewhat or very much with the opportunity to work with friends, engage in activities and learn about emotions and how to manage them as the most highlighted aspects:  and most students (74.5%, n = 35) reported finding the skills taught a little or somewhat useful: "They helped organise my mind. " (student 32) "Because it is good learning about how you can prevent some situations from happening. " (student 20) 87.2% (n = 41) described the programme as useful for students like them. There was no consensus on which skills were considered the most helpful, with all being mentioned by someone and some students describing them all as helpful: "They all help you when you get into a situation. " (student 46). For those who did not like the programme, a few students did not enjoy the lessons as at (10.6%, n = 5): "I honestly, honestly hated the course. The only good parts were a few of the coping techniques and emotion names. " (student 19) and a small number (29.8%, n = 14) reported that the skills had not helped solve any problems, and some (n = 10, 21.3%) found them not relevant for them or their current circumstances: "I knew quite a lot anyway. " (student 3). "None of this helped me because I don't suffer these problems. " (student 10). "I found them interesting even though I didn't relate to them. " (student 37). "Good but they didn't relate to my lifestyle. " (student 44).
There was no agreement on what students found most challenging or least helpful, apart from some who disliked completing the evaluation questionnaires. Opinions on what should be changed before delivering the programme again were split, with some happy with the programme as delivered (4 comments), and some suggestions it should be limited to those who need it or delivered as a separate lesson. By far the most comments (11) made were around including more activities, games and discussions, and reducing the paperwork (6 comments): "I felt as though there should be more activities in a group like discussing a problem, as you think more. I feel as though it is boring writing it down also. " (student 43).

Feedback from teachers
Feedback was given by all teachers who taught DBT STEPS-A (n = 4). Most of the teachers (n = 3, 75%) reported feeling well prepared to deliver the programme and were supported by previous experience, detailed materials, and access to support sessions: "the delivery and information was well put together for us. " (Teacher 1). "Following the 3 days training I felt quite prepared to teach DBT. " (Teacher 4).
Being confident with the material and principles took time and for some the timing of the training and start of the delivery got in the way: "having the summer break meant that I had forgotten many skills until I arrived back in September. " (Teacher 4) Teachers reported that delivery of the programme was also hindered by limited time to cover the material, interruptions to lessons, students being taken out of lessons and teachers called away to other duties.
Overall, the DBT skills were well received with all the teachers identifying the skills as relevant to the challenges typically faced by their students, whilst recognising that the universal delivery also meant that not all the students saw the relevance of the skills which impacted on their engagement: "Some said that they did not need the skills. Some realised the benefits and were very forthcoming with their opinions. " (Teacher 2) All the teachers found the skills easy to understand (3 fairly, 1 very) and described them as very useful for themselves: "after delivering and gaining experience from teaching it has improved my skills tenfold!" (Teacher 3) And three reported seeing use of skills and changes in student behaviour: "many pupils (students) have shared good and positive experiences of putting the skill to use outside of the lesson. " (Teacher 3) "some behavioural problem and behaviours such as fighting have changed how the pupils deal with the aftermath of a fight. " (Teacher 3) In agreement with student feedback, one clear difficulty highlighted by teachers was the accessibility of the materials. Whilst one of teachers described the lesson plans and handouts as poor and three as fair, all identified the need for power point presentations, as this is the standard mode of classroom teaching and these had to be developed from the lesson plans in the manual. Teachers also thought the handouts needed to be made more student friendly (3 rating them as poor and 1 as fair suggesting adaptations such as reducing the amount of information on them and having more pictures, examples and workspace were made). For some students a significant amount of differentiation was needed to make the materials accessible. Having opportunities to collaborate and share resources was suggested as an additional improvement. These challenges aside, the overall experience was a positive one and thoughts were already turning towards the next opportunity: "I have enjoyed the experience teaching something totally out of my comfort zone" (Teacher 2) "I enjoyed having the opportunity to participate in this study… exciting to see the rewards. Hoping it will continue in school and that we will teach it again. Once we have power points to use will be much easier-especially for new teachers. " (Teacher 1)

Evaluation of student outcomes
Robust Regression analysis was used to examine the differences between the two groups (intervention and control) to determine whether there were changes on the SDQ-S (n = 111). Results (see Table 3) showed a marginally significant change on the SDQ prosocial subscale (β = 0.909, p = 0.050) indicating an improvement on this subscale for the intervention group. Other comparisons were not statistically significant. Effect sizes with 95% BCa Confidence Intervals are shown in (Table 3). The largest effect size was seen for the prosocial subscale (SMD 0.92), with a small effect size for emotional problems (SMD 0.37), although not in the expected direction, and very small to negligible effect sizes for hyperactivity (SMD 0.13), conduct (SMD 0.06), total difficulties (SMD 0.05), and peer problems (SMD −0.01).
Pre-post intervention analysis on cohort B was carried out using additional measures (see Table 4). Whilst there were no significant changes on any of the SDQ-S or WCCL subscales, there were statistically significant changes on all subscales of the DERS (except for Lack of emotional awareness: AWARE), and on the CAMM. Moderate to small effect sizes were found for the DERS Goals

Discussion
The primary outcomes of interest in this study were the feasibility of using DBT STEPS-A in a mainstream school setting, the acceptability of the intervention and the measures, including questionnaire completion rates. Results from this study demonstrate that the delivery of a 15-week DBT STEPS-A curriculum in a mainstream setting is acceptable to staff and students. There was mixed feedback about the acceptability of completing outcome measures and this with the challenges posed by the school collecting data is reflected in the low completion rates. Several factors affected the variable completion rates. Measures were completed during the school timetable, administered by teachers and initial communication errors led to gaps in the data collection at the correct time point. Whilst completion rates generally improved at the later time points, some students highlighted that they disliked completing the evaluation questionnaires. Several of the measures are lengthy and a lack of student willingness to persist was one likely contributor to this decline. Having a robust system for data collection, whether for future studies or routine in-school evaluation is essential, as is appropriate orientation of students to the rationale of the programme, which if linked to individually relevant goals could help increase willingness to collect data.
Feedback from participants also highlights aspects of the programme delivery and implementation that may need further consideration to support wider implementation and evaluation. The most common feedback was about the materials used and need to adapt this for a school setting (by having lesson PowerPoint presentations) and to be creative about the content to include more activities and less reliance on handouts. Any manual is unlikely to be a one-size-fits-all option and adapting materials for lessons is a key task for teachers. Increased familiarity with the programme may feed Table 3 Linear model of predictors of SDQ-S scores, with 95% bias corrected and accelerated confidence intervals. Confidence intervals and standard errors based on 1000 bootstrap samples SD Standard Deviation *Beta coefficient associated with p <.05SDQ-S Strengths and Difficulties Questionnaire-Self-rated SDQ S scale  into greater confidence in adapting the materials and devising appropriate and relevant activities, although there is a time commitment associated with this. Several aspects of this study support the feasibility of delivering DBT STEPS-A. Prior to delivery, the programme was identified as a good fit with the existing school priorities and policies. Preparation, planning and support were key to the implementation; it was supported by the school governing body as well as several senior staff members, administration staff and the local CAMHS who attended the initial training alongside the staff delivering the programme. Information about the programme was shared with students, parents and the wider school in advance. Attention to the timing of training and ensuring access to opportunities for ongoing consultation, collaboration, sharing and developing resources, added additional support for delivery.
The universal delivery (to all the students in the identified year group) made it straightforward to plan delivery as once space in the timetable had been allocated, there was no need to screen for inclusion, any potential stigmatising effects of being identified were removed, and having a transdiagnostic focus, it was applicable to all. The manualised presentation of the programme in lesson plans designed for delivery by staff, meant that with some adjustments to accommodate timetable space, detailed guidance, suggested exercises and handouts were available, making efficient use of school resources. Overall, the skills were well received with all the teachers identifying the skills as relevant to the challenges typically faced by their students, as well as for themselves and students finding the skills interesting and helpful.
Preliminary findings relating to outcomes were also reported, with controlled comparisons showing a statistically significant positive change (and large effect size) in pro-social behaviour in the intervention group. The skills taught in DBT STEPS-A seek to increase awareness of emotions and emotional experience, increase emotion regulation, and improve relationships. Possibly by gaining skills in emotional awareness and responding, this improved awareness of others and their needs, and considering the feedback from students who particularly enjoyed the working together aspect of the programme, suggests one possible explanation for this outcome. Other SDQ-S subscales (apart from emotional problems) showed no meaningful change. The small effect size increase in emotional problems in the intervention group may represent a true increase in problems (which would be concerning), although another possible explanation is that similar to the findings of Johnson et al., [10], awareness increased, which is to be expected as the focus of the intervention is on increasing awareness of emotional experience. It is possible that detecting change immediately after the end of the skills teaching does not allow for skills to generalise and having a longer follow up might detect meaningful change This study also considered the validity of the measures used. With regard to their ability to detect change, it is possible that the SDQ-S, whilst valid as a measure of child mental health, is too broad a measure to capture the changes anticipated as a result of DBT STEPS-A. Having multiple informants (teachers and parents) could increase the sensitivity of the SDQ-S [77,78]. The DERS and CAMM are both more specific to the targets in DBT STEPS-A; the DERS specifically measures difficulties in emotion regulation, and the CAMM measures use of mindfulness skills, so improvements suggest the potential for change as a result of receiving the intervention. Regarding the acceptability of the measures, some students did not like completing the evaluation questionnaires and their length may have been one aspect of this; the DERS and SDQ-S (36 and 25 items respectively) are relatively lengthy, whereas the CAMM is only 10 items, perhaps making it more acceptable. All the measures however are available at no cost, which is a significant advantage.
Delivering fidelity within flexibility [25] remains a challenge requiring the balancing of numerous factors and challenges that present in real-world settings such as schools. As in previous studies [5] there were challenges in the implementation. A significant amount of the timetable was given over to the programme and the content proved challenging to cover and with space in the curriculum under pressure. Other priorities such as visits by outside speakers, sometimes disrupted delivery. Further consideration may need to be given to how the lessons could incorporated into the curriculum with more flexibility e.g., deliver modules across different year groups. One clear difficulty highlighted by both staff and students was the usability of the materials. Classroom teaching is frequently delivered using PowerPoint presentations and staff developed their own from the lesson plans in the manual. Whilst this is a typical task for teachers, it was clear that in this study the manual was not an off-the-shelf solution and time needed to be allocated to develop the lesson plans and handouts. Teaching a skills programme for managing emotions is unlike teaching a typical curriculum topic and can require a shift in teaching style for staff, as suggested by the comment from Teacher 2: "I have enjoyed the experience teaching something totally out of my comfort zone", which needs highlighting during staff training. Becoming confident with the material and principles take time and training for staff should focus on these aspects, using the lesson plans as guidance and reference rather than considering it a ready-made option. Ahlen et al. [8] found improved outcomes for participants whose teachers attended additional supervision and so attention to the timing of training and ensuring access to opportunities for ongoing consultation, collaboration, sharing and developing resources may enhance the quality of delivery. Whilst consultation was offered, the uptake in this study was low, perhaps reflecting the pressures on staff time. Regular consultation is a feature of DBT skills teaching programmes [39], which is perhaps different from standard teaching practice and so exploration of how to build this into a school context may need additional attention. Maintaining staff motivation and sharing skills could be an important aspect of ensuring the sustainability and integration of skills teaching in a school context.
The content of the programme has a promising evidence base, as it is broadly based on CBT principles, is transdiagnostic, and relatively short [13]. However, the universal delivery meant that not all the students saw the relevance of the skills, which impacted on their engagement. Consideration may need to be given to whether a selected or targeted delivery would be more appropriate or whether additional orientation to the preventative value of these skills as more of an 'emotional skills first aid kit' , would help students understand the rationale for learning about them, even if they do not see the immediate relevance. Deciding which level of intervention is delivered will depend on the school's aims and intended outcomes.
To ensure sustainability, the implementation of any programme requires attention from the start, ensuring staff availability for future delivery and responding to feedback from participants. Lessons from this study point towards attending to factors that build sustainability into implementation. Whilst the skills taught in DBT STEPS-A are applicable to a wide variety of situations and young people, it will not cover every need and so its implementation needs to be linked to the wider school provision and be part of a pathway to further assistance where this is needed. Students, staff, governing bodies and parents need to understand why a programme is being offered, and the changes expected, and so ongoing measurement of impact/outcomes is crucial [7,25,37]. Communication about the expected outcomes as well and rationale for relevant measures of change (standardized measures and other indicators e.g., change in referral to other agencies) used at individual and school level may support sustainability.
Teaching skills is only one part of the implementation process and being able to generalise skills is recognised as part of the learning process in DBT [40,82] and this is no different in schools. Skills need to move from being taught in the classroom, to being used in everyday life. Whilst some signs of this were reported in the feedback from teachers, the process of generalisation needs attention to ensure that the wider environment (home and school) supports the use of these skills. The homework/practice element of the programme is designed to address this and whilst this was included in the programme delivery, this aspect was not evaluated as part of the study, so would benefit from being explored in future studies. Structuring the environment may also need attention, for example, including staff other than those teaching the skills directly (e.g., pastoral staff ), so that they know to support students in using skills when needed.
This study has limitations that affect the interpretation of the findings. Only one main outcome measure was used for the controlled comparison analysis and only from a self-report perspective. Using teacher and parent ratings of outcomes could capture changes that might not have been noted through self-report. In addition, potential longterm impact on academic progress warrants attention in future research. As the programme was offered as part of the curriculum, completing the teaching was compulsory which could have affected student interest and willingness to participate; an assessment of motivation to change could be included to understand individual responses in future research. Collecting follow-up data after the end of the intervention might also detect whether there are changes once skills use has been consolidated for a period and detect whether skills use had generalized. Future studies should also include attention to teacher adherence to delivery of the programme, such as number of lessons delivered and content covered, as no such data were available in the current study.