Evaluation of a group format of clinician-guided, parent-delivered cognitive behavioural therapy for child anxiety in routine clinical practice: a pilot-implementation study

Background: Randomised controlled trials suggest that cognitive behavioural therapy (CBT) delivered by parents who are guided, in groups, by clinicians (Group GPD-CBT) is an efﬁcacious and potentially efﬁcient treatment approach for child anxiety. The extent to which these results translate to routine settings is unclear. We evaluated Group GPD-CBT as delivered in UK routine clinical services. Method: Retrospective data regarding attendance and outcomes were routinely collected for 83 children whose parent(s) had attended Group GPD-CBT. Semistructured interviews were conducted with 14 clinicians who had delivered Group GPD-CBT. Results: By 3 – 8 months (M = 5.22, SD = 1.17) post-treatment, 70% of children were discharged or referred for support for other (non-anxiety) conditions, without any further intervention for anxiety. Of the subset ( N = 20) with available parent-report symptom data, there was a signiﬁcant decline in total anxiety score from pre- to post-treatment. Clinician interviews were analysed using thematic analysis. This revealed that clinicians found Group GPD-CBT to be acceptable and described it as a helpful, practical and empowering treatment for child anxiety. They highlighted additional beneﬁts associated with group process factors (e.g. peer support, enhanced engagement), although noted that some, particularly anxious, parents were reluctant to attend a group format. Conclusions: Results were promising regarding children’s outcomes following Group GPD-CBT delivered in routine practice. Group GPD-CBT was viewed by clinicians as acceptable and helpful, and group process factors were seen to provide additional beneﬁts. Some parents may ﬁnd it difﬁcult to attend a group format, suggesting that services should give careful consideration to how groups are presented and intro-duced to parents. childhood anxiety disorders in order to improve accessibility to psychological treatments. (cid:129) Previous research shows that clinician-supported parent-led cognitive behavioral therapy (CBT) in groups (Group GPD-CBT) is an effective low-intensity treatment for child anxiety disorders when evaluated within research trials. (cid:129) This is the ﬁrst pilot-implementation study to evaluate Group GPD-CBT in routine clinical practice in terms of clinical outcomes (anxiety symptomatology and service utilization) and clinician experience of delivering treatment. (cid:129) Group GPD-CBT was associated with a signiﬁcant reduction in child anxiety symptoms and most did not require further treatment. (cid:129) Clinicians reported that Group GPD-CBT has beneﬁts over individual treatment for child anxiety such as peer support and increased engagement. (cid:129) Clinicians highlighted that it is important to consider how services present and deliver Group GPD-CBT to parents in order to maximise parental uptake and engagement.

A 'stepped-care' approach may increase accessibility of evidence-based psychological therapies (Clark, 2011). Under the stepped-care approach, less resource-intensive ('low-intensity') treatments are offered to patients first and more resource-intensive ('high-intensity') treatments are reserved for those who do not, or can be reliably predicted not to, respond to low-intensity treatment (Bower & Gilbody, 2005). One such low-intensity treatment for child anxiety is guided parent-delivered CBT (GPD-CBT), in which parents are supported by a clinician in applying CBT techniques with their child. This approach can be considered to be a low-intensity treatment as it follows a 'guided self-help' format in which parents are guided by a therapist in working through a book that describes how to apply CBT principles in their child's day to day life. This guided-parent-led approach requires considerably less clinician contact time than traditional individual child CBT for anxiety (Thirlwall et al., 2013) and is more cost-effective than another brief psychological intervention for child anxiety . Several trials have shown that GPD-CBT is effective in reducing child anxiety (Creswell et al., 2010;Lyneham & Rapee, 2006;Thirlwall et al., 2013). Indeed, similar outcomes have been reported with approximately 5½ hr of clinician contact time in GPD-CBT (e.g.  to those found in trials of 14 hr of individual child CBT (e.g. Piacentini et al., 2014).
Delivering GPD-CBT with parents in groups, rather than individually, has the potential to further increase the efficiency of this approach, as well as bringing other benefits. Research into parent groups for children with other difficulties (e.g. long-term physical health conditions, eating disorders) has highlighted benefits such as reduced parental stress, isolation, loneliness and guilt (Shapiro, 1989;Shilling et al., 2013) and increased parental self-worth and empowerment (Shilling et al., 2013). Indeed, parents report valuing the normalising experience and peer support offered by groups (Goodier et al., 2014).
Group GPD-CBT has been found to be effective in treating anxiety in 2-16-year olds after 2-3 hr (Cobham, Filus, & Sanders, 2017;Waters, Ford, Wharton, & Cobham, 2009), 6 hr (Cartwright-Hatton et al., 2011) and 8 hr (Thienemann, Moore, & Tompkins, 2006) of clinician contact per family in research settings. However, the outcomes of Group GPD-CBT delivered in routine clinical (e.g. National Health Service; NHS) settings remain unknown. Due to the range of differences between treatment delivered in research settings compared with clinical settings (e.g. measurement of adherence, comorbidity of patients), it cannot be assumed that outcomes from treatment trials will necessarily translate into routine clinical practice (Kazdin, 2008). The collection of outcome measures as standard (e.g. Wolpert et al., 2016) in child and adolescent mental health services (CAMHS) in the NHS enables the evaluation of the routine delivery of Group GPD-CBT in these services. These data can be complemented by investigating children's service utilisation post-treatment. Service utilisation data may be particularly helpful as an additional indicator of outcomes, given recent findings that questionnaire measures may only modestly reflect child recovery status (Evans, Thirlwall, Cooper, & Creswell, 2017). Furthermore, as a central tenet of the stepped care approach is to maximise efficiency by reducing the demand for high-intensity treatments (Bower & Gilbody, 2005), it is crucial to measure post-treatment service utilisation to evaluate whether Group GPD-CBT is functioning successfully as a low-intensity intervention in routine practice. Together, symptom outcome measures and service utilisation data can inform on the usefulness of Group GPD-CBT both for families and for services developing stepped care models of treatment for child anxiety.
In addition to treatment outcomes, little is known about the acceptability of the approach to clinicians in routine practice. This is an important consideration as, for example, Shafran et al. (2009) highlighted that clinicians' attitudes towards CBT are one of the key obstacles to the dissemination of CBT, as clinicians who do not feel positive about a protocol may only deliver it partially or opt for alternative treatments. Creswell et al. (2010) reported high levels of clinician satisfaction with individually delivered GPD-CBT in routine clinical practice, including high clinician-reported probability that they would continue to deliver it. However, the experience of facilitating a group format of treatment brings additional challenges for clinicians, such as public speaking and managing group dynamics. As such, it is unclear how clinicians will view Group GPD-CBT. This study set out to evaluate Group GPD-CBT in routine clinical practice, by investigating the following questions: 1. What are the outcomes of Group GPD-CBT in terms of reductions in child anxiety symptoms and service utilisation post-treatment?
2. How do clinicians describe their experiences of Group GPD-CBT and its usefulness for treating child anxiety in routine clinical practice?
To investigate these questions, a mixed-methods approach was employed. Quantitative data were collected retrospectively based on routine outcome measures and post-treatment service utilisation by children whose parents attended Group GPD-CBT as part of routine clinical practice. As relatively little is known about the experience of facilitating Group GPD-CBT for clinicians, semistructured interviews were conducted as this qualitative approach offers the potential to enable a more in-depth understanding of the relevant issues and to generate new insights (Brannen, 2005).

Participants
Children and their parent(s). All quantitative data collection was retrospective from routine clinical records. Data were collected regarding 83 children whose parent(s) had attended one of 16 Group GPD-CBT programmes in one of four NHS Trusts, between 2014 and 2017. Clinicians were asked to only provide the researcher with the details of children who were not engaged in any concurrent psychological intervention for anxiety. As this was a retrospective service evaluation, the sample was opportunistic, and the researcher had no control over the characteristics of the sample. The majority of children (65%) had been referred by their GP, with some referred by other health professionals (11%) or their school (17%). This information was missing for 7% of children. The children were 51% female and ranged in age from 5 to 12 years (M = 9.08, SD = 1.66). Families had waited between 0 and 13 months for assessment (M = 2.66, SD = 2.28) and between 0 and 23 months between assessment and treatment (M = 4.16, SD = 4.93).

Clinicians. Service managers provided the researchers with
contact details of all current clinicians in their service who had delivered Group GPD-CBT for child anxiety. All of these clinicians were sent recruitment emails inviting them to provide interviews. Of 18 clinicians who were invited to participate, 14 provided interviews, after which point theoretical saturation was reached. The clinicians reflected a range of professional backgrounds that work within NHS Child and Adolescent Mental Health Services (CAMHS), including nurses (n = 6) an occupational therapist (n = 1), a social worker (n = 1), a clinical psychologist (n = 1), a psychological wellbeing practitioner (n = 1), and clinicians with no formal registration but considerable relevant experience (n = 4). The clinicians had worked with children and families for an average of 19 years (SD = 12.19), including an average of 7 years (SD = 4.26) in their current service. The majority (n = 12) had also delivered individual GPD-CBT. The clinicians worked in primary CAMHS (n = 9) and specialist CAMHS (n = 5) across the five NHS Trusts.

Materials
Measures. Waiting times and attendance. Parental attendance at each session was recorded as indicated by electronic clinical records.
Post-treatment service utilisation. Electronic clinical records were used to identify children's service use at the time of post-treatment review (M = 1.65 months post-treatment, SD = 1.25), and at the closest available point to 6 months (within a range of 3-8 months) post-treatment (M = 5.22 months post-treatment, SD = 1.17). If a child did not have a post-treatment review status or 6 months post-treatment status recorded, because they had already been discharged or referred to an alternative service, this was recorded as their status at that time point. For example, if the child was discharged at the post-treatment review with no further update by 6 months post-treatment, this was recorded as their status at 6 months post-treatment.
Revised Children's Anxiety and Depression Scale -Parent report (RCADS-P). The RCADS-P (Chorpita, Ebesutani, & Spence, 2015) is a parent-report questionnaire measure comprising 47 items scored on a 4-point scale (never, sometimes, often, always) which refer to symptoms of child anxiety disorders and depression. The RCADS-P has been demonstrated to be a reliable and valid measure for children from 3 years of age (Ebesutani, Tottenham, & Chorpita, 2015). RCADS-P data were collected from electronic clinical records, where available.
Topic guides and interview materials. Interviews followed a semistructured topic guide, which asked about clinicians' experience of delivering Group GPD-CBT and positive and negative aspects of the treatment (See Appendices S1 and S2). Where clinicians had delivered both Group GPD-CBT and Individual GPD-CBT (n = 12), they were interviewed about their relative experience of both.

Intervention
Parents received a self-help book (Creswell & Willetts, 2007) and attended a group in which they were supported by clinicians in applying the strategies with their child. The group treatment protocol (Willetts & Creswell, 2007a;Willetts & Creswell, 2016) specifies seven sessions of 90 min each. Eight of the 16 groups were conducted in this format; however, two Trusts chose to reduce this to five (n = 2 groups) or six (n = 3 groups) sessions, or to adapt the individual version of the manual (Willetts & Creswell, 2007b) to deliver four 60-75-min face-to-face group sessions and four individual 20-min phone sessions (n = 3 groups). The treatment consisted of psychoeducation, identification and modification of anxious cognitions, graded exposure and problem solving. Parents were asked to complete various between-session tasks, both independently and with their child.

Procedure
Data collection. Permission for collection of data was arranged in accordance with each NHS Trust's procedures. Clinicians provided informed consent prior to interviews and were instructed not to reveal patient-identifiable information during interviews. Confidentiality of interviews, and the value of honest feedback, was stated verbally and in research information sheets. Interviews ranged in duration between 27 and 62 min (M = 41.87, SD = 10.58) and were conducted either in person at the clinicians' workplaces (n = 11) or via telephone (n = 3). All interviews were audio recorded and transcribed verbatim. The researcher was employed with an honorary contract for the purposes of quantitative data collection within three NHS Trusts. Clinicians from a fourth Trust opted to email the anonymised data to the researcher. For the fifth NHS Trust, clinicians provided interviews only, due to time restrictions.

Analysis
Quantitative analysis. Descriptive statistics were conducted on attendance levels and post-treatment service utilisation. A paired samples t-test was conducted on total RCADS-P anxiety score between pre-and post-treatment. Chi-square tests and independent t-tests were conducted to investigate differences between those with and without complete RCADS-P data.
Qualitative analysis. Interviews were analysed using thematic analysis (Braun & Clarke, 2006). Coding of all interviews was inductive and recursive, in that codes were led by the data and were constantly reviewed and adjusted. The codes were reviewed using constant comparison techniques, both within and between transcripts, as recommended by Pope, Ziebland, and Mays (2000). An independent researcher [DOB] provided reflective discussion and feedback on interpretation of qualitative results.

Results
Outcomes of group GPD-CBT Attendance. Sufficient attendance data to calculate overall proportion of sessions attended were only available for 69 families. On average, these families attended 79.64% (SD = 23.15) of planned sessions. Session-bysession attendance for each format of Group GPD-CBT (5,6,7 and 8-session) are shown in Table 1.
Anxiety symptoms. Revised Children's Anxiety and Depression Scale -Parent data were only available within a month prior to the start of treatment and a month following treatment for 20 children. Among this subgroup, there was a significant decline in total anxiety score from pre-treatment (M = 46.00, SD = 17.06) to post-treatment (M = 34.95, SD = 13.19), t(18) = 3.40, p = .003, d = .63.
Service utilisation. As shown in Table 2, by the time of post-treatment review, 49% of children were discharged or referred for non-anxiety assessment/intervention, without further treatment for child anxiety. At 3-8 months post-treatment, the majority of children (70%) had not required further treatment for child anxiety and had been either discharged or referred for non-anxiety assessment/intervention. At both time points, a minority of children (22%-25%) were engaged in, waiting for, or discharged following, an additional intervention for anxiety.

Clinician experience of group GPD-CBT
Clinicians' descriptions of their experiences of Group GPD-CBT and its usefulness for treating child anxiety in routine clinical practice were captured by three overall themes: 'Acceptability of GPD-CBT', 'Accessibility of GPD-CBT', and 'GPD-CBT Treatment Outcomes and their Mechanisms'.
Acceptability of GPD-CBT. Clinicians generally had positive attitudes to Group GPD-CBT, describing it as 'enjoyable', 'refreshing' and 'a good bit of the week'.
Several commented that this was due to the group format and the opportunity to facilitate alongside a colleague. However, some clinicians highlighted that they felt nervous facilitating group interventions and, in some cases, felt that the individual format of GPD-CBT enabled greater depth when working with families. Conversely, one clinician perceived a positive aspect of the group format was that that she found it easier to adhere to the treatment protocol working in a group compared to an individual format. On this topic of adherence, clinicians generally reported that they adhered to the manual (see also under 'Accessibility of GPD-CBT') although several commented that they delivered it with a degree of 'flexibility', for example, in the way ideas were communicated or the timing of content within sessions. When it's in a group there is something about that distance. . . the lack of getting really into the nitty-gritty of the system itself. . . which means I think you can get through the theory in a slightly different way which I think can be helpfulit drifted less.
In terms of acceptability to families, clinicians reflected that parents are often initially sceptical of a group format. They highlighted the importance of presenting the rationale of GPD-CBT as based in evidence and empowering parents, to counteract stigma around attending a 'parenting group'. Other clinicians reflected that their own belief in the utility of Group GPD-CBT was important to engage parents effectively. Notably, despite reports of initial scepticism of some parents, clinicians described that those parents often gave positive feedback after attending-giving further weight to the importance of the initial presentation of Group GPD-CBT to parents.  n/a n/a n/a a Availability of session-by-session data varied for this group between N = 37 and N = 49. b Availability of session-by-session data for this group varied between N = 5 and N = 6. Accessibility of GPD-CBT. Clinicians highlighted that more anxious parents may find it particularly difficult to attend a group treatment and that sensitive management of this by facilitators is required.
Clinician 12: Some very anxious children have very anxious parents who just didn't manage the group.
Clinician 3: I think it comes down to facilitating, because if you've already met them and you know that they're anxious-you get quite good at picking it up so, you know, when it comes to any sort of talking out loud, you. . .kind of guide it very gently.
Clinicians varied in their views about the efficiency of taking a group approach. Some clinicians reported that Group GPD-CBT had been delivered in response to long wait-lists, implying presumed greater efficiency of this format. However, others felt that Group GPD-CBT was not necessarily more efficient than an individual format, due to the need for two facilitators and additional preparation, particularly if groups were small.

Clinician 1:
If you can get enough people there it's more efficient but it might not be necessarily. . .because it does take two people to run it and it does take quite a lot of setting up and. . .the higher number you have in the group, the more efficient it is, because if it's a smaller one you're not really making a lot of savings on that front.
A related issue under the theme of accessibility is perceived accessibility of GPD-CBT for the clinicians who delivered it. Clinicians described Group GPD-CBT as 'clear', 'accessible' and 'explainable'.
GPD-CBT treatment outcomes and their mechanisms. Clinicians described GPD-CBT as an 'empowering' and 'practical' approach. They reported good clinical outcomes for children and families following Group GPD-CBT. However, clinicians highlighted that families with more complex circumstances may find it challenging to fully engage with Group GPD-CBT, for example, where the child has an Autism Spectrum Disorder (ASD) or the family is going through stressful life events and/or is receiving input from multiple agencies. Clinicians therefore suggested that such families may not achieve as positive outcomes as other families from Group GPD-CBT in these contexts and may require higher intensity individual support.
In terms of mechanisms, the group format was described as enabling greater engagement and motivation, compared with working through the book independently or individually with the clinician. Clinicians highlighted that parents worked together to problem solve, share ideas and roleplay conversations.
Clinician 1: I think it's really interesting that a self-help book, which is very difficult to do by yourself really. . . and then you give it some life by putting it in a group and you get people talking about their own experiences and then the format of the book is much more useful.

Clinician 4:
When you're in a group, you're getting encouragement. . . you're seeing other people who are struggling, so you don't feel so isolated and sometimes that can motivate you.
The group format was also seen as having normalising effects for parents of anxious children. Moreover, clinicians described the positive empowering experience for parents of supporting each other during the treatment.
Clinician 9: I think. . .one of the most popular outcomes is the people sharing ideas and experiences.

Discussion
This is the first study to evaluate Group GPD-CBT for treating child anxiety as delivered in routine clinical practice. We found promising outcomes in terms of posttreatment service utilisation and available symptom data which are complemented by clinicians' description of Group GPD-CBT as a broadly acceptable, practical and helpful treatment for child anxiety.
The finding that 70% of children did not receive further treatment for anxiety following Group GPD-CBT suggests that it is a helpful low-intensity treatment and one which is likely to work well within a 'stepped care' model in routine practice. This proportion of children not 'stepped up' to higher intensity treatment is similar to the proportion of children in research trials of GPD-CBT reported as 'much' or 'very much' improved on the clinical global impression scale (CGI; Guy, 1976) at 6 months post-treatment (76% in Thirlwall et al., 2013;66;. The significant decline in RCADS-P scores also suggests good outcomes, albeit from a limited subset of the sample. Moreover, parental attendance levels of approximately 80% may be considered indicative of acceptability of Group GPD-CBT to parents. However, attendance was lower than has been observed in research trials (88% in Cartwright-Hatton et al., 2011, 92;.4% in Waters et al., 2009) and the reasons for this are unclear.
The analysis of clinician interviews suggests that the approach is generally acceptable to clinicians. However, clinicians varied in their attitudes towards and confidence in delivering Group GPD-CBT, and this may influence how well they felt able to 'sell' the approach to parents who may initially be sceptical about a parent group approach. Clinicians' reports that anxious parents tended to be particularly reluctant to attend a group format is consistent with findings from adult CBT literature that when given the choice, the majority of anxious patients chose individual over group CBT (Sharp, Power, & Swanson, 2004). However, parents' initial scepticism or reluctance to attend a group format was not perceived by clinicians to determine treatment outcomes. This highlights the importance of how the treatment is initially presented to parents to enable them to participate in the groups. To maximise initial participation, services could consider, for example, video vignettes showing the experience of parents who have completed the treatment. It may be particularly helpful to include vignettes from highly anxious parents who have attended, given clinicians' reports that this group may feel particularly daunted by attending a group treatment.
Notably, clinicians highlighted the potential for Group GPD-CBT to be more efficient than an individual approach as long as a sufficient number of parents attend each group. Small groups did not bring efficiency benefits given the additional time commitments of delivering GPD-CBT in a group, including allowing time for parents to gain from each other's experiences and the additional preparation time. Administration costs such as preparation time should not be overlooked, as a recent trial has found these factors to be associated with a significant difference in cost-effectiveness between two low-intensity treatments for child anxiety . Systematic evaluation of the cost-effectiveness of group versus individual GPD-CBT, and the optimal group size, is required, as highlighted in relation to CBT research more broadly (Tucker & Oei, 2007).
Clinicians also described potential benefits of group process factors on parents' motivation and engagement. Consistent with findings from parent groups for children with other difficulties (e.g. Goodier et al., 2014;Shilling et al., 2013), they described how a group format provided the opportunity for normalising, gaining peer support and sharing ideas. However, clinicians reported that the group format may not be as helpful for families with complex circumstances or children with additional needs such as ASD. Indeed, previous trials have highlighted the need to adapt CBT for anxiety in children with ASD (Wood et al., 2009), which may be more difficult to achieve in a group format which includes children with and without ASD. Furthermore, as greater attendance and homework completion has been associated with superior outcomes in CBT for anxiety more broadly (Glenn et al., 2013), it logically follows that families with complex circumstances or those undergoing stressful life events may not find the treatment as beneficial as it may practically be more difficult to achieve this level of engagement. Indeed, it has previously been suggested that families with additional needs associated with social factors may require additional support to enhance engagement in GPD-CBT (Waters et al., 2009). As such, the positive findings for outcomes of Group GPD-CBT must be understood in the context that it is a low-intensity intervention that is not necessarily suitable for families with particularly complex needs. Interestingly, however, clinicians felt able to support some families with additional needs (e.g. in cases of high parental anxiety), to attend and achieve good outcomes. These conflicting views further demonstrate a need previously expressed in the Waters et al. (2009) trial for further exploration of factors that facilitate engagement with and good outcomes from Group GPD-CBT for families with additional needs, and how outcomes may be improved for these families.
This study has a number of strengths, most notably the combination of qualitative and quantitative data, providing a more contextually informed evaluation than could be achieved by either form of data in isolation. Collection of data from routine clinical practice across a number of services is also a strength, as it indicates applicability of results across routine clinical settings. Furthermore, the use of semistructured clinician interviews has generated a range of new insights which may not have been achieved through quantitative data. The study also had a number of limitations. The study used retrospective data collected in routine clinical practice which prevented measurement of clinician, child or parent adherence to the treatment protocol. Although interviews suggest that clinicians mostly found the treatment accessible, easily to follow and reported adhering to the manual with some degree of flexibility, it is not possible to draw firm conclusions regarding adherence from these data. Indeed, this issue is complicated by the clear variation in adherence to the manual which is reflected by differences in length/structure of treatment delivery between services. The sample size was not large enough to enable meaningful analyses of differences in outcomes for children between these formats, and future research would benefit from considering this due to its implications, for example, for cost-effectiveness. It was also not possible to collect complete attendance data for 14 families because the number of sessions they attended was not consistently recorded. Furthermore, it is possible that clinicians/services that made themselves known to the researchers represented a group with particularly positive experiences of Group GPD-CBT, meaning the results may not be representative of other clinicians or NHS Trusts. A related issue is that inclusion in the sample depended on clinicians providing the researcher details of children who were not receiving concurrent anxiety treatment other than parent Group GPD-CBT and information was not available on the number or characteristics of children excluded for this reason.
In line with experiences of CAMHS nationally (Wolpert et al., 2016), a further limitation is that only a minority (24%) of cases had complete RCADS-P data at both pre and post treatment. This may be in part because some services opted only to collect RCADS-P data at initial intake and discharge, which were several months prior to/following treatment and therefore could not be included in analyses. Notably, the average total anxiety scores for the cases with complete RCADS-P data fell within the subclinical range at both time points. This is surprising given these were CAMHS samples, although it has recently been shown that even seemingly low scores on some parent-report measures of child anxiety may not be reliable indicator of recovery from a number of common child anxiety disorders (Evans et al., 2017). As no other outcome measures were available in sufficient numbers, it was not possible to corroborate the RCADS-P results with those from other measures. In addition, those with complete RCADS-P data represented a younger group of children whose parents had attended a greater proportion of sessions than those without complete RCADS-P data. As such, it is important to remain cautious when interpreting the symptom change results, and it should not be inferred that the results from this subset represent the sample as a whole. It is also necessary to emphasise that these results should not be interpreted as representing the wider population as this is beyond the remit of a service evaluation and it was not possible to collect sufficient data to achieve a representative sample.
These limitations notwithstanding, this pilot-implementation study has provided a rounded initial evaluation of Group GPD-CBT as delivered in routine child mental health services and has generated a range of useful considerations for clinical services and future research. Group GPD-CBT is associated with good outcomes and appears broadly acceptable to clinicians. Moreover, clinicians suggest that it may be more efficient (in some circumstances) and bring particular benefits associated with group process factors, if due attention is taken regarding its presentation to families. Future research is now required to systematically compare individual and Group GPD-CBT, including investigation of which children may or may not benefit from Group GPD-CBT and how best to adapt the programme for particular groups such as those with ASD or complex needs.