Adapting compassionate mind training into guided self‐help for parents of autistic children

Abstract

Objectives

Parents of autistic children (PAC) are at increased risk of psychological distress, including shame and self‐criticism. Compassion‐focused therapy (CFT) is a recommended transdiagnostic intervention for such difficulties, but to date, there is limited research around the use of CFT with PAC. The present study aimed to explore the adaptations needed for a guided self‐help CFT intervention for wellbeing, and the potential factors influencing future implementation during feasibility testing, through engagement with relevant stakeholders.

Design

A two‐stage qualitative research design aimed to (1) extract and configure stakeholder feedback for adaptations and (2) understand stakeholder views regarding the anticipated factors influencing implementation of the CFT resource during future feasibility testing.

Methods

Five iterative rounds of stakeholder feedback and intervention refinement were facilitated, using focus groups and written commentary with PAC (n = 7), clinical psychologists working within children's autism services (n = 4) and a CFT expert (n = 1). Directed content analysis supported extraction of adaptation suggestions during each feedback round. Framework analysis was employed to all focus group data to meet the second research aim.

Results

Compassionate Mind Training for Parents of Autistic Children (CMT‐PAC) guided self‐help intervention was developed. Two key concepts from the framework analysis were recognised across focus groups: ‘personal and social context of parents’ and ‘barriers and facilitators to engagement’.

Conclusions

Stakeholders appeared optimistic about the use of CMT‐PAC within a future feasibility study. However, some potential issues were raised relating to how CMT‐PAC could be implemented amongst the target population. Implications for research and clinical practice are discussed.

INTRODUCTION

Autism is a lifelong neurodevelopmental condition characterised by differences in communication, social interaction and restricted, repetitive interests or behaviours (American Psychiatric Association, 2013). Estimated prevalence rates between 2012 and 2021 report a total 1% worldwide prevalence of autism, including 0.82% in the Americas, 2.03% in the Western Pacific and 0.63% in Europe (Zeidan et al., 2022). Within the United Kingdom, there are an estimated 700,000 autistic people and three million carers (National Autistic Society, n.d.; Department of Health & Social Care & Department for Education, 2021). However, these figures may be doubled due to age‐related diagnostic inequalities, as a possible 750,000 further undiagnosed autistic people are estimated within England alone (O'Nions et al., 2023).

Parental mental health

Parenting an autistic child can be greatly rewarding, although it may present unique challenges (Khan et al., 2016). A wealth of literature evidences the prevalence of psychological distress in parents of autistic children (PAC), including higher rates of depression, anxiety, stress and lower quality of life (DesChamps et al., 2020; Hayes & Watson, 2013; Hsiao, 2016; Jellett et al., 2015; Lushin & O'Brien, 2016; Piro‐Gambetti et al., 2023; Stewart et al., 2017; Weiss et al., 2012). Parents' responses to their child's diagnosis (Legg & Tickle, 2019; Rabba et al., 2019), management of children's differences associated with key features of autism, financial and familial pressures and barriers to service support (Crane et al., 2016) are evidenced as risk factors associated with poorer parental mental health.

Despite this, there is little acknowledgement of parental mental health needs within current United Kingdom health‐related policies, such as the National Autism Strategy (Department of Health & Social Care & Department for Education, 2021) and The NHS (2019) Long Term Plan. There are no specific guidelines for mental health treatment amongst PAC. The National Institute for Health and Care Excellence (NICE, 2021b) recommends clinicians provide advice and emotional support to PAC, yet there is minimal research regarding systemic working with this population, and services are commissioned with a greater focus on child needs (NHS England, n.d.), so parents are often missed (Griffin et al., 2024). Systematic literature reviews for mental health interventions with PAC have demonstrated that parent‐focused interventions can significantly reduce stress, depressive symptoms and parental distress (Li et al., 2024). However, several included studies indicated variable quality concerns, including high risk of bias and attrition (Catalano et al., 2018; Da Paz & Wallander, 2017; Juvin et al., 2022; Kulasinghe et al., 2023; Li et al., 2024; MacKenzie & Eack, 2022). Many lacked a conceptual framework or overall conceptual coherence, and only a small number of studies were found to use evidence‐based interventions, including cognitive behaviour therapy (CBT), acceptance and commitment therapy and mindfulness‐based interventions.

Bidirectional effects incorporating poor parental mental health and child autism‐related differences can inadvertently reduce quality parent–child interactions and their attachment relationship (Fanti et al., 2013; Neece et al., 2012). Rodriguez et al. (2019) evidenced that high parenting stress positively predicted child internalising behaviours, such as social withdrawal and low mood, and a reciprocal relationship was observed between parental stress and children's externalising behaviours, such as aggression and impulsivity. Poorer parental emotional regulation skills are associated with poorer child developmental outcomes, including their social and communication skills, ability to emotionally self‐regulate and empathy development (Crowell et al., 2019).

Furthermore, autistic children are more likely to experience public stigma, such as social exclusion and victimisation from their peers (Trundle et al., 2023) or blame for appearing disruptive (Swaab et al., 2023). Subsequently, parents often experience prejudice and discrimination due to their association with a publicly stigmatised person (Kinnear et al., 2016; Rusu et al., 2024), leading to internalisation from repeated exposure to these experiences, known as affiliate stigma (Deguchi et al., 2021; Milačić‐Vidojević et al., 2014). Research indicates that PAC experience severe affiliate stigma (Liao et al., 2019; Salleh et al., 2020; Zhou et al., 2018) and subsequent shame (Burrell et al., 2017; Papadopoulos et al., 2019; Rusu et al., 2024; Salleh et al., 2020). Shame is a universal socially focused emotion arising from negative self‐evaluation in relation to societal norms (Tangney et al., 2004; Wolf et al., 2010) and is linked to defensive strategies of inhibition (Gilbert, 2019). Shame is evidenced to mediate the relationship between individuals' adverse experiences, such as affiliate stigma, and the onset and maintenance of parental mental health difficulties (Lewis, 2019).

Potential benefits of compassion‐based interventions

Definitions of compassion are grounded within Buddhist tradition, comprising two core elements: motivation to engage with suffering and actively engaging to alleviate, reduce, or prevent further suffering in response (Gilbert, 2010; The Dalai Lama & Vreeland, 2001). Evolutionary theories suggest compassion as an antidote for shame, through enabling individuals to move from shame‐related competitive motivational systems into a soothing motivational system, resulting in affiliative patterns of emotional disposition and self‐to‐self relating (Gilbert, 2019). Cross‐sectional research indicates that compassion is positively associated with greater wellbeing factors in PAC populations, including life satisfaction, quality of life, hope and goal engagement, and it is negatively associated with parental stress, depression, guilt and shame in response to challenging parenting events, and distress arising from affiliate stigma (Bohadana et al., 2019; Neff & Faso, 2015; Sirois et al., 2019; Torbet et al., 2019; Wong et al., 2016). Greater compassion levels are evidenced to predict more psychologically facilitative parenting styles (Kirby, Grzazek, & Gilbert, 2019) and greater warmth during parent–child interactions amid challenging parenting events (Miller et al., 2015).

Compassion is recognised as a modifiable trait (Neff & Germer, 2013) and is a therapeutic intervention focus for a range of mental health difficulties, especially those attributed to shame and self‐criticism, with growing empirical support (Kirby et al., 2017; Leaviss & Uttley, 2015). Many involve compassionate mind training (CMT) exercises taken from compassion‐focused therapy (CFT; Gilbert, 2009), which can be used as a standalone intervention to support individuals to cultivate compassion, to approach and alleviate distress (Gilbert & Procter, 2006; Irons & Heriot‐Maitland, 2021). Considering the high prevalence of shame experienced by PAC, which is attributed to poorer mental health outcomes, compassion‐focused interventions could be a viable approach for this population as an alternative to other evidence‐based psychological therapies, such as CBT.

Systematic reviews provide tentative evidence towards the efficacy of compassion‐based interventions used with parents to improve self‐compassion, depression, anxiety and mindfulness (Jefferson et al., 2020; Kemp et al., 2022), although they highlight the variable methodological quality of included studies. To date, published feasibility studies of compassion‐based interventions trialled specifically with PAC are largely delivered in group format and utilise approaches such as cognitively‐based compassion training (Fernandez‐Carriba et al., 2019), mindfulness‐based self‐compassion training (Bharadwaj, 2025; Cengiz & Kılıç, 2025; Rojas‐Torres et al., 2021) or CFT (Khanjani et al., 2024). Studies have demonstrated effectiveness at increasing parental self‐compassion (Ahmed & Raj, 2023; Bharadwaj, 2025; Bohadana‐Brown, 2021; Cengiz & Kılıç, 2025; Kaçan et al., 2025), perceived competence (Bharadwaj, 2025) and social support (Kaçan et al., 2025), and reducing parental stress (Fernandez‐Carriba et al., 2019; Bharadwaj, 2025; Ahmed & Raj, 2023; Bohadana‐Brown, 2021), anxiety (Bohadana‐Brown, 2021; Rojas‐Torres et al., 2021), internalised shame (Khanjani et al., 2024) and depression (Ahmed & Raj, 2023; Bharadwaj, 2025; Bohadana‐Brown, 2021). However, many studies were limited by low intervention uptake and unclear reporting regarding intervention adaptation procedures.

Kirby et al. (2023) evidenced the efficacy of a brief CFT approach with parents using a two‐hour seminar and workbook. Findings demonstrated significant reductions in parental self‐criticism and child emotional problems at two weeks post‐intervention, sustained at three‐month follow‐up, alongside further positive changes observed in parenting styles. Brief CFT interventions such as this hold promise, as they may be efficacious for enhancing PAC well‐being. However, research indicates several barriers preventing many PAC from seeking mental health support, including having limited knowledge of service accessibility, ambiguous service pathways, familial financial pressures, service inflexibility and stigma (Bonis, 2016; Braddock & Twyman, 2014; Nealy et al., 2012; Osborn et al., 2020; Wallace‐Watkin et al., 2023). Structured support via groups may be implied as inaccessible for many PAC, and it may be reasonable to assume that shame proneness may further deter PAC from accessing support (Dunford & Granger, 2017).

Self‐help interventions could provide a flexible alternative for engagement in parental wellbeing support. Use of self‐help interventions is evidenced to overcome practical and psychological barriers described above (Döpfner et al., 2021) and could provide greater scope for PAC to access timely and evidence‐based psychological treatment. One concern highlighted amongst reviews of self‐help interventions relates to high attrition rates (Hermes et al., 2019). The addition of guided support and choice of interaction for this element is noted to positively influence retention (Day & Sanders, 2018; Wojtowicz et al., 2013). Guided self‐help interventions involve an individual working through a self‐help workbook or online course, with the support of a therapist (NHS, 2022). Systematic reviews evidence that guided self‐help can have comparable effects to face‐to‐face psychotherapy for anxiety and depression (Cuijpers et al., 2010; Moshe et al., 2021), although the modalities of included studies were dominated by CBT. Furthermore, there is tentative evidence that mindfulness and acceptance can be effectively cultivated via self‐help interventions (Cavanagh et al., 2014; Taylor et al., 2021). Currently, the vast majority of CFT randomised controlled trials are facilitated as group interventions rather than guided and unguided self‐help, although there is evidence that brief, self‐help compassion interventions can show promise when used within complex clinical populations (Craig et al., 2020). There is a need to further this evidence base to develop preventative mental health interventions that can become accessible to the broader population.

Study rationale

Given that there is a greater prevalence of psychological distress, shame and stigma within PAC, there is a need to further the evidence base for accessible interventions for this population. Initial evidence for compassion‐based interventions with PAC holds promise, although such interventions were largely tested in group format to date. The creation of a PAC‐focused CFT intervention using guided self‐help methods could provide PAC with greater flexibility in how and when they engage with the intervention. Furthermore, this intervention format may be more palatable for PAC who experience high levels of shame, self‐criticism, or perceived stigmatisation, as it is reasonable to assume that these factors may further deter PAC from seeking mental health support.

The existing CFT workbook for parents (Kirby et al., 2023) was provided to the authors with permission to adapt for use with PAC. Enhancing intervention‐context fit through the adaptation of evidence‐based interventions to increase specificity regarding recipients' cultural backgrounds (e.g., parenting an autistic child) can enhance recipient engagement, acceptance and satisfaction regarding an intervention, which may enhance outcomes (WHO, 2024). The original workbook was used for self‐directed study following attendance at the two‐hour seminar and lacked specificity regarding the unique differences and challenges when parenting an autistic child. The present study therefore aims to systematically adapt this workbook to create a standalone six‐week PAC‐specific guided self‐help intervention, where each week would involve parents reading a workbook chapter, engaging in daily CMT experiential practice, and attending a 15‐min support call with a CMT‐trained clinician to provide an opportunity to discuss the resource and problem solve any difficulties with implementation. This intervention would aim to primarily increase PAC self‐compassion, which may positively influence a range of PAC and child well‐being outcomes.

Following best practice guidelines for intervention development and adaptation, it is imperative that initial adaptation phases are duly considered (Moore et al., 2021; Skivington et al., 2021b). This phase aims to minimise potential harm to future study participants arising from issues such as mismatch between intervention materials and recipients, inaccessibility, or overprescription (Papworth et al., 2015) to mitigate unforeseen factors that may limit intervention effectiveness. A stakeholder‐informed adaptation process was warranted in the present study, particularly as the planned intervention format involves reliance on PAC engagement through self‐directed practice. It was important that relevant stakeholders, including PAC, were consulted at all stages of intervention development to enhance the anticipated relevance, usability and acceptability of the intervention in future practice.

Research aims

To explore adaptations needed to create a guided self‐help CMT intervention with PAC, and to define anticipated factors influencing successful implementation during future feasibility testing.

Research questions

• How can a CFT resource be adapted into a guided self‐help intervention for PAC, whilst maintaining theoretical coherence, for further investigation during feasibility testing?

• What are stakeholder views regarding anticipated factors influencing successful implementation of the CFT resource during future feasibility testing?

MATERIALS AND METHODS

A qualitative research design was undertaken from a pragmatist epistemological position. The study received favourable ethical opinion from the University of Nottingham Mental Health and Clinical Neurosciences ethics committee (reference: 3003).

Participants

Participants involved PAC and clinical psychologists (CPs) within the United Kingdom. PAC were aged 18 or over, could communicate using the English language, and had a children aged under 18 years with a current diagnosis of autism. CPs were registered with the Health and Care Professions Council and were employed within children's autism services, thus regularly provided psychological support to PAC. All participants required technology to engage in virtual focus groups via Microsoft Teams. Written feedback from international CFT clinicians was later requested, who had expertise around the theoretical approach and its application with parents.

Guidelines for focus group size are varied, although typically 4–12 participants per group is recommended (Kitzinger, 1995; Krueger & Casey, 2015; Stewart & Shamdasani, 2015). Smaller focus groups with as low as three participants are favoured by some (Barbour & Kitzinger, 1999), as these may be advantageous when used with participants with expertise or additional needs to ensure that each has enough time to express their views (Barrett & Kirk, 2000; Gates & Waight, 2007; Morgan, 1995). The target sample size of the present study was five PAC, as it was felt that this would support balance between promoting diversity of experiences within the focus group discussion, whilst ensuring that all participants had sufficient time to express their views about the intervention. However, only three of five PAC attended the first focus group due to participant drop‐out, so further recruitment efforts were warranted to ensure that five could attend the second focus group. The recruitment strategy for CPs was broadened to include four to eight participants, to support the balance of maximising the time available for all participants to share their expertise during the focus group, with the ability to recruit more people to minimise the impact of participant drop‐out. Seven PAC, four CPs and one CFT expert were recruited in total. All except one PAC provided demographic information, summarised in Table 1. Three PAC disclosed they were neurodivergent within the focus groups.

TABLE 1.

Sample characteristics.

Factor	Parents, n (%)	Clinical psychologists, n (%)
Age
20–29 years	1 (14.3)	0 (0)
30–39 years	2 (28.6)	1 (25)
40–49 years	2 (28.6)	3 (75)
50–59 years	1 (14.3)	0 (0)
Gender
Female	6 (85.7)	4 (100)
Ethnicity
Asian or Asian British	1 (14.3)	0 (0)
Black or Black British	1 (14.3)	0 (0)
White or White British	4 (57.1)	4 (100)
Marital status a
Single	1 (14.3)	–
Married/civil partner	4 (57.1)	–
Divorced/partnership dissolved	1 (14.3)	–
Age of child a
Up to 5 years	1 (14.3)	–
6–10 years	2 (28.6)	–
11–15 years	2 (28.6)	–
15–17 years	1 (14.3)	–
Highest level of professional qualification b
Doctorate in clinical psychology	–	3 (75)
Post‐doctoral training	–	1 (25)
Compassion‐focused therapy trained? b
Yes	–	4 (100)
Years' experience working with parent of autistic children? b
4 years	–	1 (25)
10 years	–	1 (25)
14 years	–	2 (50)

^a Question for PAC only.

^b Question for CPs only.

Recruitment

PAC were recruited via an advert circulated through social media and within regional and national parent support groups, autism charities and organisations within the United Kingdom. Potential participants expressed their interest to the lead researcher via email and were sent a Focus Group Guidelines document with a Participant Information Sheet detailing ethical considerations. All were offered a screening meeting to provide an opportunity to ask questions about the research. PAC were asked to complete a Consent Form in Microsoft Word format and email this to the lead researcher to opt into the project. PAC were offered a £10 Amazon gift voucher for each focus group attended as compensation for their time.

CPs were recruited through professional networking and snowball sampling and experienced the same recruitment process as PAC. Alternative versions of the Participant Information Sheet and Consent Form were shared with CPs. CFT experts were contacted directly through email by the lead researcher. CPs and CFT experts were not compensated for their time.

Procedure

Phase One

The existing CMT workbook was sent via email to PAC for individual review, alongside a demographic questionnaire. Individual review was chosen to reflect future use of the intervention. One week later, PAC participated in a ninety‐minute audio‐recorded virtual focus group meeting moderated by the lead researcher. The field supervisor attended to record observations about group processes, including non‐verbal indicators. PAC provided feedback on the workbook materials for future feasibility testing and to share recommendations for further adaptation. Feedback influenced the nature of intervention refinement.

Phase Two

PAC were sent the updated CMT workbook and invited to a second audio‐recorded ninety‐minute virtual focus group with the same facilitators to gain a majority agreement on changes made and to gain further feedback and recommendations for adaptation. All were provided a debrief letter following the second focus group. The resource was subsequently refined.

Phase Three

The resource was sent to CPs alongside a demographic questionnaire, with a request for written feedback within a four‐week period. An email reminder was sent seven days before the feedback deadline. Written feedback was collated to inform further intervention refinements, and the updated resource was returned to CPs. Any conflicting feedback was noted and anonymised.

Phase Four

A subsequent two‐hour audio‐recorded virtual focus group was facilitated with all CPs to gain a consensus view on each conflicting recommended adaptation and for further discussion around perceived advantages and disadvantages of the proposed intervention, the applicability of CMT to PAC, and to obtain opinions regarding whether theoretical coherence was maintained following adaptation. The intervention was refined.

Phase Five

The resource was sent to international CFT experts via email for commentary around whether theoretical coherence was maintained following adaptation, and for their opinions regarding the applicability of CMT with PAC. All participants who opted to receive research updates were emailed the updated resource with an optional invitation to provide final written comments to ensure that the updates from phases three to five were accepted by PAC.

Analysis

The present study design involved two processes: the extraction of stakeholder‐suggested amendments at each feedback phase to support intervention refinement, and broader analysis across focus groups to identify and describe key views regarding the anticipated factors influencing successful implementation of the intervention during future feasibility testing. Several stages of analysis occurred sequentially throughout the research.

Extracting suggested amendments

Focus group recordings were transcribed verbatim by the lead researcher. Deductive directed content analysis (DDCA) was chosen to answer the first research question, as it can be used deductively to compare the frequency of items in categories at different periods (Elo & Kyngäs, 2008; Vaismoradi et al., 2013). This appealed as it enabled the categorisation of amendments and comparisons between the types of suggested amendments received during each feedback phase. DDCA has been used as a rapid coding process (Neal et al., 2015) for similar studies to provide a time‐efficient methodology for extraction and enactment of stakeholder feedback (Ametaj et al., 2021; Radovic et al., 2017), which was felt to be valuable when considering the constraints and available resources throughout the research timeline. Rapid assessment for suggested amendments using DDCA was facilitated by the lead researcher following each study phase. A unit of meaning was defined as one or more sentences that clearly identify either a negative reaction towards an aspect of the proposed intervention, or when a suggestion for amendment was made. A unit of analysis was defined as a study phase (e.g., a whole focus group). To support data extraction, a categorisation matrix was developed from FRAME (Stirman et al., 2013, 2019) to systematically code relevant content which corresponded with the identified categories within the matrix. NVIVO14 was used to code the data according to structured categories. As the context of information coded was relevant, thematic units were used rather than line‐by‐line coding, as this would have risked fragmenting amendment requests. All coded data were exported to a Microsoft Excel spreadsheet for further consideration.

Adaptation protocol

During each refinement phase, the lead researcher reviewed all extractions from the DDCA. Wherever negative feedback was not already linked to a specific amendment suggestion, the lead researcher pragmatically inferred what amendment could reasonably rectify the negative feedback by reviewing the raw transcript and discussing in research supervision. To uphold intervention fidelity throughout the adaptation process, the lead researcher then reviewed each amendment suggestion to consider appropriateness, including how this may potentially influence intervention outcomes. The MADI model (Kirk et al., 2020) supported this by proposing four questions to consider possible mediating or moderating factors of intervention adaptation characteristics on potential outcomes.

1. Would the adaptation be consistent with the core functions of the intervention or future implementation strategy?

2. Is the adaptation being made for a reason or goal to address fit?

3. What could be the impact on outcomes, considering available data, theory, stakeholder opinions and best practice?

4. Is the adaptation being made due to an anticipated obstacle? (Kirk et al., 2020, p. 8)

Wherever a consensus opinion regarding an amendment suggestion was not achieved within or between stakeholder groups, suggestions were discussed within the research team to determine the appropriateness of implementing one or multiple conflicting requests, and a majority decision was sought.

Categorising adaptation types

Once all study phases were complete, a taxonomic analysis (Holtrop et al., 2022) was performed on all intervention amendments that had been facilitated throughout the study to group goals and the nature of adaptations into clustered ‘adaptation types’. This involved calculating frequencies for each adaptation component combination during each phase of refinement. This purely functioned to enhance reporting of reasons for adaptations and is hoped to enrich future investigations regarding the intended and actual impact of adaptations during feasibility testing.

Defining anticipated factors influencing successful implementation

Once all study phases were complete, a framework analysis (FA; Ritchie & Spencer, 1994) was employed to all focus group transcripts to define concepts relevant to the second research question that arose between stakeholder groups. FA was chosen as it is a flexible tool that is not associated with any epistemological paradigm and provides a systematic analytical approach for time‐limited research objectives (Gale et al., 2013). FA enables the researcher to move between stages of raw data, abstraction and interpretation to foster transparency and replicability during data management and interpretative processes (Ritchie et al., 2014; Smith & Firth, 2011). The indexing process followed an inductive‐deductive approach, where key themes were already developed through a priori knowledge that had influenced interview schedule development. However, data were coded inductively if they did not fit pre‐determined codes yet appeared to answer the second research question, and so framework matrices were subsequently refined.

Quality considerations

The lead researcher used a reflexive log and research supervision to consider the influence of personal experiences and interpretations throughout the research process. During the FA indexing phase, the second author checked 20 % of a coded transcript for plausibility of inductive coding. Charting and interpretation processes were discussed within supervision.

Inter‐rater reliability assessments were conducted during the DDCA, supported by an independent doctoral researcher. Twenty percent of a focus group transcript was separately coded by the doctoral researcher for presence or absence of a suggested adaptation per participant comment. Percentage agreement and Cohen's (1960, 1968) Kappa coefficients indicated substantial to excellent agreement (90.4% agreement, κ = 0.79; Landis & Koch, 1977). Twenty percent of amendment suggestions across all phases were separately coded by both nature and goal for adaptation. A pooled Kappa estimator (De Vries et al., 2008) indicated excellent agreement between coders' judgments (97% agreement, κ _pooled = 0.94).

RESULTS

Intervention adaptation

Adaptation occurred iteratively over four phases, as no amendments were suggested during phase five. Three PAC attended focus group one (FG1) and a fourth did not attend due to technological difficulties, so they provided brief written feedback. Five PAC attended focus group two (FG2), including two from FG1 and three new participants. All CPs provided written feedback and attended focus group three (FG3), although one arrived late and thus did not contribute to all consensus‐making discussions. Table 2 provides an overview regarding the nature of adaptations made per feedback phase.

TABLE 2.

Changes made according to the FRAME model by phase.

Process	Phase
	1	2	3	4
When did modifications occur?
Preimplementation/planning/pilot	X	X	X	X
Were adaptations planned?
Planned/proactive (proactive adaptation)	X	X	X	X
Who participated in the decision to modify?
Lead researcher	X	X	X	X
Research supervisors	X			X
Parent stakeholders	X	X
Clinical Psychologists working within children's autism services			X	X
What is modified?
Content	X	X	X	X
Contextual	X
At what level of delivery for whom the modification is made?
Target intervention group	X	X	X	X
Cohort/Individuals that share a particular characteristic	X	X	X	X
Practitioners				X
Contextual modifications are made to which of the following?
Format	X
Personnel	X
Population	X
What is the nature of the content modification?
Adding elements	X	X	X	X
Lengthening/extending (pacing/timing)			X
Loosening structure		X		X
Removing/skipping elements	X
Reordering of intervention modules or segments	X			X
Repeating elements	X
Shortening/condensing (pacing/timing)	X	X
Substituting	X
Breaking up session content over multiple sessions	X			X
Tailoring/tweaking/refining	X	X	X	X
What was the goal?
Increase reach or engagement	X	X	X	X
Increase retention	X
Improve feasibility	X	X	X	X
Improve fit with recipients	X	X	X	X
To address cultural factors	X		X	X
Improve effectiveness/outcomes	X	X	X
Increase satisfaction		X	X
Reduce cost	X
Relationship fidelity/core elements?
Fidelity consistent/core elements or functions preserved	X	X	X	X
Reasons
Access to resource	X
Available resources	X
Cognitive capacity	X	X	X
Comorbidity/multimorbidity	X	X	X
Crisis or emergent circumstances	X	X		X
Cultural norms	X		X
Literacy and education level	X	X	X	X
Motivation and readiness	X	X	X	X
Perception of intervention	X
Service structure	X

Deductive directed content analysis

Following the de‐duplication of the amendment suggestions extracted, a total of 107 of 164 suggestions were implemented across four feedback phases. Twenty‐seven suggestions were not implemented due to being already present or deemed inappropriate, as these contradicted core components of CMT. Thirty suggested adaptations were not implemented yet due to restrictions on available resources, but were noted for future consideration (see Appendices S1 and S2).

PAC feedback in FG1 regarding the original resource was largely negative. Criticisms related to perceived high intervention burden due to large amounts of text and complex language to read, lack of clarity around purpose and coherence of the intervention, and lack of specificity to parenting an autistic child. Consequently, the research team chose to rewrite the workbook material, as this was more time efficient than adapting the resource so significantly. Compassionate Mind Training for Parents of Autistic Children (CMT‐PAC; see Appendix S3) was created as a six‐week intervention with accompanying audio tracks for daily practice. It incorporated more diagrams, text boxes, colour‐coding and autism‐specific examples recommended during FG1. PAC feedback within FG2 was subsequently largely positive, and further amendments involved use of more examples to increase clarity of psychoeducation components and minor alterations to colours and formatting.

CP amendment suggestions in phase three largely involved making language informal and breaking some psychoeducation elements into smaller sections. Character stories and reflective spaces were also added to strengthen future recipients' understanding and practice. Following CP feedback in FG3, final adaptations included minor format changes; reordering chapters one and two to introduce challenges in parenting prior to the theoretical background of CMT; addition of a compassionate object exercise in chapter six; and splitting of a compassionate imagery exercise into shortened and extended versions to enhance accessibility.

Taxonomic analysis

The taxonomic analysis provided further context regarding reasons why adaptations were made per feedback phase, as shown in Table 3. This revealed that only 30 of 91 possible combinations for adaptation types were undertaken, where most were categorised as tailoring, tweaking, or refining the intervention (n = 86). Reasons for adaptation were most commonly to increase reach and engagement (n = 25), to increase fit with recipients (n = 37) and to improve feasibility (n = 15).

TABLE 3.

Why did adaptations occur during each phase?

What was adapted?	Reach/engagement	Increase retention	Improve feasibility	Fit with recipients	Cultural factors	Effectiveness/outcomes	Reduce cost	Increase satisfaction
Adding elements a (n = 13)	0:0:0:0	1:0:0:0	0:0:1:0	0:0:1:1	0:0:1:0	5:1:2:0	0:0:0:0	0:0:0:0
Lengthening/extending a (n = 1)	0:0:0:0	0:0:0:0	0:0:0:0	0:0:1:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Loosening structure a (n = 2)	0:0:0:0	0:0:0:0	0:1:0:1	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Removing elements a (n = 2)	0:0:0:0	0:0:0:0	0:0:0:0	1:0:0:0	0:0:0:0	0:0:0:0	1:0:0:0	0:0:0:0
Re‐ordering modules a (n = 3)	0:0:0:0	0:0:0:0	0:0:0:0	1:0:0:0	0:0:0:1	1:0:0:0	0:0:0:0	0:0:0:0
Repeating elements a (n = 1)	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	1:0:0:0	0:0:0:0	0:0:0:0
Shortening/condensing a (n = 4)	0:0:0:0	0:0:0:0	3:1:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Spreading content a (n = 2)	0:0:0:0	0:0:0:0	0:0:0:1	0:0:0:0	0:0:0:0	1:0:0:0	0:0:0:0	0:0:0:0
Substituting content a (n = 1)	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	1:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Tailoring content a (n = 86)	5:4:12:1	4:0:0:0	3:1:1:0	4:6:18:3	5:0:1:0	0:0:3:0	0:0:0:0	0:9:6:0
Intervention format b (n = 3)	1:0:0:0	0:0:0:0	1:0:0:0	1:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Intervention personnel b (n = 1)	1:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Target population b (n = 2)	1:0:0:0	0:0:0:0	1:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0	0:0:0:0
Total adaptations by phase:	8:4:12:1	5:0:0:0	8:3:2:2	7:6:20:4	6:0:2:1	8:1:5:0	1:0:0:0	0:9:6:0

^a Content modification.

^b Contextual modification.

Anticipated factors influencing successful implementation of CMT‐PAC during future testing

The FA involved all focus group data. Table 4 shows the development of two key concepts as a result.

TABLE 4.

Key concepts developed from initial categories and themes.

Initial categories	Final themes	Key concepts
Challenges in parenting Parental constraints Parents' concerns and priorities	Role as a parent	Personal and social context of parents
Parental mental health Negative self‐evaluation Prevalence of trauma histories in parents	Parental mental health
Bureaucracy in services Gaps in services Potential value of CMT in services Diagnosis as the gateway to support	Services around the parent
General accessibility considerations Applicability constraints Fluctuating engagement dynamics Perceived feasibility of intervention Importance of accessibility options Value of guided support Parents' learning support needs	Accessibility and learning needs	Facilitators and barriers
Influence of prior knowledge Understanding of the CMT process Preference for autism specific Perceived target audience	Knowledge and understanding
Perceptions around compassion Needing to feel understood Activating the soothing system Attitude towards intervention tone	Psychological needs

Personal and social context of parents

This concept focused on PAC's experiences of navigating parenthood and services to share insight into how future recipients' contexts may influence their motivation and ability to engage with CMT‐PAC. It also provided participants' views around the potential value of CMT‐PAC. It comprised three themes, described below with accompanying quotes.

Role as a parent

Groups highlighted various parenting demands that could impact on PAC engagement with a psychological intervention, including childcare commitments, difficulties maintaining a usual routine and likelihood of unexpected interruptions. Parents within FG1 shared having limited time and energy and considered strategies to use CMT‐PAC alongside fulfilling their parenting duties, such as through listening to an audiobook version or trying CMT exercises with their child, if appropriate. CPs recognised the likelihood that parents' engagement will fluctuate due to pressures associated with parenting.

you've got children with autism; it is difficult. Mo (parent, FG1)

there will be other times when life gets in the way and you bench it […] it will become something that sits on the kitchen table that you might dip in and out of, and forget about, then come back to. Shona (CP, FG3)

Furthermore, parents in FG1 recognised their reasons to seek CMT‐PAC would be likely related to wanting answers or a ‘fix’ to manage their stressful situational context, which is not necessarily provided by CMT‐PAC. One parent shared their prioritisation to help their children over themselves. Therefore, it was felt that the potential mismatch between PAC motivators versus aims of CMT‐PAC could lead to disengagement by some. However, CPs reflected that it is often parents that would likely benefit from intervention, rather than their children.

it's not telling me what I need to do in a certain situation, or how I'm gonna […] combat this. Jane (parent, FG1)

how many times […] that the reason that a referral is coming in isn't because we can suddenly wave a magic wand and make a child less autistic, it's because parents are just overwhelmed. Shona (CP, FG3)

Parental mental health

All groups highlighted poorer mental health outcomes in PAC, acknowledging a need for intervention whilst also recognising that some PAC may lack headspace to engage when reactively help‐seeking under significant stress. Additionally, CPs considered possibilities of PAC experiencing trauma, and how self‐compassion could subsequently be experienced as uncomfortable or ‘triggering’ by some in this context. CPs highlighted that CMT‐PAC remained appropriate in these circumstances but advised that facilitators consider the possibility of unexpected distress by PAC during CMT practices and advocated for guided support to manage this. Finally, negative self‐evaluation was evident in both PAC groups, where they described shame and self‐criticism in relation to their parenting approaches, including their beliefs around being able to complete aspects of CMT‐PAC as prescribed. This evidences potential need for CMT‐PAC but indicates a possible factor influencing future recipient dropout.

that's quite common I think that the parents we see, they've often experienced some really tricky stuff. Jan (CP, FG3)

there's a sentence that says, ‘this is not your fault, this is not your fault’ and I found that, you know, no one ever really says that to you, so, and you do blame yourself for everything! Josie (parent, FG2)

Services around the parent

PAC shared their negative experiences with mental health services, including difficulties obtaining a referral for their child's autism assessments and long waiting times to be seen. They highlighted their belief that accessing service support was contingent on their child receiving an autism diagnosis. Similarly, CPs acknowledged that CMT‐PAC could be a valuable waitlist intervention, as the absence of a child's diagnosis should not prevent access to preventative care.

that seems to be the experience of many parents that I've spoken to […] years battling and fighting. Jane (parent, FG1)

I thought the same about the […] waiting list stuff and maybe pre‐diagnosis, that it could be usefu. Tara (CP, FG3)

Furthermore, parents shared their experiences around a perceived lack of post‐diagnostic support, which was felt to have been warranted. Similarly, CPs recognised structural gaps in services due to clinical thresholds between primary and secondary mental health services, a lack of variety in therapeutic modalities offered by services, and a lack of commissioning for parent focussed work, which they felt required addressing. These factors evidence potential value of preventative interventions for this population, such as CMT‐PAC, as they could be increasingly sought and accessed by PAC as an alternative to seeking service involvement, particularly by parents with negative service experiences. However, one parent shared concerns that CMT‐PAC may become offered by services as a box‐ticking exercise before direct work is considered and advised that developers consider how the CMT‐PAC is sold to future recipients.

I think it's a really important bit of work that's often missing from services because we're not really commissioned to do parent work, but actually, that's the work that often needs to be done. Jan (CP, FG3)

I was just trying to rack my brains […] about whether I've seen anything else, but I haven't. Polly (CP, FG3)

I'd be interested in using it, if it wasn't sold as a replacement to something else or would push you down the waiting list. Rachel (parent, FG2)

Facilitators and barriers

This concept captured anticipated factors underpinning stakeholders' adaptation requests to make CMT‐PAC more successfully implementable during future studies. It comprised three themes, described below with accompanying quotes.

Accessibility and learning needs

Both PAC and CP groups shared that guided support would increase intervention uptake by PAC compared to self‐help, as this would provide an accountability check‐in space and troubleshooting if needed. All groups considered ways to generally support recipients' learning needs through the use of colour, visuals, and reducing the amount of text. All groups highlighted the likelihood that some PAC will be neurodivergent themselves, so further considerations were made around making instructions more explicit, providing choice in communication preferences for guided support elements, and providing troubleshooting information for those with interoception differences, or when imagery or body‐based exercises may be harder to access. It was recognised that no intervention can be made truly universally accessible, but that it was important to make the original CMT resource accessible to more.

parents of autistic children don't need self‐help; they need help. Rachel (parent, FG1)

sometimes you have to be careful with abstract stuff or too much imagery, […] especially if we're wondering if some of the parents might also be neurodivergent. Tara (CP, FG3)

when you're looking for an intervention that will go to the masses, there's always going to be some groups of people that don't have that sort of way of thinking or different hurdles for different people. Shona (CP, FG3)

Furthermore, all groups highlighted that parents' engagement with CMT‐PAC would likely fluctuate, so recommendations included validation of PAC's efforts and advocacy for flexible use of the tool to increase PAC's confidence to return to the intervention. Groups emphasised the importance of providing options and choice to flexibly engage with CMT‐PAC through different intervention and guided support formats, and type and length of exercises to meet a broader range of recipients' preferences and learning needs. During FG2, all PAC anticipated that CMT‐PAC could be feasible for future study participants to complete.

it's got to be really flexible, I would say. Rachel (parent, FG1)

it's a bit like when you do a yoga class and there're different options depending on where your body's at. It's your compassion muscle I guess that you're testing. Jan (CP, FG3)

I think it is definitely do‐able to do it week‐by‐week. Josie (parent, FG2)

Knowledge and understanding

PAC in FG1 requested increased clarity around purpose and rationale for intervention components, to support recipients' informed decision‐making to engage with CMT‐PAC. For example, one parent highlighted that psychoeducation elements in the original resource, such as the ‘tricky brain’, appeared irrelevant as they viewed only experiential elements as being “the therapy” rather than the whole intervention.

you're not learning about what compassionate theory is for your research project, you're doing the therapy, so I think that's what it needs to focus on, doing the therapy with the option of learning more about it alongside. Um, at the moment it's focused on the theoretical aspect of it, with the actual therapy on the side. Rachel (parent, FG1)

All groups emphasised value and need for autism‐specific interventions that would likely match PAC's experiential knowledge. PAC in FG1 shared negative emotional responses towards general content in the original resource that was perceived to not apply to their own children, such as developmental milestones of typically‐developing children. CPs later commended the autism‐specific nature of the updated CMT‐PAC and highlighted its potentially protective nature for parental mental health, arising from PAC becoming connected with the ‘autism identity’.

I want this to be talking to me as a parent, and me […] as a parent of an autistic child, and there was nothing really in there that was talking to me in that sense; it was very, very general. Jane (parent, FG1)

having ASD in the intro, or you know the title, I think that might be quite important in some ways because there's a lot of that, sort of, autism identity, and that in itself has, um, a massive protective factor for mental health. Shona (CP, FG3)

Both stakeholder groups felt that potential recipients would require a level of education to access the resource, although guided support elements were anticipated to increase engagement with a more diverse range of PAC with varied educational backgrounds. One PAC shared their prior interest around mindfulness approaches and felt that the original resource may not be helpful for those with a good understanding of these approaches already, or where PAC may exhibit learning needs around their child's autism‐related support needs.

when you've got a child that's got sensory issues, if you're maybe at a point where you don't understand those sensory issues, trying to apply the principles of compassion still doesn't cut it, because you're not knowing why they're still having a meltdown, so you, for all the compassion in the world, if you don't understand that side, this still probably wouldn't be very helpful. Jane (parent, FG1)

I think yeah for people that are educated […] say people who are like reading a lot, they would like it. Mo (parent, FG1)

Psychological needs

Parents shared varying prior knowledge about the concept of compassion, including being unaware or less experienced with self‐compassion, and held varying beliefs around the value of a compassionate mindset. Similarly, CPs shared clinical experiences where PAC's self‐compassion appeared blocked or harder to tolerate but described optimism about using CMT with this population and highlighted potential well‐being benefits for PAC and their children. Furthermore, all groups highlighted the importance of modelling compassion to PAC to increase retention during future studies. Participants shared strategies to model compassion through upholding a compassionate tone within written and guided components, and normalisation of stumbling blocks to de‐shame PAC's experiences, should aspects of CMT‐PAC seem difficult to complete.

we often focus on being compassionate towards other people but in all honesty, before I started reading this, it wasn't something that I'd thought about being a self thing. Josie (parent, FG2)

it isn't something that you're just going to get straight away, and […] things aren't going to be perfect; it's going to take time and there will be stumbling points and […] part of that is about saying to people ‘that's okay’ and […] it's a learning curve. Tara (CP, FG3)

Finally, all groups discussed value attached to PAC perceiving that they are truly understood by the intervention developers, basing this on tone and content of intervention materials. PAC felt this would increase potential acceptability for future study participants. CPs shared recommendations around further use of common themes experienced by PAC from the beginning of the workbook to enable parents to feel understood from the outset.

it sounded like you were like, putting yourself in a parent's shoes and you have that understanding, and I thought it was really, really well written, yeah, and I enjoyed it. Olivia (parent, FG2)

Theoretical consistency and final comments

All CPs during phase four shared that CMT‐PAC remained theoretically consistent post‐adaptation. During phase five, written commentaries were requested from two CFT experts regarding whether theoretical coherence was maintained. Copies were redistributed to four PAC participants with an optional invitation to provide further feedback, but none was received. Feedback was received from one CFT clinician, who held clinical and research expertise regarding use of CFT with parents:

This workbook is a wonderful resource for parents of autistic children. All the material and exercises are consistent with the theory and models that underpin Compassion Focused Therapy. […] This CMT workbook is highly applicable and likely to be very helpful for parents of autistic children. CFT expert

DISCUSSION

This study aimed to explore adaptations needed to use a CFT resource as a guided self‐help intervention with PAC and define anticipated factors influencing successful implementation during future feasibility testing. The original resource was deemed inappropriate for PAC by key stakeholders and required significant adaptation, largely aiming to increase feasibility, fit and engagement. Stakeholders described optimism about the final CMT‐PAC, and CFT clinicians deemed that theoretical coherence was maintained post‐adaptation. Several facilitators, barriers and contextual factors are anticipated to influence the implementation of CMT‐PAC. Major findings recognised broadly across analyses are discussed further, and several clinical and research implications are provided.

Compassion and the therapeutic relationship

PAC highlighted several key facilitators for engagement, including the wish to feel understood by intervention developers, perceived through validation and an ‘affirming’ rather than ‘instructional’ tone. These preferences may relate to key attributes of compassion, such as nonjudgement and empathy (Gilbert, 2009), highlighting potential value from modelling a compassionate stance on intervention outcomes. Self‐determination theory (Deci & Ryan, 1980, 1985, 2012) might further explain motivational processes from using a compassionate tone, where an autonomy‐supportive approach could be anticipated to increase recipients' motivation and perceived competence to use the resource effectively within their context. Furthermore, the shared preference for guided support may be an integral method to promote recipients' autonomy and competence to foster ongoing engagement, motivation and value of compassion cultivation (Ryan & Deci, 2022).

Importantly, this finding reveals the anticipated role of the therapeutic relationship between PAC and the CMT‐PAC workbook as a key facilitator for engagement, which has important clinical implications related to how CFT interventions are adapted and delivered with this population. Common factors in psychotherapy (Rosenzweig, 1936) such as flexibility, responsiveness and therapeutic alliance were recognised by stakeholders as crucial aspects anticipated to increase the effectiveness of CMT (Petrocchi et al., 2024) and should be interwoven into CFT‐guided self‐help materials. Richardson et al. (2010) previously hypothesised that a lack of common factors within non‐guided self‐help would impair intervention performance, and the present findings build on this by evidencing this need expressed by target intervention recipients.

Fear of compassion

Stakeholders described potential adverse effects from CMT exercises, including fears of failure or future recipients experiencing compassion as threatening or aversive. Fears of compassion (Gilbert et al., 2011) can be marked within highly self‐critical groups (Kirby, Day, & Sagar, 2019; Merritt & Purdon, 2020), particularly if people have experienced traumatic, abusive or neglectful backgrounds (Gilbert & Procter, 2006; Matos et al., 2017; Winders et al., 2020). Although high levels of self‐criticism are well‐evidenced within PAC populations (Bohadana et al., 2019; Sirois et al., 2019), no research publications have yet identified fears of compassion within PAC. Bohadana et al. (2019) also identified barriers to self‐compassion amongst PAC, although interpreted these as mindfulness difficulties and competing priorities, which may also be attributed to PAC fears of failure around completing CMT‐PAC as directed.

Fear of compassion could decrease CMT‐PAC engagement and retention if not primarily addressed. Techniques to approach and work through an individual's fears and resistance towards compassion can provide an important therapeutic effect (Gilbert, 2010). Therefore, a clinical implication concerns incorporating fears of compassion into screening procedures for PAC accessing CFT interventions to inform whether further intervention is required, echoing Merritt and Purdon (2020). This is particularly relevant for interventions that are largely self‐directed and could be a focus within guided support opportunities.

Intervention‐context fit

In accordance with previous research, stakeholders emphasised poor PAC mental health outcomes, high competing parenting demands, gaps in service delivery and barriers to accessing psychological support (Bohadana et al., 2019; Osborn et al., 2020; Schnabel et al., 2020; Wallace‐Watkin et al., 2023). All PAC attending FG2 shared that CMT‐PAC appeared feasible and described interest in using the materials, highlighting the potential value of guided self‐help resources to increase PAC access to preventative psychological interventions, in line with broader initiatives recognised in the NHS (2019) Long Term Plan.

Flexible engagement options could be viewed as another autonomy‐supportive approach, to enhance PAC alliance with the intervention and its fit alongside contextual factors. In contrast, close attention to implementation fidelity is often seen as a requirement when conducting behavioural‐intervention efficacy research, to decrease presence of confounding variables that may influence evaluation of treatment integrity (Johnson & Remien, 2003; Persch & Page, 2013). However, concerns identified by stakeholders in the present study are reflected elsewhere, with arguments that rigid psychological therapy protocols may not be suitable for complex client presentations in community settings (Addis et al., 1999; Mazzucchelli & Sanders, 2010). Furthermore, systematic reviews have recognised that therapist adherence and manualised treatment are not empirically supported as more effective than variable adherence or non‐manualised treatment, across a range of presentations including anxiety, depression and trauma‐related, and use of interventions including CBT, family therapy and attachment‐based therapy (Truijens et al., 2019; Webb et al., 2010). There is currently no empirical evidence regarding the relationship between adherence to CFT protocol and treatment outcomes, likely as the first practice manual was published in 2024 (Petrocchi et al., 2024).

Diverse learning needs

Increased prevalence of parental neurodivergence was identified by all stakeholder groups. Despite increasing evidence that PAC often exhibit sub‐diagnostic threshold autistic traits, defined as the ‘broader autism phenotype’ (Pruitt et al., 2018; Rubenstein & Chawla, 2018; Wheelwright et al., 2010), and evidence suggesting autism aetiology is partly genetic (Warrier & Baron‐Cohen, 2017), there is a lack of research regarding autistic parents (Fletcher‐Randle, 2022; Pohl et al., 2020), and no prevalence estimates for autistic PAC. Regardless, the recognised prevalence of autistic traits within PAC held important clinical implications regarding increasing the accessibility of CFT interventions for PAC, and subsequently influenced the nature of many adaptation suggestions related to broader ‘autism‐friendly’ principles (NICE, 2021a).

Potential autism‐related contraindications were highlighted which require further consideration and could result in changes to the delivery of usual CMT exercises, relating to differences in interoception and imagery abilities. However, during the present research, two small‐scale feasibility studies of compassion‐based interventions with autistic adults were published, which included imagery tasks (Edwards et al., 2024; Riebel et al., 2024). These provide tentative evidence that such approaches appeared feasible for participants, although conclusions cannot be generalised as autism‐related learning differences are known to be varied and individualised (Nader et al., 2022). Practical implications include consideration of PAC learning needs during CFT screening processes. Alternative tangible CMT exercises are emerging, such as the Compassionate Kitbag (Lucre & Clapton, 2021) which promote inclusion from recipients with a broader range of accessibility and learning needs and could be used when imagery tasks appear inaccessible.

Strengths and limitations

This study provides a novel contribution to clinical psychology literature, through the creation of a culturally adapted CMT intervention, and by documenting key stakeholder insights around the appropriateness of delivery through guided self‐help. Furthermore, this study adds to a growing body of literature to detail stakeholder‐informed adaptation processes using qualitative methods. The study adhered to recommendations for adaptation of complex health interventions (Kirk et al., 2020; Moore et al., 2021; Skivington et al., 2021a), reporting guidelines (Stirman et al., 2019) and involvement of people with lived experience to adapt mental health interventions, which may support acceptability, feasibility and intervention appeal for this target population group (Bell et al., 2023).

However, the study was limited by the failure to recruit the anticipated number of parents to attend both focus groups, and only a small number of CPs. Participants shared homogenous demographics, meaning that adaptation suggestions from those from other backgrounds may have been missed. Additionally, dominant voices were prevalent in all focus groups, despite attempts to moderate discussion to rebalance contributions from all participants, which may mean that inevitably some participant views remained unspoken or silenced (Kitzinger, 1995). There was a risk that adaptations completed from later rounds of feedback unintentionally contradicted recommendations from earlier rounds. Attempts to prevent this included thorough documentation of adaptation processes and the final resource being disseminated to all participants who had opted to receive a copy, but no further feedback was received.

Future research directions

The current research indicates several implications for research and justifies a pilot study to assess the feasibility and utility of CMT‐PAC, following complex intervention development guidelines (Skivington et al., 2021a). Findings from a pilot testing phase can inform the appropriateness of progression to a randomised controlled trial to further the evidence base regarding the utility of CFT guided self‐help interventions for PAC. Future research should investigate how fear of compassion is experienced by PAC to inform how this is assessed and managed when CMT‐PAC is studied. The request for promotion of flexible engagement options raises considerations regarding how adherence to protocol is appropriately measured during feasibility testing. Finally, the likelihood of neurodivergent adults accessing CMT‐PAC in practice is recognised, so further studies investigating the effectiveness of compassion‐based interventions with autistic adults are warranted, echoing Mason et al. (2023).

CONCLUSION

This study adds to the growing literature relating to the applicability of compassion‐based interventions for PAC and extends it by creating a stakeholder‐informed CMT intervention for this population. Studies such as this create a promising foundation for further testing of compassion‐based interventions with PAC and highlight the value of stakeholder engagement during adaptation processes, as is recommended within various intervention development frameworks (Moore et al., 2021; Skivington et al., 2021a), yet is still often missed (O'Cathain et al., 2019). Future research should assess CMT‐PAC to ascertain feasibility in practice, and to answer queries raised about how it is experienced and utilised by PAC.

AUTHOR CONTRIBUTIONS

Francesca Kemp: Conceptualization; data curation; formal analysis; investigation; methodology; project administration; resources; validation; visualization; writing – original draft; writing – review and editing. Mark Hudson: Conceptualization; formal analysis; methodology; supervision; validation; writing – review and editing. Thomas Schröder: Conceptualization; methodology; supervision; writing – review and editing. Corinne Gale: Conceptualization; methodology; investigation; supervision; writing – review and editing.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to declare.

Supporting information

Appendix S1

Appendix S2

Appendix S3

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.