I am not going to lie; some people do not even want to talk: Co-design with vulnerable groups affected by child criminal exploitation

Co-production values lived experience as it promotes individuals as assets and offers insight where little is known about a problem. This paper critically considers the pragmatic approach to co-design adopted in Wales with young people and parents affected by child criminal exploitation during the COVID-19 pandemic. It suggests that combining co-design with data collection facilitated the recruitment, sensitisation and facilitation of vulnerable groups. By placing informed consent at the forefront of co-design, young people and parents decided how they wanted their voices heard. Further, combining data gathering with co-design contextualised solution development within their lived experiences. However, embedding these solutions into policy and practice remains subject to existing power imbalances.


INTRODUCTION
Co-production values lived experience as it promotes individuals as assets rather than victims or offenders.It offers insight where little is known about a problem or issue or where a particular group do not engage with services or solutions (Latif et al., 2018).The term arose from Ostrom et al. (1978) observation that people are hidden, untapped resources who can enable the development of more effective and efficient public service outcomes and outputs.In doing so, it provides an ideal tool for developing an understanding of child criminal exploitation, as there is relatively limited evidence about the nature and scale of the problem and what works in the identification and prevention of children becoming criminally exploited.Co-production enables anti-oppressive practice and social justice as it is based on the belief that involving children and adults in service design and development adds value to public service provision (Bell & Pahl, 2018;Dudau et al., 2019).Rather than 'doing to' (Slay & Stephens, 2013), co-designed services engage individuals as part of the 'design team as experts of their experiences' (Sanders & Stappers, 2008:120).Such engagement is often underpinned by participatory approaches where individuals with lived experiences are partnered with academic researchers so their knowledge can be combined to develop optimum solutions (Cargo & Mercer, 2008).Participatory approaches offer a range of methods such as tools, tasks and activities aimed at facilitating participation and fostering genuine co-production (Vaughn & Jacquez, 2020).The use of art, drama or other arts-based participatory approaches provide immersive experiences to promote dialogue between the parties and enhance understanding of lived experiences (Wang et al., 2017).This is aligned with the United Nations Convention for the Rights of the Child (UNCRC, 1989) as co-design facilitates the child's right to have their voice heard as active participants in decisions that affect their lives (Article 12 of the UNCRC).It shifts the power away from professionals so that children are acknowledged and accepted as assets, rather than passive recipients, who can promote a better understanding of their needs and how these needs should be addressed.
It is perhaps not surprising that this has led to the belief that co-production is a 'cure for all ills' of public services (Bussu & Galanti, 2018:347).Such optimism has been questioned by Steen et al. (2018) who have warned that co-production has a 'dark side' where engagement is tokenistic, replicating social and power inequalities and where it is used to rationalise budget reductions or introduce ineffective solutions.This dark side is particularly pertinent for criminally exploited young people as they may be reluctantly involved with social care or youth justice services and their co-production may be subject to professional decisions regarding their vulnerability and the risks and benefits of their involvement.This creates a paradox where professionals may be reluctant to co-design services with vulnerable young people yet the need to design services for vulnerable consumers has been identified as a priority by research centres and networks from around the world (Ostrom et al., 2015:140).There are many reasons why youth engagement in co-design may be limited or reduced.This may be because of their perceived vulnerability including their care or offender status, or distrust of professionals.Such perceptions can lead to the marginalisation of criminally exploited young people, especially if they feel professionals do not acknowledge or value their lived experiences.Young people may also be excluded if they are unable to meet the time or emotional labour required for co-design (Duggan, 2021).Hence, those groups who are able to engage may experience research fatigue when there is increased or sustained interest in a particular issue from professionals, policymakers or university academics (Kidd et al., 2018).This is particularly relevant with county lines (a form of criminal exploitation where young people are coerced or forced by individuals or groups into transporting class A drugs from urban to rural, coastal or border towns) and child criminal exploitation research where growing recognition of the phenomenon has necessitated more nuanced understanding and practice.
With reference to Dietrich et al.'s (2017) six-stage co-design framework for vulnerable groups, this paper critically considers the co-design methods used within a larger university-led study aimed at co-producing a toolkit to enhance community responses to child criminal exploitation in Wales.As the study was undertaken during the COVID-19 pandemic, the shift towards remote communication resulted in adaptions to co-design.This was primarily because it heightened challenges in recruiting criminally exploited young people and parents.The initial strategy relied upon embedding a researcher within specialist third-sector services.This was deemed to be vital to establishing trusting relationships with young people and parents given the transient nature of those affected by exploitation and their fear of violent reprisals from the people exploiting them (Bonning & Cleaver, 2020;Shaw & Greenhow, 2020).However, the lockdown restrictions significantly affected service delivery, with 91% of charities in England and Wales negatively impacted by the pandemic and around half having to adapt their services in response to lockdown restrictions (Charity Commission, 2021).The aim of the paper is to critically consider the complex relationship between co-production with vulnerable groups and the need for pragmatic, ethical and outcome-focussed approaches.In support of Lohmeyer (2020), the paper concludes that some of the challenges of co-production may be insurmountable.Nevertheless, it provides a vehicle for children to share difficult, troubling or hidden problems, have their voices heard and contribute to policy and practice change (Smithson & Jones, 2021).

Background
Despite the renewed interest in co-production, there is no universally agreed definition.It is often used as an umbrella term for a range of activities, such as consultation, participation and involvement.There is also little consensus regarding whether co-production must include the full process of service planning, design, development, implementation, and assessment rather than one or more elements (Maxwell & Corliss, 2020).There is consensus that co-production must include the voluntary, active engagement of individuals working in partnership with professionals with the aim of adding value to outcomes (e.g.Bovaird & Loeffler, 2012).Within public management research, 'co-design' is a discrete form of co-production that brings service designers and service users together to address complex societal problems with the aim of creating innovative service solutions.While the benefits of this approach are well documented, less is known about co-design with: Vulnerable consumers […] whose individual characteristics (e.g.age) or individual states (e.g.cognitive or physical capacity) interact with external conditions (the environment) to create a state of powerlessness in consumption situations (Dietrich et al., 2017:665) This reflects the power-laden interchange between those delivering services and those receiving services, where certain groups are labelled as 'vulnerable' yet this apparent vulnerability is contextspecific and transient (von Benzon and Van Blerk, 2017).In the context of criminal exploitation, young people's 'vulnerability' goes beyond the exploitative relationship to include any existing vulnerabilities that enhance their susceptibility to being targeted.This included individual characteristics, such as age and ethnicity, personal circumstances such as poverty, history of abuse or care experience and enduring or transient vulnerabilities such as physical or mental health issues (National Crime Agency, 2017).In addition, young people experience vulnerability from the wider environment, such as poverty or marginalisation from society due to school exclusion or the lack of social capital (Maxwell & Wallace, 2021).For parents, vulnerability included feelings of helplessness to protect their children and the challenges they encountered in securing service provision and having their voices heard (Children's Society, 2019).While the notion of 'consumer' may be a misnomer given that young people experiencing social care or youth justice services may have little choice regarding their involvement, Dietrich et al. (2017) contend vulnerable groups may find it difficult to articulate their goals for co-design and have fewer resources to realise these goals.Moreover, some groups, such as criminally exploited young people, may be reluctant to engage in co-design due to the distressing nature of their lived experiences, a lack of trust or cynicism that their involvement will effect change (Factor & Ackerley, 2019).Drawing on insight from their review and redesign of an alcohol education programme, Dietrich et al. (2017) concluded that unlike other user groups, vulnerable groups require additional time and resources across six main stages of co-design: resourcing, planning, recruiting, sensitisation, facilitation and evaluation.These stages were designed to address challenges in recruitment and facilitate engagement.Such challenges include reliance on professional judgements regarding young people's capacity to co-design (Maxwell & Corliss, 2020).This can exacerbate existing power inequalities, especially when parents, carers or professionals mediate youth involvement based on judgements of the child's competence, communication skills, reliability, credibility or the risks and benefits of their involvement (Head, 2011;Shier, 2001;Tisdall, 2017).Consequently, co-production has been presented as a continuum, ranging from tokenistic inclusion, which limits meaningful engagement from young people at one end of the spectrum to full engagement at the other (Hart, 1992).According to Shier's (2001) five levels of participation, the lower three levels meet the minimum expectation under the UNCRC (1989): namely, supporting young people to share their views, listening to these views and considering them in decision-making.The higher two levels represent engagement or co-design, where professionals and young people share power and decision-making.In other words, participation can range from 'doing to', 'doing for' or 'doing with' (Slay & Stephens, 2013).Based on findings from their literature review of co-production in mental health, Slay and Stephens (2013) found that some services adopted a 'doing to' approach.
Often used in traditional service design and provision, this approach was underpinned by the belief that young people should be informed or educated by more knowledgeable, expert professionals (Booth, 2019).Adult-led approaches conflict with the young person's right to make decisions and assert independence.They can also negatively impact service engagement, with older children less likely than younger ones to engage in adult-designed and delivered services (Yeager et al., 2015).Services adopting the 'doing for' level of participation (Slay & Stephens, 2013) acknowledged young people as legitimate actors within the process rather than passive service recipients (Bovaird & Loeffler, 2012).Nevertheless, youth engagement was still constrained by professionals, for example, inviting young people to rank or comment on pre-determined options or only consulting them when it was too late to respond or influence decision-making (Mazzei et al., 2020;Tisdall, 2017).Finally, services adopting a 'doing with' approach represented full participation where young people co-designed and delivered the services they developed through activities such as peer support, mentoring or contributing to organisation decisions (Slay & Stephens, 2013).
In practice, services often outsource co-design to third-sector organisations or research institutions.While this is 'not a guarantor of meaningful service user inclusion' (Mazzei et al., 2020(Mazzei et al., :1279) ) it can mitigate these challenges and prevent siloed approaches within public service offerings as it enables the inclusion of multiple organisations, including public and third-sector organisations.Outsourcing to universities may also introduce a power dynamic where academics are deemed as experts in the field while professionals and children are the 'researched' (Pinfold et al., 2015: 221).There is also the risk that co-design becomes a standalone exercise rather than an ongoing conversation with individuals about service improvement with demonstrable developments (Vis et al., 2012;Wright et al., 2006).This poses the challenge of how to embed co-production within service design and garner organisational commitment to change its culture and structures based on the lived experiences of vulnerable groups and their views regarding how to address their needs.This paper adds to the co-production literature by critically considering how a toolkit to enhance professional and community responses to child criminal exploitation was co-designed with young people, parents and professionals.Rather than adopting co-design as a separate task, it blurred the distinction between data collection and co-production and in doing so, highlighted advantages relating to informed consent, sensitising young people and parents to the task and making co-design more meaningful to their lived experiences.Further, it offers insights into how university academics can foster equitable co-design with vulnerable young people, parents and professionals within the constraints of a time-limited study and draws upon Dietrich et al.'s (2017) six-stage co-design framework for vulnerable groups.

METHOD
This paper draws upon findings from the co-design methods used as part of a larger study that examined how child criminal exploitation manifested in Wales and what approaches were deemed most effective in safeguarding young people.The study adopted an exploratory design consisting of three phases: (1) a mapping exercise consisting of qualitative interviews and focus groups with young people, parents and professionals regarding the nature, scale and current responses to children at risk or involved with child criminal exploitation; (2) co-design of a toolkit based on qualitative findings and a literature review and (3) preliminary evaluation and refinement based on feedback from young people, parents and professionals.This paper focuses on phases two and three.
Co-design drew on action research principles where participants began by evaluating and critically reflecting on phase one findings before embarking upon an iterative process of toolkit development and evaluation.Ethical approval for the study was granted by the Cardiff University School of Social Sciences Research Ethics Committee.Pseudonyms were used for participants and information has been anonymised.

Participants
The study adopted a purposive sampling method to recruit young people, parents and professionals with direct experience of child criminal exploitation.

Young people
The original plan was for phase one findings to inform phase two regarding the range of young people affected by criminal exploitation and the services they accessed.It was anticipated that a small number of young people participating in phase one would engage in co-design.To mitigate against potential challenges in identifying and recruiting the required sample, a partnership was brokered with a third-sector organisation delivering specialist services for criminal exploitation.This organisation expressed a willingness to vouch for the researcher and support recruitment and engagement within the co-design activities.This was particularly important given many young people had been manipulated through grooming where older peers or adults had befriended them to coerce them into criminal activities (Harding, 2020).Having an organisation willing to speak to researcher's credibility and trustworthiness was deemed vital for recruitment (Lenette et al., 2019).It also had the added bonus of a safe and familiar environment in which to engage young people in participatory arts-based activities.However, in the time between study application and award, funding for the specialist services ended and was not renewed.Attempts to broker similar relationships with other specialist services were initiated but they were curtailed by the COVID-19 pandemic.This had the immediate effect of delaying participant recruitment but as organisations amended their provision in light of lockdown restrictions, challenges were encountered in participant recruitment and engagement.Consequently, study methods were amended and in light of recruitment challenges, a pragmatic approach had to be adopted for the larger study and co-design methods.Young people were recruited remotely using gatekeeper organisations, including children's services, education, housing, police and third-sector agencies.This culminated in a sample of 21 young people.Of these, 18 had lived experience of criminal exploitation and four participants were identified as at risk of child criminal exploitation.Three participants were adult males.While all three males were older than the pre-specified sample, they provided valuable insight into the journey from becoming involved in criminal activities through incarceration and transfer onto more positive pathways.The 18 participants under the age of 18 were aged between 12 and 18 years, with a mean age of 15 years.Of these, two were female and 16 were male.Regarding ethnicity, around half of young people were white.

Parents
Similar challenges were experienced in relation to recruiting parents as noted for young people.Therefore, parents were recruited through gatekeeper organisations.However, initial response rates were low.According to gatekeeper organisations, parents were wary about engaging with university researchers in case this placed their children at risk of harm from the people exploiting them.Even where parents indicated a willingness to participate, the distressing nature of exploitation meant they cancelled co-design sessions.To address recruitment challenges and in consultation with the study advisory group it was decided that recruitment would be extended to England (for further details, see Maxwell, 2023).This led to the recruitment of 15 parents, eight of whom were living in Wales and seven were living in England.The Welsh and English samples differed in three main ways: average number of years of exploitation (Wales = 2.5 years, England = 6 years), ethnicity (parents living in Wales were more likely to be white) and parents living in Wales were four times more likely to be married that their English counterparts.The sample consisted of 13 mothers and two fathers.Regarding ethnicity, one third of parents were white.

Co-design recruitment: Young people and parents
Rather than introducing the study in distinct phases, gatekeeper organisations invited young people and families to help improve practice for others in their situation.This blurred the distinction between study phases and as such, young people (n = 21) and parents (n = 15) were invited to evaluate and critically reflect on current service provision during phase one and contribute to the toolkit.They were asked for their willingness to contribute to co-design in phases two and three.While all participants indicated their willingness, challenges were encountered in re-establishing contact with young people and parents for subsequent co-design sessions.This was due to a range of factors including the fluctuating nature of child criminal exploitation and associated risks and distress, service disengagement and a lack of organisational capacity to support the study.To address the low number of young people, two additional groups contributed to co-design: an established group of young people who had experienced, perpetrated or lived in a violent area (n = 10) and a group of young people accessing a specialist child criminal exploitation service (n = 4).Hence, 18 young people and 17 parents contributed to co-design.

Professionals
For professionals, recruitment to phases two and three adopted a combination of purposive and snowball sampling based on three criteria: direct work with criminally exploited children, representation from all 22 local authorities in Wales and stakeholders from children's services, education, health, housing, police, probation, youth justice and youth services.An initial core group consisting of stakeholders who had supported the initial research tender were invited to make recommendations on whom to invite to the group.This led to an advisory group of 15 members.To ensure its independence from the research team, the advisory group was chaired by a senior stakeholder with experience across several sectors.The chair sought agreement regarding the group's role and remit.This consisted of knowledge transfer, evaluation and critical reflection, facilitating the engagement of other stakeholders and disseminating the finalised toolkits.Three co-design sessions were held with additional contributions made on a one-to-one basis via email.An additional, ten sector experts provided feedback on toolkit drafts.A total of 19 professionals engaged in co-design.

Analysis
All study participants gave informed consent for phase one and interviews were undertaken remotely using Zoom or Microsoft Teams.Phase one interview recordings were transcribed verbatim.Sections relating to co-design were coded separately and collated with notes taken from the professional co-design session.Data was analysed according to the principles of grounded theory (Strauss, 1987) using a coding framework initially devised deductively, based on findings from phase one.The coding framework was refined on subsequent co-design communication, which was primarily undertaken through email.Analysis was facilitated by NVivo software version 12.

FINDINGS Resourcing
Following Dietrich et al.'s (2017) co-design framework, the study began with resourcing, where relevant sources were selected by university academics to give co-designers an in-depth understanding of the problem to be addressed.Rather than drawing on existing literature and evidence reviews, findings from phase one were used as they sought to capture the voices of young people, parents and professionals with direct experience of child criminal exploitation (Maxwell & Wallace, 2021).This garnered insight into how child criminal exploitation affected young people and their parents and how professionals were defining and responding to this cross-cutting issue.In doing so, phase one findings yielded insight from multiple perspectives, highlighting the similarities and differences within and between participants.
To foster co-design, findings were presented in two reports (young people and parent findings and professional findings) and summarised in a newsletter format.Youth and parent codesigners had the option of receiving the newsletter and internal reports while professionals were sent the two reports.The division of young people and parent findings from professionals was a pragmatic decision due to delays in recruiting young people and parents to the study.As noted, all young people and parents were invited to co-design the toolkit during phase one data collection.Data regarding toolkit content and format was excluded from the newsletters or reports so it would not be subject to critical evaluation from other co-designers.Hence, this data were included as user-generated solutions alongside those produced in co-design sessions and communications.

Planning and recruiting
In support of Dietrich et al. (2017) recruiting co-designers was a lengthy process facilitated by the adoption of a strengths-based approach and alignment between the university academics, young people, parents and professionals' goals.The process began with the dissemination of a short summary of the study aims, phases and anticipated outcomes.For young people and parents, several recruitment strategies were adopted as some were still affected by child criminal exploitation.This posed ethical and safeguarding challenges and highlighted the tension between protecting vulnerable individuals whose lived experiences may be upsetting or distressing and protectionism where professional worries about risk and vulnerability obscure their right to participate in decisions affecting their lives (Garcia-Quiroga & Agoglia, 2020; Vis et al., 2012).Additionally, safeguarding considerations were paramount given that children and parents could be victims of repercussions from the people exploiting them (Maxwell & Wallace, 2021).Therefore, gatekeepers adopted a 'right time' approach to foster engagement from children and parents when they felt safe and able to engage (Urry et al., 2015).In doing so, co-production acknowledged the need to: keep the door open and taking account of the fluctuation in their lives allowed for young people to pick their own 'right time' (Urry et al., 2015:300) Despite overall support, many gatekeepers were reluctant to approach young people and parents due to concerns about over-intervention, especially as their own engagement with these groups was tenuous (Garcia-Quiroga & Agoglia, 2020).This reluctance was fuelled by the risk of 'over-researching' (Mulvale et al., 2019:287) this group as professionals, university academics and students vied to speak to a relatively small number of children and parents with lived experience.Gatekeepers prioritised their own information-gathering as many were engaging in exploitation strategy meetings and feeding their intelligence into multi-disciplinary approaches on the frontline.Consequently, one local authority produced its own toolkit which was rapidly implemented across Wales, thereby negating the perceived need for a co-designed toolkit based on research evidence (discussed later).
In consultation with the professional advisory group, the planning stage was extended for young people and parents.It was decided to invite young people and parents to participate in phases one (data collection) and two (co-design) simultaneously.As such, young people and parents were made aware of the potential risks and benefits of participation following the usual research ethics processes.For young people, subsequent co-design activities such as commenting and evaluating drafts of the toolkit were facilitated by close collaborations with gatekeeper organisations rather than direct communication with the academic team (Dietrich et al., 2017).This was a departure from the proposed approach and reflected the challenges of undertaking a time-limited research study during the COVID-19 pandemic and the delays encountered while organisations adapted to the lockdown restrictions and transferred their services to remote operations (Charity Commission, 2021).This created tension between what was possible in the context of the pandemic and the desired outcome of genuine co-production which required the engagement of young people with different lived experiences of criminal exploitation across different contexts in a dynamic process of co-design using participatory arts-based methods (Fine & Torre, 2019).In reality, and influenced by the time-limited nature of the study, first-hand accounts from young people became limited to phase one participation.Despite numerous attempts at re-engaging young people for phase two, gatekeeper organisations were unable to offer ongoing support for the study.Therefore, co-design activities were sent to gatekeepers who completed these activities with young people and returned notes to the researcher: The disadvantage of this approach was that it failed to ensure that young people had a 'seat at the research table; that no one can speak "their" stories for "them"' (Fine & Torre, 2019:435).While this posed the risk of marginalising young people, it highlighted the complexity and messiness of co-production and the inherent challenges of accessing young people via partner organisations (Thomas-Hughes, 2018).Moreover, it posed the risk that young people's engagement would be subject to social and power dynamics from gatekeepers.However, this approach was deemed favourable to omitting them from the study for two main reasons.First, young people were primarily recruited through third-sector organisations with whom they voluntarily engaged.These organisations were selected due to their commitment to the UNCRC (1989) and their strengths-based ethos.Second, young people were offered a range of participatory methods of engagement, including creative artsbased and audio-visual methods.To support this, a booklet of activities was developed and disseminated.These activities were completed by the at-risk group of three young people during a workshop.Conversely, young people with direct experience of criminal exploitation opted to have topic areas for discussion prior to meeting the researcher.This situated co-design between 'doing for' where young people commented on findings and toolkit designs and 'doing with' where they directed content and how the information was presented Slay and Stephens (2013).
While recruitment remained with gatekeepers, the study adopted a flexible strategy to foster engagement and mitigate against the risk young people would be reluctant to engage as one young person noted, 'Some people don't even want to talk to you guys.I ain't going to lie; some people don't even want to talk' (Jordan, youth interview).All young people were invited to contribute ideas for the toolkit.Similar recruitment challenges and strategies were adopted for parents.All parents were invited to share their views regarding toolkit content.Subsequent co-design was undertaken on a one-to-one basis through email communication as parents indicated this was their preferred method.
As noted, professionals became co-designers as part of their advisory group role.Advisory group members were selected based on their knowledge and expertise in a range of roles and across sectors.Additional stakeholders were invited to comment on toolkit iterations on an ad hoc basis.

Sensitising and facilitation
Youth co-designers While Dietrich et al.'s (2017) framework was developed using a case study approach in a school environment, this study attempted to co-design with young people who were not accessible through a shared physical location.Rather than a discrete group of vulnerable co-designers, the sample comprised young people with different ages, backgrounds and experiences.Even where young people were accessed through the same organisation, gang affiliations and tension between postcodes meant they could not be brought together safely.It also proved difficult as they participated in both phases simultaneously, blurring the distinction between taking part as information-giving participants (phase one) and active users generating ideas and potential solutions (phase two).Due to youth reluctance to engage in lengthy co-design sessions, it also prevented the use of participatory tools and art-based methods designed to sensitise young people to co-design, such as card sorting or storyboards.This rendered co-design a 'complex social intervention' (Donetto et al., 2015:240).Rather than seeking a uniform approach or following design frameworks, the emphasis was placed on supporting young people to co-design on their own terms in their preferred way.
With these limitations in mind, the researcher addressed power imbalances in phase one by developing a respectful, non-judgemental and reciprocal relationship with the young person (Wright et al., 2006).Young people were visibly uncomfortable and anxious about participation and so, explicit reference was made to the researcher's lack of lived experience and it was noted that many adults cannot claim to know what it is like to be groomed or affected by child criminal exploitation.During the exchange, young people became visibly more relaxed, for example, by removing their hoods so their faces could be seen and engaging in conversation with the researcher, for example, explaining colloquial terms and how exploitation manifests.Anecdotally, the difference was stark.Young people moved from streetwise anonymous people in hoods to young men who laughed and smiled and talked about their hopes and dreams: To mitigate against the potential pressure placed on youth co-designers to produce 'the solution', it was also made clear that there were multiple co-designers and toolkit development would be an iterative process.Young people were sensitised to the co-design task through thought-provoking questions asking them to consider what they wish they had been told, by whom and how to protect other young people from criminal exploitation.The shared goal of protecting other young people facilitated co-design between university academics and young people (Factor & Ackerley, 2019).This helped maintain their focus and encouraged them to critically reflect on appropriate solutions for young people: Because there's loads of stuff that's portrayed in books and text books about selling drugs that, 'no, I don't believe that.I'm living that lifestyle.I don't know what you're on about' So you've got to come across in different ways (Jordan, youth co-designer) Despite initial fears that combining phases one and two and a lack of participatory tools would obscure sensitisation to co-design, 'meeting young people where they were at' (Mulvale et al., 2019: 287) both spatially and emotionally, enabled young people to become active co-designers.

Parent co-designers
Parents also participated in phases one and two simultaneously.Rather than collaborating in co-design sessions as a group, gatekeeper organisations reported parent fears about sharing their experiences with others.This fear was supported by phase one findings where some parents reported that it later transpired that their children had been exploited by people known to them (Maxwell & Wallace, 2021).This reflected a distinction between individual resources and capacity and safeguarding risk.To safeguard parents and their children, co-design was undertaken individually via remote methods in the first instance, followed by email communication to gather their views, comments and feedback on toolkit iterations.
In terms of co-design sensitisation, parents shifted between information-giving and problemsolving.This manifested in narratives that moved fluidly between lived experience and what was needed to improve practice: I think any kind of support service that encourages someone to talk about their experiences will help.It just helps you get through the days, weeks, months, years, however long you go through this (Gemma, parent co-designer) University academics adopted a non-judgemental, respectful stance acknowledging parent contributions and inviting them to elaborate on their ideas.Hence, in response to Gemma's wish for support services, the researcher probed whether Gemma wanted to be signposted to such services or whether she had found any services that met her needs: So I think parents need to be made aware of what services are out there and how they can access them (Gemma, parent co-designer) While not directly comparable with collaborative co-design, this was deemed the optimum method of recognising the contribution and stimulating a more in-depth discussion of what Gemma wanted this to look like in practice.

Professional co-designers
Extending the advisory group role, co-design with professionals consisted of an introductory codesign session attended by nine professionals.Acknowledging high workloads and time commitment for the 3-h co-design session, invitations were sent 8 weeks in advance.On the day of the session, professionals were sensitised to the task by inviting them to spend the first 90 min reviewing the phase one reports.The aim of these reports, or 'tools for conversation' (Dietrich et al., 2017:667) was to sensitise professionals to how child criminal exploitation manifested in Wales from different perspectives.Allocating dedicated time to review these materials was received favourably as noted by the chair.As the session coincided with the pandemic and greater homework, several professionals reported they had read the reports outside in the sunshine.This gave them time to digest the wealth of information and distance themselves from their professional roles.Conversely, for those unused to digesting large amounts of information, insufficient consideration was given to the balance between providing detailed findings about child criminal exploitation from multiple perspectives and professional capacity to review all material in the allotted time.Potential differences between professionals were mitigated by having an experienced and independent chair who highlighted the value of professional expertise and welcomed contributions regardless of whether professionals had reviewed all materials.Professionals were given four reflective questions inviting them to consider the key messages for the toolkits, existing interventions that could be used or adapted to support young people and parents, what policy and practice guidance had to be included and what format the toolkit should adopt.To avoid researcher contamination, the chair facilitated discussion ensuring all participants had the opportunity to contribute.

Co-production cycle
All three groups of co-designers offered insight regarding the extent to which current practice enhanced or hindered responses to child criminal exploitation.Broadly speaking, they all highlighted the need for more information about child criminal exploitation and improvements to existing practices.It is possible that the process of sensitisation and facilitation may have constrained responses to the critiquing of available solutions as opposed to generating novel or innovative solutions (Trischler et al., 2019).Consequently, all three groups posited enhancements to existing provision and there was consensus one toolkit was insufficient to meet the needs of young people, parents and professionals.Rather it was suggested that a 'toolkit of toolkits' (Professional co-designer) was required encompassing resources tailored to the needs of each group.As such, co-design did not produce: …market-ready solutions.Rather, they provided user-driven idea inputs for the research team and service designers to build upon in consultation with other stakeholders (e.g.public authorities or policymakers), who are the ones responsible for creating and implementing the final concept (Trischler et al., 2019(Trischler et al., :1612) ) This revealed an unanticipated additional phase of the study, as university academics had to collate ideas and build the toolkits.It also highlighted a key weakness of the study as insufficient resources had been allocated to co-design of the final toolkit and the iterative process of refining the toolkits in light of feedback and the initial study aims.For university academics, this created tension between the realisation of these ideas and their work commitments (Pinfold et al., 2015).

Youth toolkit
On balance, young people said that raising awareness about child criminal exploitation in schools would be beneficial although a few queried whether they would have listened 'People don't learn from what others have done, they can't relate, they think "that's not me"' (Aiden) whereas others felt that hearing from those with lived experience would be a more helpful approach than listening to teachers or other professionals: Having that conversation with somebody who's lived it, been there, rather than somebody who's I don't know seen something on the internet and discussing that with you (Jack, youth co-designer) However, professional and parent co-designers expressed concern that this could be counterproductive as it amounted to presenting young people with older peers who had experienced the apparent glamour of the drug dealing lifestyle and then successfully moved into a respected, professional role.Rather than blithely accepting this view, university academics ensured that adult co-designers did not mediate or override young people's proposed solutions (McNeilly et al., 2015).To this end, Aiden's assertion above was considered with Cameron's vivid description of how he attends to information: Adults just, it's not like, it's not stupid that they talk, it's just, it's good but it just goes through one ear and it comes straight out the other … if my friend said something to me yesterday and then he asked me, "Oh, what did I say yesterday?"I would probably forget.But if he showed me like £100, he'd be like, "Oh, what did I show you yesterday?"I'd be like, "It's £100." Therefore, the potential of a film visually depicting young people with lived experience was posited.To test this proposed solution, an additional group of young people affected by child criminal exploitation were consulted about the film alongside parent and professional co-designers; all three groups were favourable.This resulted in the co-production of two films with young people about how child criminal exploitation affected their lives, either directly as those being exploited or indirectly through intimate or familial relationships.While the researcher oversaw the process to ensure that content was anonymised, young people retained control over script writing, filming and editing.Second, all three groups noted the importance of having good relationships with professionals.Such relationships took time and necessitated a patient approach as young people were wary of professionals.As noted, attempts to re-engage youth co-designers proved impossible.Therefore, an accessible group of young people working as peer researchers for a youth violence prevention project were invited to a co-design session.The aim of the session was to explore how young people wanted professionals to engage with them.Within the session, peer researchers received a brief overview of the study before being asked to brainstorm their ideas of what good relationships and communication look like, for example (Figure 1): Brainstorming notes were collated and analysed thematically.This led to the generation of 19 statements.Once the peer researchers reviewed and approved the statements, they were presented as a 'Youth Charter' for inclusion within the toolkit's youth resources.

Parent toolkit
For parents, the need for clear, accessible information was highlighted.Phase one findings highlighted that some parents had not known about child criminal exploitation or perceived it to be something that would not happen to their child.Hence, parent co-designers wanted their resource to be aimed at information raising alongside detailed information for parents whose children were being criminally exploited.Parent co-designers provided specific guidance as to what information would be most useful: I think it's important that when this information in the toolkit is out there, the parents actually understand why that toolkit is in-place.It's one thing having resources, but if a lot of parents are actually either, one, turning a blind eye to it, or genuinely don't understand, there's got to be something that alerts them to why they need to take that onboard (Fiona, parent co-designer) This was supplemented with feedback from youth and professional co-designers.Proposed content for the parent resource overlapped with that of the professional resource therefore, the parent toolkit included content from the professional toolkit and new material.The key challenge was how to make the information available and accessible to parents.Due to the technicalities and limited capacity to develop the toolkit, a parent website was developed.To do this, a learning and development manager was invited to join the university academics.Their expertise proved invaluable in developing and presenting information on the parent website.Additionally, a z-card was produced to raise awareness of child criminal exploitation and signpost parents to the website.Despite expressing interest, only two parent co-designers commented on iterations of the resource.This risked amplifying existing parental disadvantages as highlighted in phase one (Maxwell & Wallace, 2021).These findings demonstrated that parents were a vulnerable group as they experienced secondary effects of child criminal exploitation as well as often being blamed by professionals and stigmatised by community members.Further parents reported the difficulties they encountered accessing information about services and processes so they could advocate for their child.To avoid 'doing to' (Slay & Stephens, 2013) parents it was vital that the parent resource addressed their needs.Hence, infographics of systems and processes were created and two additional parents were approached and encouraged to feedback on the proposed content.The content was modified based on feedback from all four parents.

Professional toolkit
Reiterating the findings for the youth and parent resources, all three groups highlighted the need for an improved professional understanding of child criminal exploitation: So almost like a guide, like you can get a little placard and you can get your little bullet points and it's just like signs of exploitation, say, for argument's sake, and then if they need a bit more information, at that point, when you say, 'if you've noticed all five of these areas, or all five of these things, then move on to this or alert a member of the safeguarding team where they can pick it up'.Because I think those that are in the frontline are crucial to the children being safeguarded (Fiona, parent co-designer) This was supported by youth and professional co-designers.For example, phase one demonstrated a tendency for professionals to view child criminal exploitation with reference to county lines and trafficking young people into Wales from England.Whereas, young people and parents reported locallybased exploiters and young people being trafficked within Wales (Maxwell & Wallace, 2021).In addition to a more nuanced understanding of child criminal exploitation, parent co-designers called for professionals to work with them rather than assessing and placing them under surveillance: All I know is that I've been parenting him for the whole of his life and now he's about to turn 18, I'm being assessed whether I can parent him, I think it's a joke.I don't think the court or social services processes are set up to deal with it (Ella, parent co-designer) All three groups called for service change, with phase one findings providing evidence-informed proposals for what was required.Professional co-designers called for sector-specific information written in the appropriate language for each different sector with reference to relevant policy and practice guidelines.This led to the creation of a practitioner toolkit separated into two sections.Section 1 included definitions and guidance for multi-agency working and working with young people and families.Section 2 included sector-specific guidance.Collation of this wealth of information was a huge task, exacerbated by the differing perspectives among the professional co-designers.This was compounded by the drive for inclusivity as the researcher called for feedback on early iterations rather than spending time reflecting and developing proposed solutions.In this regard, the university academic's expertise in producing and presenting evidence should have had a more prominent role in co-design.This suggests that rather than adopting a facilitation role, university academics should have acknowledged and more clearly defined their role as a co-designer based on their knowledge and expertise of this issue.Nevertheless, university academics played a prominent role in writing and presenting the content, receiving and acting upon feedback and ensuring that all co-designers had the opportunity to comment and have their views addressed in the final outputs.

Evaluation
The study culminated in the production of a toolkit of resources for young people, parents and professionals affected by child criminal exploitation.Early feedback from co-designers was positive about the extent to which final outputs included their views: [Young person] and his Mum think that the toolkit is useful for families and parents and has good bullet points to look out for (Email from professional gatekeeper) I've gone through the toolkit and think it is really comprehensive.I love the graphics and how it includes everything we talked about during the meetings (Email from professional co-designer) Co-designers contributed to the development of resources for different audiences ensuring the toolkit was relevant to the needs of criminally exploited young people and parents.Due to time commitments for the youth films, these are yet to be finalised.The parent website attracted over 1500 visitors in its first 3 months.While this is not evidence that parents are using the resource it does demonstrate that early attempts at dissemination are generating visitors to the resource.Finally, as noted, the presence of a similar local authority-led toolkit has resulted in significant challenges in encouraging the adoption of the co-designed toolkit for professionals.It has been suggested this is because local authorities await official confirmation from the government to use the co-designed resource.It is tentatively suggested it also demonstrates the extent to which organisational factors can hinder which co-produced outputs are embedded into practice.Usubillaga et al. (2023) realist review found that implementation is complex and dependent on it being succinctly communicated to professionals and sufficiently different to local policies and current practice before it will be adopted.As the co-designed toolkit complements the local authority toolkit and existing policy it is possible that the co-designed solutions were not sufficiently innovative to encourage adoption.

DISCUSSION
Drawing on lessons from the co-design methods pragmatically employed following the restrictions imposed by the COVID-19 pandemic, support was found for Dietrich et al.'s (2017) sixstage framework.This paper posits that combining co-design with data collection facilitated the recruitment, sensitisation and facilitation of vulnerable groups.Challenges in recruitment meant that rather than six distinct stages, resourcing, planning and recruiting became an iterative process with parents and young people recruited as both study participants and co-designers simultaneously.This was not unique to this study.Findings from Amann and Sleigh's (2021) literature review revealed difficulties in initiating collaboration with vulnerable groups, especially where there were no existing connections.This required a substantial time commitment and necessitated reliance on gatekeeper organisations who were also experiencing difficulties engaging criminally exploited young people and parents.Recruiting vulnerable individuals, still subject to distressing and potentially harmful effects of their lived experiences, required adherence to rigorous ethical and moral considerations.This was facilitated by adopting a flexible approach that enabled young people and parents to participate on their own terms in their own way.Combining co-design with research participation minimised their time commitments and reduced the overall number of young people and parents needed for the study.In doing so, it obscured opportunities to engage in ongoing, active conversations with young people and parents aimed at equitable collaboration.This heightened the potential for power imbalances as young people's communication with the researcher became fragmented and mediated by professionals (Vis et al., 2012).It is suggested that the diverse nature of young people and parents recruited to the study would have been impossible had co-design been undertaken collaboratively in a specific physical location due to safeguarding concerns at the individual and organisational levels.
This highlights the unique nature of the study as it sought co-design solutions for the crosscutting issue of child criminal exploitation rather than a specific need and service response.This gave rise to a broad sample of young people and parents with different experiences of service pathways and engagement.Consequently, the potential group of service users and providers was extended beyond one service to encompass a vast array of co-designers with relevant expertise.This necessitated a re-imagining of co-design with levels of engagement considered on a caseby-case basis (Nind, 2014).It was not co-production per se but the aims, purpose and potential to effect change that was of importance.This required the adoption of a pragmatic stance that acknowledged the extent to which co-design was influenced by the wider environment in which it occurred and where: despite our best efforts to ensure that risks are minimised in the field -unforeseen events can lead to increased risks or ethical dilemmas, both for community-based and academic researchers (Lenette et al., 2019:162).
The COVID-19 pandemic initiated a sequence of challenges to co-design, including the establishment and maintenance of relationships with gatekeeper organisations and restricted access to young people and parents.This was compounded by the constraints of academic study as the time available for undertaking participatory approaches was reduced (Rowley et al., 2022).
Co-design offered a valuable tool that enabled young people and parents to share experiences that were difficult, troubling or hidden and sought to involve those who were: Fundamentally altered by child protection involvement [and] have first-hand knowledge of how policies and practices work, and the most to lose when they do not (Black et al., 2023:3) But co-design had the potential to trigger painful memories and experiences young people and parents may not have anticipated (Isham et al., 2019).By including them as research participants subject to informed consent procedures, young people and parents had the time and space to consider how much personal information they wanted to disclose to university academics about their lives.It also meant that rather than merely participating as research participants, their engagement resulted in tangible outputs.This necessitated a reconceptualisation of the ethical process to include co-design activities so that young people and parents were able to consider the potential constraints imposed by ethical guidelines and decide what roles they wanted to adopt.Co-design aligned with youth and parents desire to help other young people and their families, which emerged as the primary motivation for engagement (Engström & Elg, 2015).Where sensitive and distressing topics are concerned, this is a crucial moral consideration.Young people and parents are more than just experts by experience; they have lives, interests, and responsibilities (Factor & Ackerley, 2019).Combining data collection with co-design empowered young people and parents as experts who were uniquely equipped with the knowledge and skills to create an appropriate, relevant toolkit.This facilitated sensitisation as it prepared them for co-design and removed the need for participatory tools for engagement purposes.This deviated from Dietrich et al.'s framework which recommends the use of creative, participatory tools to empower and engage vulnerable users.Yet, these tools are often selected or designed by expert adults.Hence, Lohmeyer (2020) has argued they hide adult-centric power-knowledge dynamics and are based on the assumption that young people are incapable of engaging in research.It is posited that giving criminally exploited young people and parents the freedom to co-design in their own way both enhances and restricts their ideas.First, in terms of enhancing co-design, relatively little is known about child criminal exploitation on which to create relevant participatory activities.Attempts at creating suitable activities could have constrained or deterred young people and parents from participating.Second, co-design could have been restricted as participants were immersed in their own lived experiences rather than being encouraged to think more laterally about young people, parents and services.Co-design reduced the power differential between young people, parents, professionals and university academics and provided a space where they could co-produce more appropriate services and interventions to address their needs (Isham et al., 2019).However, the study was limited by the siloed nature in which co-design took place.Rather than a fully collaborative process between young people, parents, professionals and university academics, it was undertaken separately with each of the three groups.This was an obvious disadvantage as it restricted creative flow between co-designers.It also placed ethical and moral responsibility on university academics to ensure co-designers had an equitable role in developing the final resources.This required considerable skill on the part of university academics as they managed different pre-conceptions, priorities, values and attitudes (Amann & Sleigh, 2021).It is posited this enhanced youth and parent voices as rather than being subjected to professional appraisal, university academics held their views in mind ensuring they were included in final outputs.The downside was the substantial time and resources required for facilitation.In support of Trischler et al. (2019), a seventh stage, that of building for change, was required to include refinements to toolkit drafts and encourage adoption of the toolkits into practice.Despite the support from professional co-designers across sectors, embedding the toolkits into practice has, thus far, proven difficult.While there are indications it will be endorsed within forthcoming policy, this has taken considerable time.In this respect, it is understandable why varying levels of co-design are employed: Reflecting on the challenges identified across the seven steps, it becomes clear why many public sector organisations still design for rather than with service users (Trischler et al., 2019(Trischler et al., : 1614) ) Embedding co-designed solutions into practice requires adoption from local government, buy-in from senior service managers and adoption from frontline professionals who believe it will affect positive change and be prepared to change their practice.Without this, co-design risks identifying what matters to vulnerable groups with no subsequent actions or attempts at service improvements.University academics play a vital role in the process.At a pragmatic level, they must ensure that co-design is adequately resourced within research applications and the proposed impact is realistic.At the change level, they must establish relationships with key stakeholders who can ensure that codesigned outputs are adopted.Without this, the inclusion of vulnerable groups remains tokenistic regardless of their level of engagement.
At first, YP [young person] began by calling the person who was quoted [in the newsletter] 'if you're dealing weed locally' a snitch.[We] encouraged YP to explain what he meant and that's where the YP opened up about what big drug dealers would do and say to persuade young and vulnerable people to join the crime, 'They'll buy you food first' 'They'll buy you trainers n clothes' 'Then they'll expect you to do favours for them' [extract from notes sent from gatekeeper organisation] If I do work right now until next year I could literally buy a car straight away because that's what I want, I want to buy a car.I want to open my own place, my own barbers or something like that (Isaac, youth co-designer)