Child sexual exploitation and the adoption of public health approaches to prevention: critical reflections on evolving processes and practices

In recent years, the use of public health approaches to address complex social problems has gained popularity. In England and Wales, the rise in low-volume, high-harm crime has accelerated this shift, with calls for public health interventions to tackle knife-crime, extremism and sexual violence made by politicians, policy-makers, welfare workers and the police service. Notwithstanding such appeals, how public health approaches are both operationalised and impact remain largely unknown. Drawing on findings from a qualitative study focused on the implementation of a specific initiative in the UK designed to reduce the risk of Child Sexual Exploitation (CSE) amongst young people, this article attempts to address tangible gaps in these two key areas of knowledge. Although generally supportive of a public health approach to CSE, an analysis of in-depth interviews with members of a multi-agency team reveals a number of quandaries and thorny issues when implemented within a specific policing and criminal justice context.


Introduction
In the UK and elsewhere, public health approaches to crime prevention have become increasingly popular and have recently been applied across a range of areas of social concern. Harm reduction methods rooted in a public health model initially became popular during the 1980s for adults with substance abuse problems (Newcombe 1992;Stimson 1992) and have since been effective in reducing morbidity amongst 'at risk' populations (Stimson 1995;Stevens and Hughes 2010). The success of harm reduction initiatives has also been observed in efforts to reduce the risk of physical violence and prevalence of sexually transmitted diseases among sex workers (Cusick 2006). It is of no surprise then that strategies informed by the public health model are now being used to address issues such as youth violence, radicalisation and child sexual exploitation (see Levy 2007;Bhui, et al. 2012;Public Health England 2017). In October 2018, the then UK Home Secretary launched a consultation process intended to inform a new legal duty underpinned by a public health approach to tackle serious violence (Home Office and Javid 2018). In-line with these general undercurrents, consecutive UK Government Action Plans to combat CSE have encouraged risk-focused preventative interventions consistent with the public health approach and geared towards raising awareness of risk among young people, parents, carers and potential perpetrators (see Weston and Mythen 2020).
Within these public health frameworks, the police service has assumed key roles in leading and directing a range of preventative activities, including working alongside partners to problem-solve, building cohesive communities, encouraging personal resilience amongst young people, improving data sharing and developing 'whole systems' approaches (see Christmas and Srivstava 2019;van Dijk, et al. 2019). While multi-agency working is necessary to address the diverse and interdisciplinary concerns of public health (Hunter and Perkins 2012), little research has focused on its success within those partnerships that have a criminal justice presence and those that have identify difficulties in aligning the interests and priorities of agencies from different cultural and ideological backgrounds (Weston 2014;Davies, et al. 2023).
For policy-makers, the major appeal of public health frameworks is the promise of clear, evidence-based, practical steps to reduce levels of offending via interventions that focus on the causes of crime, promote use of data and enable working across organisational boundaries (see Catch-22 2019). While there has been recent interest in the UK about whether these types of strategies 'work'-as witnessed by national and local evaluations of Violence Reduction Units (Caulfield, et al. 2021;Quigg, et al. 2021) and plans to investigate the effectiveness of public health interventions led by or involving the police (NIHR 2021)-the myriad ways in which they are both formally and informally implemented remains under explored. As Reimann (2019) notes in his discussion about the 'Cure Violence' initiative implemented in the US, by employing the language of medical science and placing emphasis on the role of data collection and analysis to inform preventative activity, public health approaches 'presuppose the superiority of a quantitative, evidencebased epidemiology over other approaches ' (Reimann 2019: 147). These observations are also reflected in the central focus of research thus far, which has remained fixed on outcomes, seemingly overlooking how public health approaches are qualitatively operationalised on the ground (Bhui, et al. 2012;Public Health England 2019a;2019b).
To begin to fill this knowledge gap, in this article we concentrate on the implementation of a specific awareness raising initiative in the UK designed to reduce the risk of Child Sexual Exploitation (CSE) amongst young people. In so doing, it is our intention to consider how a public health-oriented approach was operationalised by practitioners and potentially impacted by its participants. Drawing on qualitative empirical data-including field observations, in-depth interviews with members of a multi-agency team and focus groups with young people involved in the CSE awareness raising initiatives-we critically analyse the general approach used by the multi-agency team to 'make visible' young people 'which would not otherwise be seen' (see Amoore and de Goede 2008:6). On this basis, we suggest that modes of pre-emptive regulation that attempt to address so-called 'dangers' may identify unnecessary risk and thereby encourage subsequent techniques of surveillance. Our argument is that although population level public health approaches might be necessary and have utility in some areas-for instance in preventing the spread of infectious diseases, such as Covid-19-applying preventative measures under the umbrella of public health within broader contexts of crime prevention may subject individuals to increasing forms of securitisation and unduly expose them to criminal justice processes which they may not otherwise have encountered. While our study is context specific and thus generates findings that are not generalisable, we wish to raise concerns of principle and practice around the need to uphold civil liberties and safeguarding provision for those most in need of protection.

Unpacking public health approaches to crime prevention
Before presenting an outline of the research design and methods, it is necessary to provide an account of the emergence and maturation of public health approaches in areas of crime prevention. Importantly, the application of public health models within crime prevention strategies has a long history. As Tonry and Farrington (1995) suggested almost two decades ago, a public health approach should not be considered a distinctive strategy in itself, as development, situational and community strategies to crime prevention all incorporate facets of public health agendas. It was during the 1990s that public health approaches to crime prevention emerged as an alternative to traditional methods of crime management, through social development and crime prevention via environmental design. In emphasising the value of public health approaches, Moore (1995) posited that violence is a threat to community health which emerges from a complex causal system that goes beyond an offenders' motivations, intentions and character. Ergo, interventions that take place at the level of 'primary' prevention-mitigate against harms before they occur-and at the level of 'tertiary' and 'secondary'-when the risk of violence has already been identified or already occurred-are essential in reducing violent crime. Notwithstanding the above, it is only relatively recently that the public health vocabulary of primary, secondary and tertiary prevention become focal in formal criminal justice discourses. It is only since the publication of the national Policing, Health and Social Care Consensus (NPCC 2018) that all police force areas have been tasked with embedding and advancing public health approaches that emphasise the need to focus on prevention and collaborative working (see Christmas and Srivstava 2019). Following on from this, in July 2019, the UK Government announced that it would introduce new legal duties requiring public services to work together to prevent serious violence. Forming part of a new 'public health approach', funding was made available to set up Violence Reduction Units in the eighteen police forces deemed to be most affected by violent crime. Notwithstanding these developments, the key principles underpinning the public health approach are not new to policing, having previously appeared in various guises and under different nomenclatures, including adopting a population approach, working in partnership, focusing on prevention, using data and evidence to inform practice, addressing causes and evaluating implementation and impact.
The benefits of public health approaches to prevent child sexual abuse (CSA) have been acknowledged for some years (Plummer 2001;Finkelhor 2009;Letourneau, et al. 2014) but it is only relatively recently that they have been used for the prevention of CSE. Fuelled by a series of historical revelations and subsequent prosecutions documented by Jay (2014), in the UK there has been a spike in regulatory activity addressing Child Sexual Exploitation (CSE) and a series of policy imperatives promoting pre-emptive action. Following the exposure of high-profile cases of CSE, risk regulating institutions and agencies have come under increasing pressure to develop new strategies informed by the public health model. Post the publication of the Jay Report (2014), the then Prime Minister, David Cameron, referred to CSE as a 'national threat', directing police forces to collaborate more effectively across regional boundaries to better safeguard children (HM Government 2015). Following this, national guidance for practitioners, local leaders and decision makers was introduced encouraging the use of risk-focused interventions channelled towards raising among young people, parents, carers and potential perpetrators (Department for Education 2017: 18). Within these public health frameworks, the prevention, identification, and reduction in risk became a key objective.
The limited evidence produced thus far suggests some progress in applying public health approaches to a breadth of areas within policing (College of Policing 2021). However, the way in which such approaches are experienced by those engaging with them, and also how they are understood and operationalised by practitioners implementing them, has evaded analytical scrutiny. Often-and particularly within the crime reduction arena-the implementation of public health frameworks necessitates modes of surveillance and intervention. As proponents of the governmentality thesis have illuminated, the valorisation of risk as a driver for governance can incite a panoply of techniques of regulation, education and monitoring, ushered in under the rationale that the safety of both 'risky' and 'at risk' individuals must be prioritised (see O'Malley 1992;Rose and Miller 1992). These assorted processes of surveillance not only define individuals-and indeed communities-as potentially threatening, but effectively 'makeup' subjects and objects that require tracking and surveillance (Kelly 2001; Rose 1999). As we shall elucidate, with recourse to a specific empirical study gauging the efficacy of an awareness raising initiative about CSE with young people, approaches directed towards pre-emptive interventions and undergirded by the assumption that 'anybody can be at risk'-may result in exposing young people to increasing forms of securitisation. Before scrutinising the ways in which public health approaches to the prevention of CSE were interpreted and operationalised in this specific case, we outline below the study design.

Methods, study design and key findings
The data presented below are drawn from a qualitative study involving 65 participants. The study was designed to evaluate the effectiveness of a CSE prevention programme implemented by a large police force area-referred to hereafter as Midshire. Rooted in a public health approach (Public Health England 2017), the key objective of the programme was to raise awareness of CSE among young people and parents via education, targeted prevention and support for victims and offenders. Deploying age-appropriate content, adapted from materials designed by CEOP (Child Exploitation and Online Protection) and children's charities such as the NSPCC (National Society for the Protection of Children) and Barnados, a universal education and prevention programme was delivered across a variety of settings, including schools, colleges and alternative educational providers. Large group presentations were undertaken in schools, primarily focused on online safety and raising risk awareness among potential victims of CSE. These presentations were combined with more targeted work with smaller groups of young people identified as being 'at risk' of becoming either a victim and/or a perpetrator of CSE. The programme was delivered in tandem across Midshire by a multiagency team, led by the police service and involving experienced practitioners working within child/youth welfare-oriented positions, including police officers with 12-25 years' service, a family support worker, a residential care worker, a teacher, a youth worker and a voluntary sector worker, previously counselling children at risk of CSE.
The purpose of the study was to provide an evaluation of the programme as a whole and to conduct a more detailed assessment of practice focusing on the feasibility and effectiveness of prevention techniques used. Data were collected using semi-structured interviews with practitioners, field observations of the delivery of prevention methods and focus groups with young people engaging with the programme. To capture the dynamics of the process of implementation during the span of the programme, a longitudinal follow-up design was adopted. This paper draws exclusively on the data collected from the semi-structured interviews with practitioners. In total, 17 participants involved in delivering the programme were selected, comprising the nine members of the multi-agency team and a further eight members of the steering group. As with the multi-agency team delivering the programme, members of the steering group committee were from diverse contexts and included senior police chiefs, county and city-wide safeguarding leads and commissioners for children's services.
Each participant was interviewed twice over the duration of the project. The first interview took place within six months of the programme's implementation and the second, twelve months later. Interviews lasted between 45 and 90 min and followed a semi-structured topic guide about the CSE initiative and the participant's role within it. Interviews were conducted at either the office of the interviewer or that of the interviewees. Principles established in the code for research endorsed by the British Society of Criminology were adhered to throughout and formal approval for the study was granted by the home University's Ethics Committee.
Post-data gathering, all interviews and focus groups were recorded and transcribed verbatim. The broad principles of grounded theory were followed throughout the process of data analysis (see Glaser and Strauss 2017;Strauss and Corbin 1990). The first phase of analysis involved the researchers independently becoming familiar with the content of the data set and scanning for emergent patterns and themes. Initial open codes were attached to the interview and focus group transcriptions to organise and structure the data. The second phase of data analysis consisted of a more refined analysis of the data, geared towards determining key themes and clustering of sub-themes. The third phase of deliberation and negotiation involved collaboration between the researchers to compare and contrast prevalent themes and sub-themes and to determine dominant and prevalent themes. Following the principles of grounded theory, axial encoding was conducted throughout the process, enabling reflective, iterative and dynamic data analysis (see Ralph et al. 2015).
Outwith the initiative examined here, universal education approaches to the prevention of harms is by no means new. For many years, emergency services have delivered population level awareness sessions to children and young people about various topics including drugs, road safety and firework use. Yet, recent public health developments within policing simultaneously demands that users go beyond universal education and make use of surveillance data to monitor and detect potential irregularities. Within general public health contexts, these methods of data collection and analysis are relatively benign. However, within contexts that include a heavy police presence such approaches may be used in ways that produce iatrogenic effects. As we illustrate below, the universal approach adopted by practitioners in this study drew their attention away from young people who may be in need of safeguarding while simultaneously raising concerns around the encroachment of civil liberties. Rather than adopting a public health approach that makes strategic use of reliable data to make decisions about specific individuals that might be most at risk, and designing preventative activity accordingly, the universal approach adopted was directed towards reaching as many people as possible. Aside from jarring with a more concentrated focus on 'at risk' individuals and/or groups, the universalising principle under which all children and young people were deemed at risk-generated several problems within this specific context.

Shared mission or mission creep? Operationalising 'public health' aims within a policing context
The overall population level approach adopted by the Prevent CSE team was designed to engage all young people. Nonetheless, preventative activity was also aimed at those considered to be on 'the periphery'. Although practitioners carried out some one-to-one intervention work, the initiative was directed at general population level and intended to assist both potential victims and/or potential offenders and encompassing young people thought to be 'lower risk cases': Predominantly we're aimed towards dealing with victims … Vulnerable people who need prolonged and sustained support, we aren't around to do that.
We're also doing some work with potential offenders… I don't think we're dealing with the extreme, or the people who are already offending… it's like the periphery group. So, it's either they're potential victims on the periphery or they're potential offenders on the periphery. That's what we're aimed more towards (Mark, Prevent CSE Team Leader). It's to try and prevent those lower risk cases escalating to a more serious level… If we get the early intervention, and the awareness done, there will be hopefully fewer numbers of children being involved in exploitation. You know, if they are sent some images, or chatted online, or whatever has happened, hopefully, we can stop that progressing any further. So that they're not sort of drawn into this world in a more serious way (Susan, Prevent CSE Worker).
As the leader of the prevent CSE team suggests, the education awareness initiative is rooted in a public health model, but also oriented towards identification of potential cases at a stage early enough to intervene and prevent future harm. Despite being employed by a criminal justice agency-Midshire police-the team leader did not perceive of the team's role as responding to crime nor rehabilitating offenders. Rather, his team's purpose in the awareness raising initiative was predominantly pre-emptive: to identify and divert future harms. This approach chimes with the previous deployment of 'pre-crime' strategies in other contexts of regulation, most notably counter terrorism (see Amoore and de Goede;2008;Mythen and Walklate 2010). The prickly question that surfaces, however, is at what point does preventative activity become a form of disciplinary governance, not so much directed at observing and monitoring those 'at risk' or 'risky', but those outwith this ambit? The team leader's self-correction is also worthy of remark, illustrating the imagined proximity between those defined as presently potentially 'vulnerable' and those that are classified as 'victims'. This has resonances with the seemingly creeping societal spectre of the 'universal victim' and associated thorny issues of what becomes obscure if and when victimisation becomes a classificatory norm (see McGarry and Walklate 2015).
As with traditional public health and crime prevention approaches, the rationale underpinning the initiative to prevent CSE was to raise awareness of risk among young people, parents, carers and potential perpetrators, with reduction in risk being a focal objective. This aspiration was also foregrounded by the strategic leads we interviewed at the onset of the project, as one of them emphasised, that data collected from local police and safeguarding systems would direct the multi-agency team to 'the soil where children vulnerable to CSE breed'. Aligning with a public health model of using data collection and analysis to inform preventative activity, the aim was to 'profile specific geographical areas as potential sites for intervention in accordance with locally produced information systems'. Yet, the approach adopted in practice rarely appeared to take account of information gathered by local safeguarding systems: One of the things that I agreed in one of the early planning meetings for this service was that the Prevent team would attend CSE panels, where we're trying to gather not just information on individual children, but aggregated local based … and we could directly give them pieces of work to do in local areas, based on the findings of the panel. So, I went to observe one of the panels last week and it was clear we've got a cluster of young people who are attending one school. So, the ideal would be for the Prevent team to go in. Not with those young people already identified, but with probably the year group below, and look with the school and the local social work team, what they could target at that group of young people to prevent them becoming the next victims…. I only found out last week that the Prevent team weren't going to the CSE panels … So, by this team not coming, they undermine their capacity to get out quickly to where we need them to go … I think if they lose their link to the CSE panels, that's a really dropped ball. And what it will become is a police initiative with the police sent out to do work in areas dictated to by police led intelligence (Victoria, Strategic Lead).
Despite being a multi-agency team set up to deliver primary and secondary interventions that would help to prevent CSE and respond to it in situations where it has occurred, the population approach deployed by practitioners inclined towards interventions aimed at potential victims and/or potential offenders alongside those thought to be at 'lower risk'. As such, their attention was focused towards reaching as many people as possible. As the strategic lead indicates above, there are clear benefits to universal education initiatives and brief interventions, not least of which being that they are able to reach large proportions of the population. However, as the prevent CSE team leader acknowledges, there are drawbacks with this approach in terms of identifying and supporting young people considered 'more vulnerable': We've been in it only for a couple of months, straight away we're in with all these people. Brilliant ... but those people aren't always the ones who've got the kids who are more vulnerable, so it's how we then tackle that and at the moment, I don't have the answer (Mark, Prevent CSE Team Leader).
Arguably, this focus on universal education prevented practitioners from working to effectively identify those more likely to be at risk of CSE, but not quite meeting the thresholds for a safeguarding referral: And that's the bit we are not suitably skilled at knowing how to deal with. We know how to intervene once these kids have effectively been held hostage and they're trying to escape. We can fall back on our traditional methodologies at that point and they serve us quite well. In the intervention point where the risk is clear, but the young people, at least on the face of it, appear to be engaging in that abuse, we have very few tools in our artillery at the moment, to know how to work with that (Victoria, Strategic Lead).
Although elements of the suite of education and early interventions adopted by the Prevent CSE clearly align with a public health approach to crime prevention, apparently absent were those interventions aimed at the tertiary level. The multi-agency team directed attention towards the delivery of primary prevention, while those who had already become a victim of CSE were receiving interventions (secondary) from other services. Noticeably missing were interventions that responded to individuals who were considered 'more' at risk-as in the example given in the strategic lead's account above-but did not meet the threshold for a safeguarding referral on to secondary and tertiary support. This disconnection, between the aspirations of the strategic leads and the operationalisation of the prevent CSE initiative, further supports the difficulties previously acknowledged of aligning priorities within partnership approaches (Weston 2014;Davies, et al. 2023).

Data gathering, management and storage: public health or police surveillance?
Embedded within the prevent CSE team's response was the belief that intervening early can prevent some young people from being 'drawn into' the criminal justice system 'in a more serious way'. Yet, the collection and storage of data without prior consent of individuals that had been in contact with the team appeared to be standard practice. Information about young people receiving an early and brief intervention-despite having only been identified in the 'lower risk' category-were collated and stored using case management software. In the interests of having an 'audit trail', each conversation completed-including both face to face and over the telephone sessions-was recorded and logged on police information systems. Moreover, this information was subject to assessment: We put it on [case management software]. So, each session we have, the phone calls, the conversations, whatever it is, around the awareness work, I'll record it (Cathy, Prevent CSE Worker).
Practitioners of the multi-agency team often referred to this type of data gathering as 'intelligence' and, although others had negotiated understandings of their work as not being directed by data collection, emphasised that the initiative was 'a police project' and connected to police work 'on intel': It's not about intel, it's about protecting young people from the risk of exploitation, you can't move away from that. But at the same time, it's a police project and at the police service, you know, we work on intel. And if you're doing education and you're not getting much intel back, surely that's telling you something? (Sadia, Prevent CSE Worker).
While public health approaches place emphasis on the role of data collection and analysis to inform preventative activity, the language used by some of the prevent team suggests a more problematic objective-albeit relayed as secondary-that permit the use of data to identify criminal activity. To this end, the caveat made by one of the strategic leads that if the team 'lose their link to CSE panels' the initiative will become a 'police initiative with the police sent out to do work in areas dictated to by police led intelligence' comes to fruition.
The collection and storage of data in this fashion raises a range of issues relating to permission, data storage and data sharing. More specifically, it opens up portals for data to be accessed by those with relevant permissions across the force. As one of the prevent CSE team recalls below, all entries made on the case management system are stored. The rationale for this was to ensure an audit trail that would allow retrieval in the event of a future incident involving a particular individual: We have to put this information on [a case management system], because then there's a record. Obviously, if it's been called in through 101 there's a record anyway, but you do have to remind them of that… And I just say to them, you know, because it's a vulnerable child, we need to make a record that this has happened so that if anything happens in the future, or any other incidents hap-pen… it's linked (Amanda, Prevent CSE Worker).
Alongside these reflections, another member of the team talked about the possible eventuation of a future incident occurring which involved those having previous contact with the Prevent CSE team. In such an instance, the understanding was that data previously collected would be cross-referred to in order to inform future decisions and actions. The aim of this process, therefore, appeared to be about identifying those perceived to be ignoring the advice of the prevent CSE team: So in a way, you won't be able to say, well I didn't know, you won't be able to use that as an excuse -because you did know, you had an intervention and I'll make sure that this is recorded on [the case management system]. It's about me giving them information, definitely, but it is also about me saying, you can't just get away with it next time. It won't be acceptable for you to just play the 'I don't know card', or 'I didn't know that's what the law meant'. If a police incident is rung through and it's regarding a person under the age of 18, they should be able to get a record of if, that child has been known to [case management system]. So, I think if a police officer is needed … had got somebody under the age of 18 they'd be able to access [case management system] to find out had there been any work done with them (Hannah, Prevent CSE Worker).
The collection of 'intel' is a longstanding and ubiquitous practice within policing (Ratcliffe 2016) yet, when operationalised within this particular context, some barbed issues do emerge. Given that the programme evaluated was presented as a general educational initiative rooted in a public health approach, there are some questions that can be justifiably raised around transparency of purpose and the protection of human rights. While formally characterised by practitioners as 'preventative' and 'supportive', a combination of net-widening intent and data gathering under the rubric of crime prevention brings to the surface issues of monitoring and surveillance.

Reflections and discussion
In recent years, public health approaches to crime reduction have gained increased relevance, focusing attention on prevention, multi-agency working, data sharing, evidence-based practice and whole systems approaches (Christmas and Srivstava 2019). Within this framework, the police have become key players in developing preventative activities intended to both identify and reduce risk. Reflecting this direction of travel, Midshire's intervention strategy to prevent CSE comprised of primary (universal education), secondary (brief intervention) and tertiary (wraparound services) interventions. Within this arrangement, the police had responsibility for the deployment of primary and secondary level measures and, in so doing, were pivotal in identifying those at risk of becoming a victim and/or a perpetrator of CSE.
While such a set of structures and principles is defensible in theory, the way in which these public health approaches were operationalised raises a number of important issues that are worthy of further consideration. Central to a public health approach is the role of surveillance data, which provides the scientific and factual data essential to inform decision-making and appropriate action (Nsubuga, et al. 2006). Often used to monitor trends, detect irregularities and inform practice (Declich and Carter 1994), surveillance data allow practitioners to implement evidence-based interventions that are aligned to the prevalence, incidence and related risk behaviours of a given problem. However, rather than using data in a way that would focus their activity on communities that are statistically more likely to include victims, the approach adopted by the multi-agency team referred to in this paper was to assume that each and every life was at risk in some way or other. Observed was the implementation of a population level approach that deemed every young person 'at risk' or 'risky' and thus potentially in need of some level of intervention. As the data presented above illustrates, the implications of this approach are twofold, pertaining first to inconsistent trajectory and objectives and, second, to issues around transparency, data sharing and surveillance.
Firstly, the multi-agency team appeared to overlook evidence that would help to focus their intervention activity in 'the soil in which the children vulnerable to CSE breed'. Despite being a multi-agency team set up to deliver both primary and secondary interventions that would help to prevent CSE and respond to it in situations where it had occurred, the population level approach adopted by practitioners favoured interventions aimed at potential victims and/or potential offenders and those thought to be at 'lower risk', detracting practitioners' attention away from those who may actually be in tangible need of safeguarding.
Secondly, and equally as problematic, was the team's approach to data collection and storage. Aligning with a public health approach necessitated the acknowledgement among practitioners about the importance of data collection in informing and driving activity. Yet, the implementation of this approach focused the multi-agency team on the collection of 'intel'. Within the context of the CSE prevention initiatives we examined, primary and secondary (brief) interventions were thus used as apparatus to collect information, not on the basis that such information would help inform preventative activity, but rather on the basis that it may help to identify criminal activity. This approach encourages the subjectification of all young people to increasing forms of surveillance and may unduly expose them to the apparatus of the criminal justice system.
There is no doubt that public health approaches to crime prevention might be beneficial, particularly when used to identify and address the complex causal system of crime that goes beyond an offenders' motivations, intentions and character (Moore 1995). Yet, in this context of application at least, the activity adopted in Midshire to prevent CSE suggests that the operationalisation of these approaches does not readily align with achieving these goals. A public health approach suggests that a much more developed insight into the deep-rooted causes of crime, smarter use of data and working across organisational boundaries will take place but, in the case under study, there was little evidence to suggest that this was occurring. While emphasis was placed on data gathering, how the data were collected and whether transparency was adhered to in terms of informed consent for data storage and data sharing remains a moot point. Further, in order to be put to effective use, data gathered from young people should be evaluated alongside systemic factors, such as discrimination, poverty and structural disadvantage, which are acknowledged widely as salient factors associated with youth crime and (sexual) violence (Goldson and Muncie 2015). Instead, the population approach adopted by the multi-agency team, alongside the working assumptions that each and every life was at risk in some way, served to individualise, rather than be appreciative of the material conditions and inequalities that inform, shape and structure young people's behaviour.
Although often perceived negatively, the demand that public health approaches place on surveillance data that systematically report on cases may work in the interests of both care and control (Lyon 2006). Although surveillance is effectively being watched with a certain purpose in mind, which can often be controlling and disciplining the subject into certain behaviour or set of norms, it might also work to protect and care for the subject. Despite having long-established constitutional systems of government that are ostensibly committed to liberal democracy, responses to COVID-19 in both the UK and US have not only included a withdrawal of freedoms but revealed the extent to which people are willing to give up their civil liberties in favour of perceived security and protection (see Arceneaux, et al. 2020;Alsan, et al. 2020). The way in which such data collection techniques are implemented, particularly where there is an underlying project to impose certain moralistic behaviour, is vexing. Although the implementation of public health approaches more often than not demand for both the collection of surveillance data and subsequent informed intervention, such techniques can be ideationally legitimised in areas of societal concern on the basis that so-called problematic behaviour is an outcome of dysfunctional individuals who need to be identified through assessment and educated and managed accordingly to ensure both their safety and the safety of others (Kelly 2001;O'Malley 1992;Rose and Miller 1992). This process is one in which individuals become the object of government whilst also encouraging the growth of surveillance inclined towards particular groups (Rose 1999). The narrative, for example, that all young people need to be protected from the risk of sexual exploitation-and the resulting population approach that was adopted by the multi-agency team here-permits state intervention into the lives of all young people and their families. A question to be asked, therefore, is to what extent are people willing to trade off their civil liberties in the interests of a wider public good?
We would suggest, however, that the relinquishing of civil liberties should not be unlimited particularly where surveillance data are used to legitimate particular types of governance and control that discriminate between people. Given that people seem quite willing to provide data, particularly where there is a promise of safety and security as we observed during the Covid-19 pandemic, the challenge becomes how to encourage resistance within contexts where civil liberties are being compromised, but specificity about how data is used is absent. As Mariner (2007) suggests, surveillance for the common good should only be pursued without sacrificing the rights of privacy and, we would add, unnecessarily criminalising already marginalised sections of the population.
Although we would acknowledge the benefits of a public health approach to crime prevention, we have illustrated that when operationalised in a particular way and within a particular policing context, they harbour the risk of engendering iatrogenic effects that fuel both the securitisation of populations and induce responses that may not be supportive or warranted. Despite these issues, public health approaches to crime prevention are gaining momentum at the level of policy implementation, not only for the criminal justice system but also for health care and social work. Already well-established are Violence Reduction Units which have a core aim of deploying a public health approach to tackle serious violence and its root causes, and working closely in partnership to develop a more prevention-focused, holistic and coordinated approach to addressing serious youth violence. There is a need, therefore, to focus more closely on how such approaches are understood and operationalised in practice, taking into account the impact at the boundaries of traditional domains, such as criminal justice, social work and health care. As we have elucidated, although there may well be advantages to reconceptualising the complex social issues that straddle these boundaries in health terms, to do so unilaterally and uncritically runs the risk of marginalising contributions made by established fields of analysis which have traditionally informed practices, such as criminology, sociology and social policy. As Reimann (2019: 148) suggests, a public health approach may promote treatment models that become 'disentangled from socioeconomic inequalities and explained by reference to individual pathology alone'. Thus, individual behaviour becomes a priority for change, reducing broader structural factors that might be driving complex social issues, such as youth violence, radicalisation and CSE, to simply 'modulating factors'. An analysis of how public health approaches to crime prevention are being understood and operationalised, therefore, needs to also consider how, if at all, factors such as social inequalities are being addressed.

Secondly, and equally as problematic, was the team's approach to data collection and storage
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.