CONTROLLING DRUG USERS: FORMS OF POWER AND BEHAVIOURAL REGULATION IN DRUG TREATMENT SERVICES

This article examines the control practices used in drug treatment services to regulate the behaviour of people with drug problems. Drawing on an extensive qualitative study, we developed a conceptual framework, integrating the notion of responsive regulation with Wrong’s sociology of power. The picture that emerges is of a complex ‘web of controls’, combining diverse forms of power and control techniques, used to steer action and shape behavioural outcomes. It is argued that we can understand these control practices within drug treatment as part of broader strategies for the social regulation of the poor, built on deep-rooted hybrids of punishment and welfare. The article concludes with the suggestion that drug treatment represents an important site for understanding penal power today. a scary time for a service user … if they’re coming to a service, kick off, and then leave, es-corted out, police … whatever form that leaving takes, that’s a really vulnerable place for them. What happens then? Do they ever come back? Can they come back? How are they going to be received? And that’s for us to be mindful of and manage … But our parting message is still, please do come back, if you need to (Site 1, Service Manager (#103)).


Introduction
The idea of drug treatment-providing medical interventions to help people with drug problems-dates back to the mid-19th century. It developed alongside the emergence of the notion that addiction was a disease (Berridge 1979). Kerr's (1886) celebrated textbook on Inebriety sought to compile the 'state of the art' of treatment approaches, which ranged from prescribing bromide sedatives through to abrupt or gradual withdrawal. From the start, an important strand running through ideas about drug treatment was the coercive coninement of 'addicts' within either prisons, lunatic asylums or specialist inebriate reformatories, where treatment would often be an austere 'combination of food, work and religion' (Berridge 2012: 22).
Contemporary drug treatment is in some ways not so different from its 19th century forerunners (for an overview of its evolution, see Mold 2008). It remains largely a medicalized enterprise-typically funded through health budgets and/or insurance-with prescribing of substitute medication and supervised detox amongst the mainstream of approaches. Coercion is also still an important part of the treatment repertoire. The need for control can be partly explained by the fact that treatment is often an intense emotional experience, which, when combined with drug dependence and withdrawal, elevates the risk of conlict or disorder. Studies of health professionals' attitudes towards and interactions with people who use drugs show that they are generally viewed as aggressive, dishonest, impatient, manipulative and uncooperative (van Boekel et al. 2013). Stated thus, there are obvious pragmatic reasons for imposing restrictions and controls. The available evidence indicates, however, that 'the majority of drug users do not behave in this way and yet they are often tarred by the same brush' (Lloyd 2010: 32).
This raises a critical question. Why might control practices be so important within a health care setting such as drug treatment, particularly when evidence suggests that the potential for disorder may be less acute than some fear? What does this tell us about the drug treatment enterprise? One way to understand this is through the literature on the relationship between punishment and welfare. As Garland (2017: 94) observes, a foundational insight from the study of this relationship is that institutions of punishment and welfare are jointly involved as 'front-line, street-level modes of governing [the poor]'. From this perspective, it is unsurprising that drug treatment settings are characterized by both care and control, as the hybridity of penal-welfare strategies has a long history. Indeed, we can understand the birth of drug treatment as part of the wider story that Garland (1985) tells of the assembly of the institutions and ideologies of penal welfarism in the late-19th century and which Seddon (2010: 56-99) then traces through the 20th-century evolution of liberal governance. In recent decades, the conjoining of crime control and health concerns within treatment has intensiied, shaped by the transition to neo-liberalism and the ensuing criminalization of drug policy (Seddon et al. 2012: 23-45). Today, the populations in drug treatment and criminal justice systems continue to overlap signiicantly and are characterized by high rates of unemployment, low income and other indicators of socio-economic disadvantage (Jones et al. 2007;Prison Reform Trust 2017). Both systems are elements in the social regulation of the poor.
It is these control practices that are the focus of this article. They constitute a signiicantly under-researched area in the literature. There is a small body of research on associations between methadone maintenance treatment and social control, which emphasizes the repressive functions, power imbalances and institutional stigma of methadone regimes (see Neale 2013). In an unsettling account, Bourgois (2000: 167) argues that methadone 'represents the state's attempt to inculcate moral discipline into the hearts, minds, and bodies of deviants who reject sobriety and economic productivity'. However, aside from a few studies of therapeutic communities (De Leon 2000;Rawlings and Yates 2001), we know remarkably little about the mechanisms and practices of control in treatment settings, especially when we look more broadly than just methadone prescribing.
This article draws on the indings of an extensive qualitative study across a range of drug treatment services in England, including community-based and residential provision. Its analytic focus is on the control and ordering activities of drug workers that are aimed at regulating the behaviour of service users. We adopt Black's (2002: 26) definition of regulation as the 'sustained and focussed attempt to alter the behaviour of others according to deined standards or purposes with the intention of producing a broadly deined outcome or outcomes'. Our critical ethos was to conduct a concrete empirical study of practices, asking the question of how drug workers actually get things done in treatment services. In so doing, we contribute to and advance a promising area of scholarship that applies regulatory theory in the drug policy context (e.g. Ritter 2010;Seddon 2010: 100-121). This literature is principally concerned with better approaches and strategies for regulating drug markets. It has not involved signiicant empirical research, with the exception of Seddon et al. (2012: 106-117), which uses a regulation perspective to understand how courts direct drug-using offenders into treatment services and the dificulties experienced in securing compliance.
In the irst section of the article, we set out a conceptual framework for the analysis. After outlining our research methods, we turn to the indings to explore the nature, purposes and functioning of control practices. We use the building blocks from our framework-rules and principles; discretion and responsiveness; forms of power and control techniques; ladder of escalation-to structure the presentation. In conclusion, we consider the wider criminological signiicance of our indings, arguing that drug treatment represents an important site for understanding the connections between penal and social control that make up contemporary penality.

Forms of Power and Behavioural Regulation: A Conceptual Framework
For our study, we sought to develop a framework for understanding how control was exercised in drug treatment settings. The social science literature contains a rich seam on the concepts of power and control. Rather than navigating through this extensive and complex body of work at the outset of the project, we proceeded primarily on an inductive basis, attempting to build a framework up from the data, and moving back and forth between concepts and data as we sought to make sense of our empirical material.
Adopting this approach, the notion of 'responsive regulation' emerged as particularly useful for our purposes, as it 'it' closely with the control practices we were observing in our ieldwork sites. The most inluential presentation of the theory is by Ayres and Braithwaite (1992), who pulled together insights from earlier studies on corporate crime, mine safety and business regulation. The core idea is that regulators need to be responsive to the behaviour of those they are seeking to regulate and highly attuned to the context in which they are operating. Discretion is, therefore, of central importance to effective regulation. The 'regulatory pyramid' offers a heuristic that captures a key premise: the strategy of escalating the severity of interventions until compliance is achieved. The presumption is that dynamic regulatory efforts should begin at the base of the pyramid, with the most restorative, dialogue-based remedies available. Only if these persuasive measures fail, should one move up the pyramid to increasingly more punitive or coercive sanctions, culminating in the biggest 'stick' at the apex. The knowledge that escalated enforcement is the inevitable consequence of non-compliance and that the sanction at the apex exists should be suficient to drive down most of the regulatory action towards the lower levels of the pyramid.
Although the pyramid heuristic is a guide to strategies for regulatory enforcement, it is important to note Braithwaite's (2011: 491-2) reminder that this does not mean that regulation simply concerns rule compliance. In an earlier paper, Braithwaite (2002a) sets out an account of the interplay between rules and principles in effective regulation, suggesting that the more complex the phenomenon, the less useful it is to rely on rules alone. He argues for regulators to develop rules-principles conigurations that are responsive to their speciic context.
The theory of responsive regulation has been tested and reined across diverse sectors (see Braithwaite 2011). As our analysis progressed, however, it became clear that we needed to supplement the regulatory pyramid with a more detailed set of categories to describe the speciic forms of power utilized by drug treatment workers. Inluenced by its employment in some of the policing and prisons literature (e.g. Crewe 2009; Bottoms and Tankebe 2012), we turned to the sociologist Wrong's work on power, as set out in his 1979 book, which has subsequently gone through several editions. Although Wrong's work pre-dates that of Braithwaite and his colleagues, it includes a very similar notion of the 'ladder of escalation', referring to the need sometimes to move up the ladder 'from the mildest and most consensual to the most punitive forms of power in order to obtain compliance' (Wrong 2002: 71). This congruence with the regulatory pyramid encouraged us to explore how we might bring the two together.
Wrong differentiates force, manipulation, persuasion and authority as distinct forms of power. Authority is deined as 'successfully ordering or forbidding' (Wrong 2002: 35) and divided into ive subtypes: coercive, inducement, legitimate, competent and personal. These eight forms of power constitute his basic framework and are explained in more detail below. Although the forms of power are presented as discrete entities, the divisions between them are blurred in reality as 'most power relations are mixed, exhibiting qualities of contrasting types interwoven into an apparently inseparable blend' (Wrong 2002: 67). Wrong (2002: 73) argues that it is advantageous to be capable of exercising multiple forms of power, especially in exercising control over a plurality of subjects with diverse motivations for compliance.
Wrong does not organize the forms of power hierarchically but we argue that there is a tendency in practice for some forms to be drawn on more within 'softer' approaches to securing compliance and others more within 'harder' strategies (Crewe 2011;Skinns et al. 2017). Situated primarily-but not exclusively-at the base of the regulatory pyramid, are competent authority and personal authority. The former is based on specialist knowledge and skills, as in the prototypical example of the doctor-patient relationship. In drug treatment contexts, it exists when a service user complies with the advice or instructions of the drug worker out of belief in the worker's expertise. The latter involves drug workers exercising power on the basis of their personal signiicance to the service user, typically stemming from the affective bond that develops between worker and client during the treatment process (Meier et al. 2005). For drug workers who have competent and/or personal authority, service users comply with a recommendation because of its source rather than its content. Persuasion, on the other hand, the foundation of responsive regulation, occurs when the service user acts on the advice of the worker having independently assessed it in light of his/her own goals. Deliberate attempts to inluence the behaviour of another where the desired effect has not been explicitly communicated constitute manipulation. Alternatively, drug workers may appeal to the self-interest of service users by offering incentives for compliance. Authority by inducement includes the provision of services and the power to prescribe substitute drugs and what are termed 'contingency management' techniques (e.g. providing inancial incentives for attendance at treatment) (Prendergast et al. 2006).
Legitimate authority is a power relation in which there is 'an acknowledged right to command' and an 'obligation to obey' (Wrong 2002: 49). It is deployed at all levels of the pyramid but may be particularly useful when persuasive efforts have failed. A legitimate authority can induce voluntary compliance by appealing to shared norms. For Wrong, these norms prescribe obedience to the authority rather than the content of the commands issued by the authority. This resonates with notions of legitimacy and procedural justice, which, building on the works of B eet ha m (19 91) and Tyler (19 9 0), respectively, have become inluential within criminology (e.g. Crawford and Hucklesby 2013). A key inding of this work is that perceptions of legitimacy and fairness are critical to regulatory effectiveness.
At the 'hard' end of the power spectrum, corresponding to the apex of the regulatory pyramid, drug workers have coercive authority and may use (threats of) force to gain compliance. Potent means of force available to drug workers include the withdrawal of prescriptions and the exclusion of service users from treatment. If people are accustomed to, or dependent on, the resources provided by treatment services, the possibility of cessation can be employed as a threat and transform a power relation initially based on inducement into one based on coercion. 'Lighter' forms of coercion may also be built into the everyday routines of drug treatment services. Examples include mandatory drug testing and daily pickups of substitute drugs. Figure 1 summarizes how Wrong's taxonomy of forms of power maps onto the regulatory pyramid. Conceptually, we can understand the pyramid as describing the strategy for behavioural regulation, whilst the forms of power are the speciic techniques deployed to execute that strategy.

Research Methods
Fieldwork was undertaken in six drug treatment services in three regions of England between October 2011 and July 2012. Drawing on indings from a national survey of Drug and Alcohol Action Teams (DAATs) (Bacon and Seddon 2013), we selected regions where various techniques of behavioural regulation appeared to be widely used and then sites within each of these regions to capture the spread of service providers and treatment modalities. Our focus was on residential and structured communitybased services. Table 1 provides a breakdown of the research sites and participants in the interview strand of the research.
The study employed a mixed-methods design consisting of three interconnecting elements of data collection. First, service user handbooks, various written agreements and other documents were analysed to identify the principles, rules and sanctioning processes of the treatment services. Second, the main strand of the methodology comprised 108 semi-structured interviews with staff members and service users. Interviews explored a wide range of topics relating to behavioural control. We interviewed service managers and frontline practitioners. Service user interviewees were selected purposively to ensure a range of views were represented and that there was a balance in the sample in terms of age, gender, ethnicity, treatment history and type/severity of drug problem. There was a mix of users subject to court orders and those there 'voluntarily', although the latter formed the substantial majority. All interviews were audio recorded and transcribed with each participant assigned an anonymous unique identiier. Third, over 300 hours of direct observation were carried out in waiting rooms, communal areas, group sessions and staff meetings. This method provided a fruitful way of deepening our understanding of staff-client interactions and the everyday operation of power and regulation in treatment settings. Interviews provided an opportunity to ask about observed incidents.

Regulatory goals: rules and principles
In each of the research sites, there were systems of rules and principles in place to regulate behaviour. Following Braithwaite (2002a), we conceive rules as speciic prescriptions and principles as prescribing unspeciic actions that can justify rules and bring about mutual orientations. Service provider 'rules'-which were interchangeably referred to as 'boundaries', 'expectations', 'guidelines' and 'responsibilities'-were contained within the terms and conditions of entry into treatment. To gain access, service users had to agree to abide by the rules and accept the consequences of non-compliance. This involved the signing of various contract-like agreements and is best conceptualized as a form of contractual governance (Bacon and Seddon 2013). It implies a reciprocity in which the service user is required to assume responsibility for their own treatment, a strategy of 'responsibilization ', which O'Malley (1992)   2 Public sector community-based service located in a borough of London 4 5 3 Voluntary sector community-based service located in a borough of London that not only specializes in alcohol treatment but also provides a 12-week abstinence-oriented day programme for people who use drugs 8 10 4 Voluntary sector therapeutic community located in a city in the West Midlands that provides a 12-week residential rehabilitation programme 6 14 5 Public sector community-based service for female sex workers located in a city in the West Midlands 8 19 6 Voluntary sector community-based service located in a city in the West Midlands that provides a 12-week aftercare day programme 5 11 Tot a ls 38 70 Rules were codiied in a range of documents. Staff explained that these documents-'code of conduct', 'house rules agreement' and 'service user charter'-incorporated national policy and clinical guidelines. They had also often been developed in consultation with service users to address the speciic needs of the service. Such engagement could improve regulatory design, enhance the legitimacy of rules and encourage empowerment (Van Hout and McElrath 2012). The responsibilities placed on service users were set out alongside their rights and what the service committed to providing for them: [T]he charter … is very much about a dual responsibility, an expectation from both sides … It's about what we will do, what they can expect from us, and what we would expect in return (Site 1, Service M a n a ger (#10 3)).
In addition to the right to treatment-which acted as an inducement-service users had the right to be dealt with in a respectful manner, to receive full information about their treatment, to be involved in decisions affecting them and to have their views listened to. This commitment to fairness-in effect, to procedural justice-helped generate a form of legitimate authority.
'Cardinal rules' prohibited behaviour that caused actual or potential harm and could result in legal proceedings against the service. They included no violence, no stealing, no discrimination, no breaches of conidentiality and no illicit drugs in or around the premises. The 'no drugs' rule covered use, possession and supply, as well as organizing deals, exchanging information about dealers and 'inappropriate' drug talk.
When someone comes in intoxicated you are not going to issue the script because it puts them at risk of overdose. It's that simple and that's a boundary that's not going to get shifted … [I]f someone did go and have an overdose we wouldn't have a leg to stand on (Site 2, Drug Worker (#110)).
'General rules', on the other hand, contained a mix of principles and expectations relating to the day-to-day operation of the service. For example, service users were expected to behave 'in a manner that is respectful and consistent with the therapeutic approach', 'cooperate with staff' and 'engage fully with all aspects of your treatment package'. They were also expected to attend appointments on time and provide urine, saliva and/or breath samples as requested. As is characteristic of residential services, Site 4 was markedly different in terms of the volume of general rules. They covered everything from dress code, meal times, telephone calls and visits to restrictions on literature, music and television.
Rules were designed to create a safe environment, impose conditions of order and structure engagement with the treatment programme. Staff and service users frequently commented that, without the rules, the service would descend into 'chaos'. Moreover, the rules acted as a form of communication system, which provided service users with a guide to social values, norms and expectations both within and outside treatment settings. As De Leon (2000) suggests, by internalizing models of pro-social behaviour, drug workers hoped that service users would make positive changes to their lifestyles, take active responsibility for their futures and become functioning members of society: What you're doing with all this, beyond the methadone, beyond the group work, and everything, you're grooming people for reality. Now, if they think they can go out and talk to people in the Job Centre, or their boss, or a policeman like that, their feet aren't going to touch [the ground] (Site 1, Drug Worker (#98/9)).
[The rules] are about the best way to live in harmony in a house … inding your role in society and being able to contribute something back. So all the rules are about that really: that you don't hide under your hat; that you spend time with your peers eating; that you spend time mingling with them rather than just watching the telly and mumbling to each other; that you get up at a set time so you're not interrupting group by walking in late; that you're clean and keep the house clean and make an effort (Site 4, Service Manager (#92/3)).
Although some described them as 'common sense', service users tended not to immediately grasp the purpose of all the rules when they irst entered treatment, not least because they were feeling somewhat disorientated, vulnerable and inundated with information. To help develop their understanding, rules were explained as part of the induction process, embedded within key work sessions and discussed during group activities. As service users engaged with the programme, embraced the justiicatory principles for rules and experienced positive treatment outcomes, the majority recognized the ways in which compliance contributed to their recovery: I can understand people thinking, 'Oh, bloody rules! It's all rules, rules, rules!' I can understand that because we're all like that the irst week … [T]he longer you're here, you get to realise, 'Well that's why there is that rule. That's why there is that rule' (Site 4, Service User (#28/9)). I just see a routine, no rules; I don't see it as a rule system, I see it as an opportunity to fucking sort my head out, sort my life out and stay off drugs (Site 3, Service User (#64/5)).
Service users rarely complained about the rules or contested their legitimacy (though they did occasionally 'kick off' when staff withheld methadone on the grounds of lateness or intoxication). Only the highly regimented restrictions of the therapeutic community were called into question on a regular basis. Residents who took issue said they were 'childish', 'dictatorish' and in need of explanation: When you eat, you can't leave your table for like … half an hour -not even to clear your plate … You have to sit there, at your table … sometimes it's like jail. And I've got a bad back, so just sitting there at the table all the time, that just does my back in … I think that's one ridiculous rule (Site 4, Service User (#38)). Some of the rules, you know, if there'd just be like … an explanation just to say we're doing this because, you know, it's a lot easier for people to accept it. Because you can, if you're feeling a bit paranoid, which some of us do from time to time, you can think they're doing that out of spite (Site 4, Service User (#37)).

Discretion and responsiveness
There are often disparities between the rulebook and how rules are applied in practice. To understand how rules operated in drug treatment services, therefore, consideration was given to the discretion exercised by staff and the various factors that shaped their decision-making processes. Discretion was also central to the power relations and sanctioning processes discussed below. What we are concerned with here are the 'rules of thumb' that enabled drug workers to 'make contextual judgements as to when it is safe to get the eficiency beneits of following the rule and when it is not' (Braithwaite 2002a: 68). Our indings are very reminiscent of Ugwudike's (2010) research on how probation oficers enforce community sanctions, illustrating the 'family resemblance' between penal and social control.
Service managers acknowledged that rules and enforcement practices were lexible. Discretion was necessary because drug workers were expected to 'know the client group', understand their needs and be responsive to the unique circumstances of individual transgressions: We always have to look at what was going on for them that day, because girls come in here who've been attacked the night before, their heads are a mess; so you have to know the client group and know why they have behaved in a certain way (Site 5, Drug Worker (#44)).
As long as they operated within 'professional boundaries', staff members had the autonomy to decide when and how to make allowances or challenge non-compliant behaviour: I'll put up with someone that threatened to cut my throat but if they try to sell drugs, no (Site 1, Drug Worker (#81)). Each staff member here's very, very individual, works in very, very different ways … Some are very 'right, this is the line, you will not cross it' sort of thing. And that's not criticising anybody; they all work in massively different ways, but everyone seems to have great outcomes (Site 6, Drug Worker (# 4 5)). Some are quite strict and will come in and say it's 11 o'clock, the television's going to go off, it's bedtime. And some will come in and say well, you know, I'm really sorry guys, it's 11 o'clock, you know I've got to turn this television off … And some will sit at the desk, and when the ilm ends, they'll come and say I didn't realise the time … It's down to the individual, that consistency; I ind varies quite a lot (Site 4, Service User (#23)).
In addition to training and experiential learning, the varying norms and values of drug workers affected how they exercised discretion. Take the following interview excerpt, e.g., in which the service manager is explaining her approach to challenging service users about their language: I think if you don't allow people to use the service because they swear a lot or because their manners are different from yours then you are making huge judgments on individuals that actually need support and that it isn't our place to dictate to somebody how they should be in relation to the way they speak or the way they behave unless they're harming somebody else (Site 6, Service Manager (#48)).
For this manager, people 'can only be themselves' and treatment services should aim to be places of 'total acceptance to everybody who walks through the door'. Aside from racist, sexist and homophobic language and verbal exchanges that were intimidating or caused offence to others, she would only pull service users up for swearing if they had chosen to address it as part of their care plan.
Drug workers emphasized the importance of discretion when dealing with new entrants and service users with chaotic lifestyles, learning disabilities and/or mental health issues. Relapses and using on top of substitute medications were viewed as part of the process. Flexibility was, therefore, imperative in order to lower initial dropout rates and allow service users to adjust to the rules: You can't start giving [service users] rules, cos they'll just rebel against it and they'll just tell you where to stick it and you'll never see them again (Site 6, Service User (#11)).
When deciding on how best to act, staff would take into account whether the service user in question was a irst-time, occasional or habitual rule breaker. They would also assess their motivation and readiness to change: There's times when people, e.g., have had a drink a couple of days ago and want to come into our relapse prevention group and in theory they shouldn't have had a drink for ten days … But then you've got somebody sitting there who wants to come to a group and needs support. Now what do you do? Do you kick him out and say no, no we're going to hold that; sometimes you might. If you know the person you think well actually you've been doing this a lot. Somebody's just started on that road, you can see they're really trying, you know; it's not always black and white. So you need those rules, yes, but you need commonsense to apply them (Site 3, Drug Worker (#59)).
Staff had more discretion when addressing minor incidents and enforcing general rules. Cardinal rules, on the other hand, were strictly enforced because the prohibited acts could make the environment unsafe or undermine the treatment process. Nevertheless, while the scope for discretion was limited, we observed discrepancies in enforcement practices. This is evidenced in the following examples, in which staff explain why they decided not to uphold the zero tolerance approach to drug offences: We have had a client once, go to give another client some drugs … I didn't ban her, because it's that kind of street mentality that they're trying to help somebody else. She can't get her script, 'Here you are, have some gear', you know what I mean? (Site 5, Service Manager (#50)). If someone's clean or trying and they've got cannabis on them, due to the intense smell that cannabis gives off, I would rather give them an airtight container to put it in than let them walk down the street and be arrested, or let people give them funny looks, or ask them whether they've got any on them, be approached. It's just over safety (Site 5, Drug Worker (#42/3)).
Although discretion was seen as necessary for providing personalized care and dealing with incidents in a responsive manner, drug workers strove to adhere to the rules and follow procedure so as to avoid inconsistencies or differential treatment. In reality, however, these unintended consequences were inevitable. Disparities in the application of rules left some service users feeling uncertain about what was expected of them or as though they had been treated unfairly. Perceptions of favouritism or 'special treatment' could lead to arguments, complaints and resentment. In this sense, there was a continual tension between responsiveness and perceptions of legitimacy, as Irwin-Rogers (2017) also found in the context of the penal system in his study of post-custodial supervision in Approved Premises.

Forms of power and control techniques
Drug workers exercised multiple forms of power to regulate the behaviour of service users. Although drug workers viewed their role as fundamentally deined by the provision of care and 'collaborative capacity building' (Braithwaite 2011: 475) through support and education, it was also acknowledged that treatment contained a signiicant element of control. Furthermore, whilst drug workers did not explicitly identify as controlling authorities, they recognized that there was a power differential between themselves and their clients. Some described the performance of care and control functions as a balancing act: I don't want a service user to get in their head 'Oh, I'm going to come in here and they're going to be my mates and it's ine', because they're not going to get anything from that … I think there's a ine line; I want them to come here and be happy, enjoy it and be comfortable and feel safe, but at the same time they need to be developing … So it's inding that balance without being too much of a dragon (Site 6, Drug Worker (#45)).
Striking the right balance between 'mate' and 'dragon' was a crucial factor in successful engagement and dialogic forms of control at the base of the regulatory pyramid. Indeed, service users regularly attributed much of their progress to relationships with staff members and how they were treated: I really believe that fucking I've ended up relating with staff members, that's what's made me decide whether I'm going to throw myself into this or not. Like I says, I can sign whatever they wanted to give me to sign [code of conduct], and it don't really matter. It really does matter how you're welcomed when you walk through that door (Site 6, Service User (#09)).
Service users valued drug workers who 'don't judge you', are 'prepared to listen if you need to talk' and will 'go into bat for you if you play ball'. Such workers were present in all of our research sites. In Site 2, however, which had up to 600 service users on the books, drug workers had heavy caseloads, which meant they spent less time with their clients and found it harder to built rapport and establish dialogue. It also meant that meetings were sometimes preoccupied with paperwork and overly instructive in nature: A few people I've met working in the service … didn't give me the feeling that they really want to help. Not a relation with friendship of course, but really listen to you and follow your progress … [T]hey're just illing charts: 'You're screening okay. Got your sample or not? Okay so next time I'm going to take your script off you'. Sometimes it just seems that they get bored with talking … It's 'Fill that, ill that, why don't you do that' or just like reading a book and telling you (Site 2, Service Us er (#113)).
Working at the base of the regulatory pyramid, drug workers had competent authority, derived from their professional knowledge and skills. Service users complied because they believed in the worker's ability to decide which actions would best serve their interests. This type of authority is evident in the following interview excerpt-mixed with legitimacy-in which the service user is relecting on being told by a staff member that he was not allowed to wear his baseball cap indoors: You've got to respect their knowledge and, you know, they're doing it for your beneit, although you can't actually understand how they're doing it, but you know they're doing it for your beneit … [I]t's gratitude as well; they're treating you with respect by helping you, so you know, the least you can do is treat them with respect and stick by their rules (Site 4, Service User (#37)).
In practice, competent authority tended to be combined with persuasion and the involvement of service users in decisions about their treatment and behaviour. Drug workers provided service users with advice, which they then assessed and acted on in light of their own goals: A lot of the work that we do is about that … therapeutic relationship between you and the client, them trusting you and you guiding them -not telling them, but guiding them -where they want to go (Site 5, Drug Worker (#44)). Yeah, he's my key worker … he talks to me and he says, like, he gives me advice, but the rest is up to me, isn't it? He can't do nothing about it; he could talk to me all day long, it's up to me if I listen, or I don't (Site 2, Service User (#114). The next quotation demonstrates how a service manager used a blend of competent authority and persuasion to censor service users. When compared to the above example on the question of swearing, it further reveals how responsiveness and discretion almost inevitably lead to inconsistent practice: They're always told not to swear when I hear it. And then we had a massive debate last week about, 'Well, why? It's my tongue, I'll speak what I like, and that's how I express myself'. And I said: 'But I ind it offensive and unnecessary. More to the point you just don't sound very good; you don't sound like you're coming across very well.' And from there I'm taking it even a step further, like an NLP [Neuro Linguistic Programming] step, of if you can start to control the way that you speak and how you come across, you're going to start to control and have power over your triggers, your cravings, your behaviours (Site 4, Service Manager (#92/3)).
When a therapeutic relationship had been established, it was possible for drug workers to shape the behaviour of service users through personal authority. Many service users said they 'look up to' members of staff. Owing to this affective bond, breaking the rules felt like they were letting their drug worker down. These feelings could be harnessed to gain compliance. However, exercising 'tough love' had to be carefully managed as it could have negative emotional effects and increase the risk of relapse. Here, we can see how a drug worker used personal authority in response to a positive drug test and the impact it had on the service user: It was the D-word. And she went in, she said -she started crying -'I'm so sorry'. I said, 'I'm just so disappointed with you'. It's 'disappointed' -they can't bear it. She was sobbing (Site 5, Drug Worker (# 8 8)). I just think they've done so much to help me, so by letting them down I'm letting myself down and if I let myself down then I feel shitty and I'll probably go out and smoke (Site 5, Service User (#18)).
Persuasion was also supplemented with authority by inducement. In fact, it could be argued that inducement is an inherent part of care planning and key work sessions in that drug workers are able to inluence how service users go about realizing their goals. A powerful incentive for compliance in prescribing services (Sites 1, 2 and 5) was the prospect of moving from the supervised consumption of substitute drugs on a daily basis to take-away doses. A system of 'privileges' was used in Site 4 to reward residents for their length and level of engagement. Privileges included outside visits and a choice of house duties. Rewards were widely used in the form of praise and the celebration of success, which Braithwaite (2002b: 24) suggests have 'unequivocally positive effects' on compliance. For those who completed treatment, there was also the possibility of applying to ambassador schemes: It's an incentive to get clean. I can be an ambassador in a year and people will see me clean… If I can get this under my belt people are going to be thinking, 'Fucking hell … if he's done it, there's hope for us all' (Site 1, Service User (#56)).
In a broad sense, manipulation was widespread in treatment settings in that drug workers manipulated the ield by acting on information, rewards and deprivations (Wrong 2002: 28). Their understanding of the importance of fairness, however, meant that intentions were not deliberately concealed from service users in an attempt to manipulate them. Nevertheless, the reasons behind rules, decisions and procedures were not always explicitly or uniformly communicated, which could result in misunderstandings and complaints: The only problem I've got like with the system is, the rules ain't passed down right. You don't have a member of staff going 'Come here you lot, this is a new rule'. We hear half off residents who've overheard a staff member, and you're trying to glean what this new rule is or whatever. And, you know, the irst time you know about it is when you're getting told off for it (Site 4, Service User (#37)).
If efforts to exercise control through these techniques failed, drug workers would refer back to the documents service users signed upon entering treatment. These acted as leverage to persuade service users to honour their 'contractual' obligations. They also reminded them about the prospect of sanctions and provided drug workers with legitimate authority to take enforcement action. From a staff perspective, contractual governance was beneicial because it enabled them to distance themselves from the intervention and thereby maintain the therapeutic relationship: The thing that you don't want is client/practitioner breakdown of trust … You want to be able to put it back on policy. So you don't want to say … 'I as a key worker don't think you are doing what you should be doing therefore your ile's going to be closed'. It's about saying, 'Hey look, really sorry Bob, but to go back to the stuff that we signed up to at the start, you did agree it's part of my job as part of the policy … to come back and look at this and it seems that you're not changing. I'd love to keep you here forever but I can't' … So if they do come back into the service there's no hard feelings about you (Site 3, Drug Worker (#63)).
Drug workers had a range of sanctions at their disposal for dealing with non-compliance. Above all else, coercive authority was based on the ability to stop providing service users with services. This is starkly illustrated in the following interview excerpts, which show how the prescribing of medication could be utilized as a control technique as punitive as any to be found within the penal system: Yeah, there's elements of a power relationship sometimes between the key worker and the client; for instance, I've got the power if I don't give you your script to make you feel very ill or to make you have to go out and ind £20 through crime (Site 2, Service Manager (#112)). 100ml of [methadone] is quite a lot of juice to be on every day. Then they just take it away from you. It's sort of like they've got you over a barrel … I don't think it should be like that (Site 2, Service User (#111)).

The ladder of escalation
Models of escalating sanctions were remarkably similar across the research sites and bore the hallmarks of the regulatory pyramid: It's like being at school, I guess, in some ways. You give them a little bit of a warning, and then you kind of escalate that warning till they understand (Site 5, Drug Worker (#87)).
It was expected that service users would breach rules. The rationale for applying sanctions was to help service users to learn from their mistakes and develop the ability to comply: If you don't have consequences, or if nobody tells you, how are you supposed to learn? [T]hat's what we say anyway, when we do give warnings, we give them the opportunity to talk about it, and to offer them support into the reasons why that actually happened (Site 6, Drug Worker (#46)).
In the next interview excerpt, a service user is talking about receiving a sanction for lateness. He was outside having a cigarette, lost track of time and missed the start of a group session. As this had happened once before, he was required to design and run a session for the house on the consequences of being late. The excerpt opens with his initial frustration but goes on to emphasize the value of self-relection and peer learning: Oh, my head fell off, I growled, I thought, you know … three minutes and you're going to make me do that; at the time, I was proper pissed off and I told them about it … But I went ahead and done it and it made me feel great afterwards, because I started the meeting and I wrote down, for instance, you know, consequences of being late: I missed the chemist one day, I couldn't get my methadone, because it was a weekend, so the irst thing I had to do was use drugs to mask, you know, the methadone, because I couldn't rattle all weekend. So that was a consequence of that; I had to go back on drugs and that going back on drugs led me to keep on smoking again … [T]hen everybody else had to give their for instance and, like, everybody else give a different scenario… I can remember one saying, oh, he was late to sign on, he never got his money, he got kicked out of his hostel … [I]t was a really, really interesting group (Site 4, Service User (#80)).
Verbal warnings were used to interrupt minor incidents, remind service users about the rules and signal that more serious sanctions existed: I believe warnings and stuff like that are there to sort of snap people, like, whatever it's out of, denial, or whether it's to wake people up, this is what you're doing, and a chance to put them back on the straight and narrow (Site 6, Drug Worker (#49)).
Written warnings were a step up the ladder of escalation for failure to heed verbal warnings or more serious infractions. Such warnings typically took the form of a personalized letter to the service user, which communicated the reasons for the warning, explained the consequences and stated that any reoccurrence would result in further sanctions. The increased formality of the procedure was designed to exert authority and send a more forceful message about rules and sanctions: As soon as you hand them that piece of paper saying 'Due to your behaviour …' they instantly become remorseful. Because sometimes they've never had boundaries or rules before, and because they actually think of us as family, if we issue them with that, then suddenly it kicks them up the arse and makes them realise that no, we don't tolerate that (Site 5, Drug Worker (#42/3)).
In four of the six research sites (2-5), behaviour contracts were used as an alternative to discharge when service users continued to break rules. They were not a regular feature of regulatory efforts as compliance was usually achieved through persuasion and warnings. Contracts documented the incidents and concerns held by staff and explained the link between the behaviour that gave rise to the sanction and the associated consequences. They also set out speciic goals and were accompanied by more intensive one-to-one support: We use [behaviour contracts] as a way … to outline what it is that we're noticing is a problem and what we are aiming to do together to support that … So it should almost be like a last chance contract, in a sense, because it really … these come up as basically the last straw for them and, yeah, it's almost like setting out quite a speciic plan for them that if they don't keep then we take that as maybe they're not so motivated (Site 3, Drug Worker (#76)).
Contracts were generally supplemented by further regulatory devices. The following example describes the use of an 'honesty journal' for a resident who had abused the privilege of outside visits and committed an act of deception by trying to smuggle chocolate into the service: He's to do a relective piece, an honesty journal … [I]t's starting from the point where he's given the contract, how he felt about the contract. Because I know at the time he thought it was all being blown out of proportion. So it's really looking at how he felt at the time, and then looking at the reasoning the staff gave him, how he feels about them … He's kind of accepted that there is an issue, but we wanted a bit more evidence on how he's working on that. And for this individual it's been decided that he will work on this journal for the rest of his time here (Site 4, Drug Worker (#91)).
In the services where contracts were not used in this way (Sites 1 and 6), service managers explained that, following warnings, problematic behaviour was addressed through care planning and key work sessions. They viewed 'last chance' contracts as 'putting up barriers' in that they 'could be a bit of a stick to beat somebody with', which was at odds with their notion of the therapeutic alliance.
Discharging service users was a last resort when efforts to gain compliance by other means had failed. Service users could also be immediately discharged without prior warning for breaking cardinal rules. Staff often found the decision dificult because they were aware of the risks associated with cessation of treatment. In most instances though, discharge actually meant a ban period and referral to another service rather than permanent exclusion: That's a scary time for a service user … if they're coming to a service, kick off, and then leave, escorted out, police … whatever form that leaving takes, that's a really vulnerable place for them. What happens then? Do they ever come back? Can they come back? How are they going to be received? And that's for us to be mindful of and manage … But our parting message is still, please do come back, if you need to (Site 1, Service Manager (#103)).
Besides verbal warnings, sanctions were decided on a case-by-case basis following deliberation between staff. Moreover, the inal decision would not be reached until staff had met with the service user in question. The aim of the meeting was to determine the nature and implications of the incident. This involved investigating the circumstances around the incident, whether there was any provocation and the impact on other individuals, as well as the stage in treatment, care needs, level of engagement and attitude of the service user. Honesty and expressed regret inluenced decisions about the response: We had someone here once … it was time for her to go out for the day with her family, or whoever she was meeting. She rang up and said she had used … and it was decided she could stay because she was honest about it. Yet on another occasion, a guy went out -he was supposed to be going to a [Narcotics Anonymous] meeting, I think -used, came back, didn't tell anyone, and then it came out that he had used, and was asked to leave (Site 4, Drug Worker (#94)).
Another key purpose of the meeting was to provide the service user with an opportunity to voice their views and ensure that they understood what the intervention meant and why it was taking place. They were asked how they felt about the handling of the situation before they were asked to consent to the sanction: You don't want residents thinking, well, this is unfair. They still will to a degree, but it's about us also looking after ourselves as staff. There needs to be reasoning behind why we're giving someone a written warning. They need to be open and everyone needs to be clear why this is happening (Site 4, Drug Worker (#91)).
Service users did occasionally express dissatisfaction when they were given sanctions. Some said they were disgruntled because they thought the rules being enforced were puerile or they disagreed with the interpretation of events. Others simply said they had 'problems with authority'. Feelings of resentment emerged when service users were given different sanctions for the same or similar rules violations. Lack of transparency was an issue here. While decisions about speciic cases may well have been based on just grounds, the responsibility to protect client conidentiality meant that other service users were not privy to the factors that contributed to the outcome. Without an understanding of the process, ostensibly reasonable decisions could be perceived as unfair and inequitable.

Conclusion
Our focus in this article has been on the control practices used in drug treatment settings. This focus should not be interpreted as casting doubt on the value of the care and support provided to drug users in treatment nor on the integrity and professionalism of drug workers. Indeed, during our ieldwork we saw countless examples of compassion, empathy and care that were inspiring and remarkable in equal measure. Nevertheless, our study also found that power and control were central to the everyday routine interactions within drug treatment services.
The overarching picture that emerges from the study is of a complex 'web of controls' or 'control space' in which the behaviour of drug users in treatment is governed. The strategies and practices of control amounted to a form of regulation that combined a complementary mix of diverse forms of power (Wrong 2002) and control techniques to steer action and shape speciic behavioural outcomes. Deployed in a broad framework of responsive regulation (Ayres and Braithwaite 1992), the exercise of power sought to facilitate not only control and order maintenance but also care and support. This hybridity might be usefully conceptualized as what Gomart (2002) termed 'generous constraints', i.e., those which aim to induce positive action rather than being solely aimed at oppression. Further research could interrogate the implications of such power and behavioural regulation for theories of addiction, treatment practice and the processes underpinning recovery (see e.g. West and Brown 2013;Fomiatti et al. 2019). Regulatory theory could also be used to enhance understandings of drug user motivation and engagement with treatment.
The ways in which drug workers balanced care and control functions were remarkably similar to those found in penal settings. In Fielding's (1984: 109) classic study of probation practice, e.g., he demonstrates that '[c]ontrol is not incompatible with care, and, indeed, it can be seen as a part of the caring process. Caring becomes the oficer's overall stance, control a technique forming an element of it'. Our integrated conceptual framework-bringing together Wrong's sociology of power with the theory of responsive regulation-provides an original analytical toolkit that can be used for research across all sites where human actors seek to regulate the behaviour of others over whom they have some authority.
It is perhaps a banal observation that the power and control we have described would not occur in the same way in many sites of health care. The key difference with drug treatment settings is self-evident: the social proile of its client base. In this sense, the strategies for controlling drug users in treatment can be understood as a form of social control and class discipline delivered through health care. As Garland (1985) has shown, this type of fusion of punishment and welfare is built into the historical origins and foundations of the welfare state, underpinning strategies for the social regulation of the poor. Drug treatment is, therefore, an important site for exploring the contemporary restructuring of welfare states (see DeVerteuil and Wilton 2009), and this is one lens through which to view, e.g., the complex contemporary landscape of treatment service commissioning and provider conigurations. In a recent review of theoretical advances in the sociology of punishment, Garland (2018: 16) argues that thinking about punishment and welfare and about penal control and social control, together, as linked phenomena is an important step forward in our understanding of penality in its broadest sense. Again, drug treatment potentially offers a particularly fruitful arena for exploring and making sense of these connections. We suggest, in conclusion, that Garland (2018: 23-25) himself perhaps misses a trick when he discusses the 'war on drugs' only in terms of penal and criminal justice. The casualties of that war appear in drug treatment services as much as they do in prison cells, and the deliberate inliction of pain is to be found in the former just as compassion is in the latter. A research agenda which embraces this idea will uncover drug treatment as a critical site for transforming our understanding of penal power today.

Funding
This work was supported by the Economic and Social Research Council (RES-062-23-3039).