WHAT IS THE HARM IN ADDICTION? AUTONOMY, VULNERABILITY, AND THE CASE FOR HARM REDUCTION DRUG POLICY

Abstract

Successful drugs policy must be driven by thoughtful principle and intergovernmental consensus, not by departmental or legal inertia, nor by public (mis)conceptions about drug use. Perhaps the most pressing choice for drugs policymakers at present is between harm reduction and abstinence approaches to drugs policy. To choose between these two approaches, we need to know addiction's normative status: is having an addiction a misfortune or a harm in its own right, even setting aside knock-on health and wellbeing consequences? We argue that the harm of addiction is driven by poor policies, but that harm is not inevitable.

I. Introduction

Since the mid-twentieth century, criminal law has been at the core of governments’ regulatory approach to drugs in most, if not all jurisdictions globally. Drug use, however, is by nature a public health issue in the first instance: criminal law is used as a vehicle for advancing public health aims, while also addressing the fallout from criminalisation itself, such as black-market-associated violence. Inevitably, other areas of law and policy must be engaged to handle (what, because of criminalisation, is) illicit drug use, since many issues arising from the use of illicit substances simply cannot be handled as matters of crime. In the UK, for example, responsibility for drugs policy sits with the Home Office, who have competence in criminal matters; but other government departments, such as the Department of Health and Social Care, and the Department for Education also must play a role in the overall strategy.¹

A policy framework that cuts across disparate fields of law, expertise and responsibility runs the risk of tensions, even legal vacuums, with different departments having different outlooks, aims, priorities and competencies. Furthermore, in federalised states like Canada and the US, and states with devolved systems of government like the UK, subsidiarity structures can compound these tensions. In the UK, for example, Scottish regional health authorities are currently in conflict with UK national criminal law in their efforts to tackle drugs deaths; just as British Columbian health policy was in tension with Canadian federal criminal law a decade earlier over the same issue.²

If drugs policy is to be coherent and successful, it must be driven by thoughtful principle and intergovernmental consensus, and not by departmental or legal inertia, nor by public (mis)conceptions of the problems or the aims. Perhaps the most important question for drugs policymakers in the UK and elsewhere at present is what role, if any, the criminal law should have in drugs policy aimed at promoting public welfare. More concretely, the question is whether states should pursue a harm reduction or a prevalence reduction approach to drugs policy.

Harm reduction strategies seek to reduce the harms associated with drug use without aiming to reduce overall drug use.³ Needle exchange programmes for people who inject drugs are probably the most well-known harm reduction strategy. Others include pill testing at festivals, drug consumption rooms (DCRs) and prescription heroin-assisted treatment (HAT). By contrast, a prevalence reduction approach – often framed as an “abstinence” or “recovery” approach, as in the 2010 and 2017 UK Drugs Strategies – seeks to reduce the number of people using drugs, primarily through punitive measures like drugs criminalisation.

One of the primary conceptual issues in need of resolution in order to assess the choice between harm reduction and prevalence reduction approaches is that of the normative status of addiction. Is addiction a misfortune in its own right, even setting aside knock-on health and wellbeing harms? Addiction is, after all, one of the primary things we typically seek to avoid when proscribing drug use: we must stop people from using drugs altogether, rather than simply make drug use safer, so the thinking goes, in order to save people from becoming addicted to drugs. In this paper, we seek to explore the question of how to understand addiction. What, if anything, is bad or harmful about addiction itself?⁴ Answering this question is necessary to guide good drugs policymaking – and importantly, to help clarify whether harm reduction or prevalence reduction should be our policy aim. If addiction itself is necessarily harmful even when separated from the many health and social harms with which it is associated, then that is a reason to prefer prevalence reduction over harm reduction, and an important factor to consider when designing drug policies.⁵ We will focus, illustratively, on UK drugs policy, but our normative question, and the wider question around harm reduction, is applicable to and in need of attention in virtually every jurisdiction.

II. Background

Harm reduction was once a defining feature of UK drug policy.⁶ For example, the unique “British method” of treating persons with heroin addiction, from the 1920s onwards, was to prescribe injectable heroin, first through private doctors, and later (from 1968) through NHS drug clinics.⁷ The UK was also a trailblazer in the area of needle exchange programmes, with 15 facilities by 1987 and around 300, nationwide, by the early 1990s.⁸ However, with a series of legislative acts in the 1990s and early 2000s that functioned to increasingly amalgamate health and criminal law responses to drugs⁹ – and more so with Cameron coalition government policy from 2010 – the prevailing winds have shifted away from measures like needle exchanges and HAT towards a straightforwardly prevalence reduction paradigm, embracing “complete abstinence” as its primary health goal. Whilst some harm reduction measures are still in place, and some new grass-roots initiatives have been granted short-term approval in recent years, government policy is now focused squarely on getting people off drugs – even prescription therapies like methadone treatment – to the detriment of measures aimed at minimising harms like overdose deaths.

Drug overdose deaths have risen dramatically since 2010.¹⁰ Responding to this increase, the Advisory Council on the Misuse of Drugs (ACMD) recommended a raft of harm reduction measures in their 2016 report, including further investment in methadone and other opioid agonist therapies and a lifting of non-evidence-based completion targets;¹¹ a scaling-up of take-home Naloxone provision;¹² and re-instatement of central funding for HAT.¹³ None of these recommendations was taken up in the 2017 UK National Drug Strategy.

In 2019, the House of Commons Health and Social Care Committee made a similar set of recommendations – urging government to consider measures such as drug consumption sites, drug checking services and a review of existing drugs legislation.¹⁴ Government's 2021 reply reiterates that abstinence-directed treatment “provides value for money, improves public health and reduces crime” – and promises a forthcoming “UK-wide Addictions Strategy”; harm reduction recommendations are essentially ignored.¹⁵ This policy dispute reflects a deeper disagreement about the harms of addiction.

Addiction is clearly linked to many significant harms, both in the UK and abroad. In England alone, over 77,000 people died of tobacco-related illness in the period 2016–19, and there were more than half a million hospital admissions related to smoking in 2018–19.¹⁶ Elsewhere, more than 500,000 Americans die of a tobacco-related illness every year.¹⁷ More than 70,000 Americans died in 2020 from an opioid overdose; for England and Wales in the same year, the figure is 2,263 and for Scotland alone, 1,192.¹⁸ It is clear from the astonishing health toll of problematic drug use that being addicted is, for many people, a very bad thing. But is being addicted to a substance in and of itself – setting aside preventable, separable health consequences – a form of harm? If it is, then addiction might give us a reason to prefer a prevalence reduction approach over or at the expense of harm reduction strategies.

We can see the importance of this question by considering the example of Opioid agonist therapy (OAT), a harm reduction strategy that stands in opposition to a strict abstinence approach. In OAT, patients take an opioid agonist (substitute) – usually methadone or buprenorphine – under medical supervision, instead of heroin. The effects of these opiate alternatives last a whole day or more, unlike heroin, which takes one on a rapid upswing and then rapid decline within a few hours. OAT drugs are also safely and legally produced with standardised dosing. Switching from heroin to OAT makes it possible in principle for people who use drugs (PWUD) to lead a more stable, productive, healthier life. Among other things, patients are able to structure their days around work, family, leisure, rather than around acquisition of typically four-hourly doses of heroin. They can work towards leading a normal, productive life even despite continued dependence on opiates.¹⁹

In the UK, successive governments have made it a goal to shrink OAT patient numbers by avoiding it in the first place or by rapidly moving patients off it. On this line of thinking, we should be curing people of addiction, getting them “clean” and off drugs entirely; simply moving a patient from one opiate addiction to another, regardless of improvements to wellbeing, is not addressing the problem and should not be an endpoint in itself.

OAT and harm reduction advocates, of course, take a different view. OAT works. It saves lives and allows people to be productive and well. It gives PWUD a chance to join mainstream society and move past the social isolation, poverty, homelessness and poor health that so often accompany chronic heroin addiction. OAT patients do not need to have “cleanliness” as the end goal; stability on OAT is a good end in itself. Advocates for harm reduction measures like OAT point out that the harms of addiction are overwhelmingly harms that are attendant on addiction, rather than harms of addiction itself. PWUD face many risks, including death from overdose, blood-borne diseases from shared needles, lack of housing and medical care, criminal prosecution, stigmatisation and social exclusion. Many of these harms can be avoided without reducing drug use, or even rates of addiction.

Opponents of harm reduction policies, however, argue that addiction itself is a harm, distinct from the knock-on health and social harms that can accompany it. OAT by itself does not address this harm. Opponents therefore see programmes like OAT as, at best, a temporary and transitional measure that leaves a significant harm of drug use (i.e. addiction) entirely unaddressed. They therefore reject harm reduction policies in preference for abstinence-focused policies, on grounds that only abstinence can prevent the distinct harm of addiction.

Whether or not addiction itself is harmful, then, is a normative question with important policy implications. Which end goal – freedom from addiction or mere freedom from the separable consequences of addiction – is preferable depends on what value we assign to addiction, and this matters to how we design drug policy. So in order to settle the debate between abstinence advocates and harm reductionists, we need to get a handle on the normative value of addiction, setting aside health and social harms. Is addiction a harm in itself? Would it be bad to have an addiction even if poverty, ill health, social isolation did not accompany it? Is it simply antithetical to wellbeing to be in perpetual need of a drug? If so, why?

The salience of this question is by no means limited to OAT or to the UK context. Whether addiction itself is a harm bears on policy questions surrounding promotion of e-cigarettes and other nicotine replacement products for smoking cessation; the efficacy of safe stimulant supplies; the ethics of managed alcohol programmes; drug testing as a condition of access to public services like housing; DCRs; the decriminalisation or legalisation of drugs, and a whole host of other drug policy issues. Nor is the question limited to the debate between supporters of harm reduction and more prohibitionist, criminal justice-based approaches. It is also relevant to debates internal to harm reduction about such topics as the design of OAT programmes and other “safe supply” approaches, and the extent to which treatment and recovery services should be integrated into harm reduction programmes such as DCRs and supportive housing. If addiction is among the central harms of drug use, then eventual abstinence, rather than safer drug use, should be the ultimate goal of all harm reduction policies. So the status of addiction matters for thinking about the appropriate relationship between prevention, treatment, recovery and harm reduction. We focus here on the goals of OAT because it is here that the answer to this question about addiction is most obviously pressing.

We consider the question of the normative import of addiction by exploring two potential harms. First, we consider the argument that addiction is a harm because it undermines autonomy. We argue that much of the autonomy harms standardly associated with addiction are separable harms: that is, thoughtful public health and drug-regulatory policy could reduce or eliminate those harms without reducing drug use or addiction. Second, we consider addiction as an avoidable form of vulnerability that can itself constitute a harm.

After explaining the ways in which addiction can render someone vulnerable and exploring the nature of vulnerability, we argue that not all forms of vulnerability should be construed as a harm, and that addiction vulnerability should be understood as a vulnerability that is only as bad as we make it. While vulnerability is a more promising framework for understanding the potential harms of addiction, the magnitude of the vulnerability attendant on addiction is, as with autonomy harms, largely dependent on policy decisions.

The upshot of our argument, then, is that the harms of addiction itself are both less significant and less common than opponents of harm reduction policies often assume. In fact, these harms are separable from addiction itself, and with improvements in drugs policy many cases of addiction need not feature these harms in any substantial way. We conclude by considering what lessons public policy should draw from this view of the harms of addiction.

III. Separable Harms

At the heart of the objection to harm reduction we are considering is the distinction between the harms that are merely “associated with” drug use and addiction, on the one hand, and the harms of addiction itself, on the other. Harm reduction policies might be effective in reducing the harms associated with drug use, but (or so goes the argument) since such policies do not have abstinence as their ultimate aim, they cannot be effective in addressing the significant harms of addiction itself. But what does it mean to say that a harm is merely “associated with” addiction, as compared to being a harm of addiction itself?

One tempting approach is to distinguish between harms that are directly and indirectly caused by drugs. Those that are indirectly caused by drug use would then be the harms that are merely associated with addiction, while the harms that are directly caused by drugs would be counted as the harms of addiction itself.

It is certainly true that policy interventions can reduce the indirect harms of drug use. Where drugs are illegal, PWUD face the risk of criminal prosecution and are forced to participate in a black market that leaves them open to exploitation and violence. Imprisonment, assault and exploitation are therefore harms that are indirectly caused by drug use, but are more directly caused by the socio-political conditions in which drug use occurs. Such harms can be reduced by reforming the laws governing drugs. The same is true of some of the other significant harms associated with drug use. Both HIV and homelessness, for instance, are more common among PWUD, and these harms are often linked to drug use.²⁰ But HIV is not caused by heroin; it is caused by injecting heroin with infected needles. Drug addiction does not lead to homelessness without poverty and inadequate social programmes playing a role. Part of what makes these harms indirect, then, is that they are mediated by the social and material conditions in which drug use occurs. Policies that intervene in those social and material conditions can reduce the indirect harms of drug use without reducing drug use itself.

The benefits of harm reduction are not, however, confined to these indirect harms of drug use. Even those harms that are the direct physiological result of drug use can be reduced or even eliminated without reducing the extent of that drug use. Consider perhaps the most serious and obvious example of a physiologically direct harm from drug use: death by opioid overdose. This is a direct harm of drug use, but it is also a separable harm of drug use. OAT dramatically reduces the risk of death by overdose, even though people prescribed OAT remain dependent on opioids. DCRs in which people can inject heroin in the presence of trained staff also significantly reduce overdoses and overdose deaths, in addition to reducing the transmission of infectious diseases.²¹ Naloxone is an opioid antagonist that near-instantly reverses the effects of overdose; distributing it to PWUD can prevent overdose deaths.²² Each of these three programmes can significantly reduce the most significant direct physiological harm of drug use without reducing drug use or drug addiction at all.

In our view, the morally relevant distinction is not between direct and indirect causal pathways of the relevant harms, or whether the harm is a direct physiological effect of the drug. Rather, the relevant distinction is whether or not the harm is separable from drug use and drug addiction.

We have argued here, however, that many of personal, social and health harms that are associated with addiction – including the most significant direct physiological harms – are separable from habitual drug use. If so, then this undermines the most common objection to harm reduction policies. But perhaps our argument overlooks the importance of a potentially inseparable harm of addiction: the loss of autonomy.

IV. Autonomy

A. Addiction as Slavery

The idea that addiction is a form of slavery that strips away autonomy, freedom and control is common to many descriptions of addiction, both among people who struggle with addiction and those who do not. It is at the heart of the 12-step movement, and it is a frequent cultural trope, used by figures as divergent as Tupac,²³ the Pope²⁴ and the French public service.²⁵ This association between addiction and slavery might suggest that addiction and dependence itself is among the most significant harms of drug use. It is not just that the person who uses heroin risks overdose and disease, or that someone who smokes regularly has a life expectancy that is 10 years shorter than a non-smoker, though these are of course real harms.²⁶ Given the high value Western societies place on individual autonomy, the idea that people with substance use disorders are not free may be the most significant and frightening harm of addiction.

According to the objection we are considering, then, there is no way to be fully autonomous while dependent on a drug, and this represents a significant and inseparable harm of drug addiction that harm reduction strategies are powerless to address. If so, then harm reduction will only ever be a second-best, band-aid solution that leaves the most significant harms of addiction entirely untouched, and should only ever be adopted when other, more promising approaches have been shown to fail.

The overall justification of harm reduction policies therefore depends on the extent to which this autonomy objection can be answered. If it cannot, then this is a reason to prefer abstinence-based policies. Whether or not addiction does inevitably undermine a person's autonomy is therefore a question of considerable ethical and practical importance.²⁷

B. External Autonomy

We believe this objection is misguided. To begin, it is worth noting that harm reduction programmes often promote the autonomy of PWUD. This is not an accidental by-product of a programme aimed at preventing death or disease. Harm reduction programmes typically place a high value on respecting the autonomy of PWUD, and aim to engage with them as equal partners, rather than passive patients.²⁸

We can see concrete evidence of this in several ways. Harm reduction programmes such as OAT, “safe supply”, and DCRs all aim – in different ways – to protect PWUD from the criminal justice system. OAT and safe supply replace illegal drugs with a legally regulated product, while DCRs provide a space for people to use drugs free from the fear of arrest. Harm reduction activists who advocate for the legalisation or decriminalisation of drug use have a similar aim in mind. A criminal conviction can significantly reduce a person's autonomy. Being confined to prison reduces the autonomy of those convicted of crimes, but so too does a term of probation, which can place significant constraints on a person's freedom of movement and association. Harm reduction programmes that promote legal ways of consuming drugs therefore protect the autonomy of PWUD.

Legal and regulated sources of drugs such as OAT and prescription injectable heroin can enhance autonomy in other ways as well. Someone who uses heroin might need to find and consume the drug many times a day. Since OAT is longer-lasting than injectable heroin and is designed to be taken once a day, it gives people the freedom to organise their lives around activities of their choice rather than searching for and taking drugs.²⁹ While prescription injectable heroin might need to be taken several times a day, the supply is stable and secure, which means that people do not need to spend many hours a day “grinding” to earn money required to pay for drugs. This gives people more choice and control over their lives. Having secure access to a reliable supply of drugs, then, can be autonomy enhancing in ways that have nothing to do with avoiding arrest or imprisonment.

Homelessness and addiction are closely linked.³⁰ Housing First programmes are aimed at people experiencing homelessness who also struggle with mental health or addiction. In contrast to traditional “treatment first” approaches, Housing First reflects the principles of harm reduction by not requiring abstinence from drugs or alcohol to secure stable independent housing. Such programmes reduce the harms associated with addiction and homelessness,³¹ but their main benefit is that they significantly improve housing stability.³² This is important in large part because stable access to housing plays an important role in securing individual autonomy.³³ Housing expands the capabilities – the freedoms and genuine options – open to people, while homelessness dramatically restricts those capabilities.³⁴ Someone without a home – a space that they can control – is in an important respect radically unfree, since they are always in principle at the mercy of others.³⁵ There is nowhere that they are simply allowed to be without the permission of others. In this sense, as Jeremy Waldron puts it, “homelessness consists in unfreedom”.³⁶

It is clear from these examples both that addiction can significantly undermine the autonomy of PWUD, and that many of these autonomy harms of addiction are in fact separable from drug use, because we can see the many ways that harm reduction can restore autonomy without ending addiction. That suggests a direct response to the objection with which we began; far from sacrificing autonomy in order to reduce health harms, a concern for autonomy is central to harm reduction. If we care about autonomy, then this is a reason to support, rather than oppose, harm reduction policies.

But perhaps this conclusion is too quick. Perhaps implementing effective and systematic harm reduction can enhance the autonomy of PWUD relative to the status quo ante by giving them much more control over their lives. But comparing the autonomy of PWUD with and without harm reduction might not be the most relevant comparison. Maybe we should instead compare the relative autonomy of PWUD under the ideal set of harm reduction policies to the autonomy of people who do not use drugs at all. After all, the objection we are considering is that harm reduction does not do as well as more abstinence-focused approaches at restoring the autonomy that is lost through addiction. We might find that addiction – even in the context of systematic harm reduction programmes – inevitably compromises a person's autonomy when compared to an existence entirely free of drug dependence. If so, this would mean that even the autonomy-enhancements of harm reduction fail to protect fully the autonomy of PWUD.

C. Internal Autonomy

To see why addiction might inseparably undermine autonomy, we need to turn to a different conception of autonomy. To not be able to do what you want because you are in prison, or homeless, or compelled by the threat of violence, is to have your autonomy limited by external forces. There is also, however, an important internal or psychological dimension to autonomy. To be autonomous in this sense involves more than just having a range of options; it is to be in control or the author of one's own actions. This autonomy can be constrained if one's actions do not reflect one's choices and values, and it is often suggested that addiction undermines this internal form of autonomy. That is the point of the slavery metaphor; it is not that PWUD are under the control of another person, or that their options are radically constrained by economic and political forces; it is that their own desire for drugs enslaves them. The source of their unfreedom is ultimately internal, not external.

1. Autonomy as control

Just what this internal dimension involves is the subject of considerable debate. There are at least two broad approaches; the first links autonomy with control over one's actions, and the second with self-expression. Addiction features prominently in the philosophical debates over both accounts.

On the control account, autonomy involves the ability to exercise genuine control over our actions,³⁷ and the problem with addiction is that it involves compulsive and irresistible desires that the person is powerless to control.³⁸ The desires, rather than the PWUD, are in control and the person cannot stop taking drugs no matter how hard they resist. If this were an accurate account of the reality of drug addiction, then it would seem as if drug dependence really would undermine a person's autonomy; someone who is literally powerless in the face of their desires to use drugs is not acting autonomously.

The problem is that this is a mistaken understanding of the nature of addiction. We can see evidence of this in several ways. First, many people with substance use disorders do eventually stop using drugs. Rates of substance use disorder peak in early adulthood and decline with age.³⁹ For some, addiction can be a life-long condition in need of constant monitoring, and it can even end in untimely death. Frequently, however, addictions simply resolve – use eases off or ceases over time without outside intervention, even in the case of “hard” drugs like heroin.⁴⁰ Firm figures are difficult to come by, since most addiction research takes addiction treatment patients as its subjects, thereby missing so-called spontaneous recoveries. Estimates of the proportion of individuals who recover from alcohol or other drug addiction without formal treatment, however, range from 26 per cent⁴¹ to “most”.⁴² Further, whether treated or not, as Heyman (2013) puts it, “most addicts quit” – the vast majority within a single-digit number of years.⁴³ Some stop using through clinical treatment or 12-step programmes, some do so on their own, but people who are dependent on drugs are often able to resist the desire to use, control their drug-taking and stop using. If the desire to use drugs were literally irresistible, this would be surprising.

Second, chronic, long-term dependence is significantly more common in people who suffer from additional psychiatric disorders.⁴⁴ Hannah Pickard hypothesises dependent drug use is often a “way of managing the severe psychological distress typically experienced by patients with comorbid psychiatric disorders and associated economic, social, and relationship problems”. Rather than being overwhelmed by irresistible compulsions, many people with psychiatric disorders “use drugs purposefully: to alleviate severe psychological distress. Consumption is the chosen means to desired ends”.⁴⁵ On this view, drug use – even dependent drug use – is often part of an autonomously chosen self-medication strategy to deal with psychiatric distress. The strategy might be imprudent or harmful – though it might also be the best of a bad set of options – but that is not at all the same as being non-autonomous.

Third, even among people with significant substance use disorders, drug use bears all the marks of intentional, reason-responsive behaviour. Securing a supply of drugs and finding a place to use them often requires complex planning, and among even heavy users with substance abuse disorders, rates of drug use are cost-sensitive. For instance, “contingency management” programmes are among the most effective treatments for substance use disorder. Participants in such programmes are given vouchers and other rewards for negative drug tests; the experimental evidence strongly suggests that such programmes are more effective than control at promoting abstinence.⁴⁶ There is also experimental evidence that habitual users of cocaine will choose small sums of money over using drugs, even when the money is to be paid in the future and the cocaine is immediately available.⁴⁷ These results would be hard to explain if addiction involved irresistible desires that the person who uses drugs was powerless to control.

Even if we accept that autonomy involves the exercise of self-control over one's actions, then, this does not show that addiction undermines autonomy, because drug dependence is compatible with control. However, it could be that a different account of the internal dimensions of autonomy would be more successful at explaining why addiction is autonomy-undermining.

2. Autonomy as self-expression

A competing model of autonomy identifies it with acting in ways that express our “deep selves”, or our values, or our rational commitments. This involves the ability to step back from and reflect on our desires to ask whether they express our values. Different theorists offer different accounts of this; it might involve asking yourself whether you endorse those desires, or whether you want to have them, or whether they are compatible with your stable long-term plans and policies, or whether you take them to be good reasons for action.⁴⁸ But common to many accounts is the idea that autonomy involves only acting on desires that are truly ours. Even if the desire to use drugs is not compulsive or irresistible it might still not be autonomous, because it does not express the genuine values of the person who has it.

In fact, it is common for philosophical accounts of autonomy to use addiction to make this very point. The philosophical character of “the Addict” acts out of a strong desire to take drugs, even though they do not identify with the desire or endorse it, and even though they might fervently wish not to have it.⁴⁹ It is in this sense that addiction might be thought to undermine a richer internal conception of autonomy. The problem is not that the desire for drugs is irresistible, but rather that the person rejects it; they do not see it as expressing their own values.

Let us assume that this is an accurate characterisation of both the nature of autonomy and the reality of drug addiction. Even so, this still does not show that addiction necessarily and inseparably undermines autonomy. To see why, assume that the person with a substance use disorder has a strong desire to use drugs that they do not endorse and would strongly prefer not to have. If they nevertheless use drugs, then on this view they lack autonomy. But to know whether this lack of autonomy is an inevitable and inseparable harm of drug dependence, we still need to ask why they would prefer not to have the desire to use drugs.

If the desire to stop using is driven by a desire to avoid disease, death, exploitation, police harassment and homelessness, and by a desire to regain a degree of control over their lives, then this desire can be satisfied while continuing to use and be dependent on drugs. After all, a desire to avoid the harms of drug use can be satisfied in at least two different ways. First, by not using drugs. Second, by using drugs in a social and material context where those harms can be avoided. As we have seen, harm reduction strategies help people to avoid those harms, and so can enhance the internal, self-expressive autonomy of PWUD. In other words, someone with a substance use disorder might not endorse their desire to use drugs precisely because they recognise that satisfying that desire, given the social, material and political context in which they live, leads to harms that they feel powerless to avoid. But this powerlessness is not an inevitable feature of drug dependence; it is a separable consequence of policy choices.

The person receiving OAT, for instance, remains dependent on drugs. But their dependence – their strong desire for drugs – is much less disruptive and harmful, and more easily integrated into their lives, than that same dependence would be if they only had access to illegal, unregulated “street” heroin. So the person taking OAT might be able to endorse their desire to take drugs, since satisfying that desire is no longer incompatible with other things they care deeply about. The fact that they do not endorse their desire to use drugs when that desire is extremely harmful does not show that they would not endorse it were drug use much less harmful. Harm reduction can enhance the autonomy of PWUD in external ways by keeping them out of prison, offering them a safe and stable supply of drugs, and providing them with stable housing. But it can also enhance their autonomy in this richer internal sense as well, by allowing them to endorse their desire to use and to integrate it into their lives.

To be clear, we are not arguing that addiction never undermines autonomy. Nor are we claiming that successfully implementing systematic harm reduction programmes would make all drug use fully autonomous. Some PWUD clearly do have abstinence as a goal and experience their drug use as a constraint on their autonomy, and this would no doubt remain true for some even if governments systematically and successfully adopted harm reduction drug policies.

Our claim is simply that autonomy impairment is not an inevitable and inseparable harm of addiction itself, because the sense in which it leads to impaired autonomy is, for many PWUD, largely a function of the clearly separable harms of drug use. Drug dependence can impair a person's autonomy, but it does not necessarily do so.

Within the context of the debate over the goals of drug policy, what matters is that whether – and how much – addiction undermines autonomy depends a great deal on the social and material context of drug use, and this context can be affected by drug policy. Moreover, this is true of both the external and the internal forms of autonomy. PWUD can be given much greater control over their lives through well-designed policies without having to reduce or eliminate their drug use. No policy approach will be able to ensure that all people are fully autonomous; there are simply too many external and internal barriers to autonomy for any set of public policies to be able to ensure that complete autonomy is available to all. Instead, the question before us is what basic approach best protects and promotes the autonomy of PWUD.

We have argued that prevalence reduction strategies, particularly when linked to prohibition, carry significant costs to the autonomy of PWUD. We have also argued that harm reduction policies can enhance their autonomy without reducing their drug use. If we are right, then opponents of harm reduction drug policies who worry about the autonomy costs of addiction have significantly over-emphasised the autonomy costs of harm reduction while at the same time over-emphasising the autonomy benefits of prevalence reduction. So a concern for autonomy need not be a reason to reject harm reduction in favour of abstinence-focused approaches; in fact, a concern for autonomy can be a reason to support harm reduction. If addiction involves inseparable harms, the problem lies elsewhere.

V. Vulnerability

In this section, we turn to what we take to be a more promising framework for understanding the harm of addiction. We will argue that mere addiction is best understood as a form of vulnerability. Since vulnerability is both relational and contextual, the moral valence and magnitude of addiction are neutral and nul, respectively, in the abstract; and only take shape in and from a particular context, under particular circumstances. This means that mere addiction, in the abstract, should not factor into our moral reasoning about drug use; addiction itself does not make PWUD vulnerable in a harmful way. Rather, it can be harmful in certain contexts, and our thinking about drug policy should be informed by the often complex and subtle ways in which addiction can lead to vulnerability.

When someone becomes dependent on a substance, she thereby acquires a new susceptibility to harm: inter alia, the harm of going without that substance and suffering the physiological and psychological effects of that lack – a lack which can sometimes have severe and traumatic physical, affective and practical consequences. Whether she actually suffers this lack – or alternately, can access the substance on which she is dependent – is largely out of her control. She is dependent on the actions of others, and on law and policy, market forces and luck. Even if the stars align for her and she is able to reliably access the substance, she still experiences excess vulnerability as a result of her addiction, since there is always the chance that her luck will change – that policing priorities will change; that prices will rise; that she will be displaced, made redundant, incarcerated or have her vulnerability otherwise compounded. All humans, simply in virtue of being mortal, are vulnerable; but a person who is substance-dependent suffers an additional layer of vulnerability. And this seems like a harm to care about.

A. The Nature of Vulnerability

To understand the import of vulnerability, it is important to first identify what it is – and is not. We will argue that vulnerability is relational and contextual, and that it can vary along two (orthogonal) axes. It can be more or less extreme, and it can separately be morally bad, morally neutral or even morally valuable.

It is common in medical contexts to talk about vulnerability as an attribute of certain individuals or groups: we refer to “vulnerable populations” in considering ethical issues like informed consent, for example. But this is a convenient short-hand that can obscure the nature of vulnerability. Vulnerability is relational. It describes a relationship between the vulnerable individual and her environment and the other actors in it. Joel Anderson's working definition is apt: “a person is vulnerable to the extent to which she is not in a position to prevent occurrences that would undermine what she takes to be important to her. … [V]ulnerability is thus a matter of effective control, understood as a function of the relative balance of power between the person in question and the forces that can influence her.”⁵⁰

Vulnerability is also contextual. It can be present in one domain but not another; or on multiple fronts that interact with one another. A person may be vulnerable in educational settings, say, yet not with regard to safe housing; or vulnerable with regard to health in a way that compounds her vulnerability to unfair employment practices. One might be more or less vulnerable within a given domain. A housed person living in an area with a high prevalence of acquisitive crime may be vulnerable to theft, while an unhoused person living in that same area will be much more vulnerable to theft.⁵¹

Most of us are more vulnerable, relative to our vulnerability in other contexts, when we step into a doctor's surgery, where our understanding of the complexities of our care is often minimal and dwarfed by that of our professional carers: we are, perhaps more than we would like, at the mercy of the doctor's expertise. While many of us are physically vulnerable in the face of physically powerful nefarious actors, all of us are physically vulnerable in the face of nefarious actors with guns.

On this understanding of vulnerability, it becomes clear that it is a pervasive feature of the human condition. We are all vulnerable in some ways, contexts and times, to some degree. It is not strictly true, then, that there are clearly delineated “vulnerable populations” that can be neatly contrasted with “us” – with non-vulnerable “normal” people. What sets “vulnerable populations” apart is, first, the degree of vulnerability, rather than its existence; second, the moral salience of that vulnerability; and third, that the context of vulnerability is relevant to policy, either because it is the result of policy, or because it has the possibility of being exacerbated or alleviated by policy.

So, when addiction is conceptualised as a vulnerability, what this means is that in being dependent on a controlled substance, a person has an additional layer of vulnerability. She is unable or less able to prevent a particular sort of occurrence that may harm her. Specifically, she lacks control over her own access to the substance on which she is dependent, and in turn is susceptible to morally salient harm that may attend lack of access, and that vulnerability is in turn sensitive to policy influences. Our question is whether this vulnerability is a reason to prefer abstinence-focused drug policies to harm reduction ones.

B. The Moral Valence of Vulnerability

Even acute vulnerability may be morally non-problematic and not an apt target for policy interventions. The moral valence of vulnerability varies along a continuum, with morally problematic vulnerability at one end (e.g. that of a lone migrant child in a deportation hearing), morally-praiseworthy vulnerability at the other (e.g. the vulnerability we practise in love relationships); and relatively morally neutral vulnerability somewhere in the middle (the vulnerability one experiences going for a swim in the ocean, say). Guidry-Grimes and Victor describe morally problematic vulnerability as the “disadvantaged placement of an individual within the context of social practices, which translates into threats to the agent's holistic well-being. … an individual is vulnerable [in this way] when he or she is in a position that threatens his or her ability to develop and achieve the most fundamental dimensions of well-being”.⁵²

In other words, whether primarily accounted for by internal features of the person, external features of the context, or both, the excessively vulnerable person is disadvantaged, disempowered in ways that threaten her wellbeing. This in turn can constitute a state of affairs that is morally wrong and in need of remedy.

That said, vulnerability, even excess vulnerability, is not morally problematic simply in virtue of being a vulnerability. Some vulnerability is morally valuable and contributes positively to well-being. Most notably, vulnerability is an inextricable aspect of the sort of openness required to form and experience meaningful interpersonal relationships. One cannot experience love, among other things, without experiencing vulnerability.

Likewise, as Anderson points out, it is often necessary to be vulnerable to rejection and exclusion to achieve self-realisation, agency and thus autonomy. Anderson argues that self-trust, self-respect and self-esteem are crucial for agency, since without them, one cannot know one's mind, trust one's judgment and ability, and thus form and enact plans to reach one's own aims.⁵³ One's sense of self and trust in oneself is partially constituted by – developed with – the recognition of those in our families, communities, friendship circles and practical orbits. Without these social relations of recognition, developing the sense of self necessary to act as an autonomous agent is at best difficult, and at worst impossible. Such relations, in turn, necessitate making oneself vulnerable to denial of respect, of mutual love and to rejection.⁵⁴

The sort of vulnerability required for love and for autonomous agency shows that vulnerability qua vulnerability does not have a moral valence: the simple fact of being vulnerable is morally neutral. More needs to be said in order to establish that a given vulnerability is morally bad, or an apt target for public policy.

At first glance, addiction might seem much closer to the clearly bad vulnerable migrant child end of the spectrum, rather than the clearly valuable vulnerable lover end. In the popular imagination, of course, addiction is a life-long downward spiral that inevitably ends in overdose or death from drug-related complications in a dark alley somewhere, homeless, friendless, helpless. (We have all seen this movie.) So it is easy to assume that substance-dependence vulnerability is a necessarily bad vulnerability to have.

People with a dependence on an illicit substance often do suffer poor health, isolation, joblessness, homelessness, stigma and moral condemnation from their peers and communities, and the risk of incarceration. For some, addiction can be a life-long condition in need of constant monitoring, and it can even end in untimely death. But as we pointed out above, many addictions simply resolve – use eases off or ceases over time without outside intervention. Disordered substance use does not lead inevitably to a lifetime of harm and suffering, not only because addictions resolve, but also because whether and to what extent people with addiction suffer in the ways described above is highly sensitive to the support, care and accommodation they receive, both through policy initiatives and community and familial support. Because of the mismatch between how everyday people tend to picture addiction and how it actually looks, it is important to consider addiction on its own, stripped of all its many separable harms. What should we make of the mere vulnerability that one experiences with regard to acquisition of the drug on which one is dependent?

Vulnerability, again, is relational: “a function of the relative balance of power between the person in question and the forces that can influence her.”⁵⁵ The vulnerability that addiction represents, then, is not merely a feature of the person with substance dependence: it is a feature of the interaction between that person and the world around her.⁵⁶

The lone migrant child's vulnerability is morally problematic not merely because the child is alone, or because she is a child, but because she is alone and a child in the face of a comparatively powerful legal-political entity that is structured so as to deemphasise her wellbeing, and she faces the possibility of a negative outcome that she is powerless to affect and that could be catastrophic for her. Her vulnerability depends on the imbalance of power between herself and relevant agencies and the wider system of laws and policies.

So too for the person with drug dependence. Her vulnerability to loss of access to the drug is not internal to her; it exists in the complex relationship between herself, the state and her society. The extent to which she is vulnerable to lack of access to the substance on which she is dependent depends on (inter alia) policies surrounding access; enforcement of those policies; social norms around access; inclusion in or exclusion from privilege-conferring communities and resources; as well as her own economic, physical and social power.

The mere fact that her access is dependent on these things, and that she, in turn, is dependent on access, constitutes a vulnerability, and one additional to what we often take to be the “normal” human condition. But this mere fact, on its own and stripped of contingent particulars, constitutes a vulnerability too negligible to warrant moral attention or a policy response. It is, after all, of a piece with the need for any necessity that is not procurable on one's own in a “state of nature” – medical care, public infrastructure, social support, etc. This sort of vulnerability is only of note when wider conditions are such that access is called into question. Our vulnerability in the face of need for water, for example, is hardly even worth picking out as a vulnerability in the normal run of things in well-functioning, high-income areas. On the other hand, for a person living in Nibinamik First Nation, Ontario or in Flint, Michigan this vulnerability is pressing and bad.⁵⁷ There simply is no fact of the matter, with regard to the moral status of water vulnerability, in the abstract.⁵⁸

As for water, so for illicit substances. It might sound strange to compare addiction to cocaine, for example, to the need for hydration, since after all, water is very unlike cocaine. However, in the context of threat of going without, the comparison seems apt. While the vulnerability has no moral valence in the abstract – it is neither good nor bad outside of a particular context⁵⁹ – there are things we can say about the likely direction of spread of the moral status of a given vulnerability. Going without water – actually being denied it, and indefinitely – is itself a grave harm, one that is guaranteed to result in death in days or weeks. So even fairly minor water insecurity is probably a significantly bad thing. All the same, the mere fact that one needs water is neither morally bad nor morally good in the abstract: there is nothing bad nor good about the physiological fact of needing water to survive any more than there is anything good or bad about the fact that most humans need to blink. We just do; it just is.

Meanwhile, while cocaine may be significantly more dangerous than water, the lack of it is not, even for those dependent on it. Death from drugs withdrawal is rare and unlucky, whereas death from water withdrawal is guaranteed.⁶⁰ If water vulnerability is neutral in the abstract, so surely is addiction vulnerability – since in virtually any particular circumstance we can imagine, the potential harm of water vulnerability makes it more clearly problematic than that of addiction vulnerability.

Further, As Ben-Ishai points out, many of us are dependent on licit drugs – prescription drugs – and this dependence, in decent medical circumstances, does not seem to constitute the same sort of vulnerability that addiction to an illicit substance does.⁶¹ But the difference in vulnerabilities is not a difference in the risk of harm from going without; after all, a lack of access to many medicines can be fatal, while withdrawal from illicit drugs typically is not. Rather, the difference in vulnerabilities is a difference in secure and stable access to the drug in question.

For someone with reliable, affordable access to healthcare and medicines, dependence on pharmaceutical drugs might simply be an annoyance that constitutes no significant threat to wellbeing. Someone dependent on heroin, by contrast, is vulnerable to adulterated drugs, interrupted supply, violence and exploitation, and incarceration, all of which can constitute an ongoing threat to well-being even when they are able to access a supply of heroin. The salient difference between dependence on licit substances and dependence on illicit substances qua vulnerability is that in the normal course of things, access to needed licit substances is usually facilitated, whereas access to needed illicit substances is sharply circumscribed or even prohibited, and carries (policy-driven) knock-on risks.

The comparison with medicines – that is to say, with licit drugs – is revealing in another way as well. When someone needs medicine to treat a chronic condition, we tend not to think that the aim of the treatment should be to get them “clean” of medicine, or that their need for medicine represents a lamentable moral failure. The need clearly justifies the provision of the medicine. With illicit drugs, however, there is an unfortunate tendency to see the need for the drug as justifying polices that restrict access in ways that heighten, rather than alleviate, the person's vulnerability and so their risk of harm.

In some very real sense, then, the vulnerability of addiction in the abstract is not morally problematic: it is simply neutral. Of course, we do not live in the abstract. We live in a particular world, under particular conditions. In our particular, actual world, the vulnerability that addiction constitutes is a bad thing, one we should care about and seek to remediate. But the fact it is neutral in the abstract shows us two things.

First, it shows us that the badness of addiction vulnerability is not inevitable.⁶² It might have been otherwise, had the forces that can influence one's access been different. This shows that even if addiction is a harm, the magnitude and import of that harm is changeable without changing the fact of addiction. Second, we – we society, we policymakers, we voters with views – hold the reins on the moral valence of addiction vulnerability. How we conceptualise and respond, morally and practically, to addiction shapes the moral nature of it. And third, it highlights a disingenuousness of the prioritising of addiction alleviation over more immediate, pressing harms like drugs overdose: even if we think the current vulnerability effects of addiction are very bad, we should not prioritise the alleviation of it over very bad harms like death, since we have it in our power to make addiction vulnerability less bad than it currently is.

Addiction, then, can fruitfully be understood as an additional layer of relational, contextual vulnerability. Such a vulnerability is neutral in the abstract. Its badness in actual fact depends on power relationships in which the vulnerable person finds herself with regard to the relevant vulnerability layer; as well as wider contextual features (legal, social, policy). In the actual world, drugs policy and law what it near-universally is, this vulnerability is a pronounced and morally problematic one. But it need not be so. While no policy or set of policies can fully alleviate vulnerability – vulnerability in one form or another is just a fact of being human – We have a choice, as a society, about the extent to which mere addiction is a misfortune for the one who suffers it. And if we care about the harm of addiction, this implies that we should craft compassionate and caring policies and practices that reduce the badness of this vulnerability.

VI. Implications for Policy and Law

Law and policy are necessarily at the core of our approach to illicit drugs, given their status as both health hazard and tradeable commodity. Whilst criminal law is the primary locus for drug control in most jurisdictions, our arguments point to a refocusing of law and policy around harm reduction measures.

Much of the opposition to harm reduction policies seems to stem from the belief that harm reduction is a sticking plaster approach to drugs harm, since it cannot reduce or eliminate one of the most significant harms of drug use: addiction. We have shown that this is a mistake. Mere addiction is not necessarily harmful. Addiction itself need not undermine autonomy or increase vulnerability in problematic ways. This means that addiction is not a reason to reject harm reduction approaches. Put another way, it cannot serve as a justification for foregoing opportunities to avoid more grave harms. Furthermore, to the extent that addiction does present separable harms, harm reduction approaches⁶³ are best situated to reduce or remove those harms.

Our account of addiction has clear implications for UK drugs policy, and below we spell out three of these. First and currently most pressing, in line with Nicholls et al,⁶⁴ it is clear that the UK government must help end Scotland's drug-related death crisis by allowing Scottish government to institute harm reduction measures such as DCRs.⁶⁵ While addiction may, in the right (bad) circumstances, lead to health and wellbeing harms, these harms simply do not compete with the ultimate harm of preventable death; and can be minimised through harm reduction measures such as DCRs.

In its January 2021 reply to Health and Social Care Committee recommendations, Government acknowledge empirical evidence that DCRs are successful in “addressing problems of public nuisance and reducing health risks in a very specific set of circumstances (for example, where open drugs scenes present a significant risk to public health)”.⁶⁶ What they fail to acknowledge is that DCRs are highly successful at reducing preventable death. Whatever other harm may attend addiction, it can be better addressed through robust harm reduction measures; and it simply cannot and should not compete with avoidable death for policy priority.

Second, this argument has important implications for the design and aims of harm reduction programmes when they are adopted. If addiction itself is not a harm, then setting “full recovery”⁶⁷ or “freedom from all dependency”⁶⁸ – that is, cessation of monitored opioid agonist use – as the benchmark for success in OAT is inappropriate from a wellbeing perspective. Continued reliance on a safe and easily procurable opioid agonist does not represent a morally salient increase on background vulnerability; and for those patients who still feel they benefit from it, does not represent an autonomy harm. Winstock, Eastwood and Stevens are right to applaud the Home Office's abandonment of arbitrary OAT time limits in the 2017 Drug Strategy;⁶⁹ but the retention of “full recovery” as the benchmark for OAT success is both inappropriate and a potential threat to important efforts to reduce the separable harms of addiction.

Finally, as Winstock, Eastwood and Stevens note, the 2017 Strategy “never allows that people who use drugs can be rational, informed, and interested in their own health and wellbeing”.⁷⁰ We have shown that not only can and should the average (i.e. non-dependent) person who uses drugs be treated as an autonomous decision maker; but that there is no reason to think persons with recalcitrant addiction cannot as well. Addiction does not rob one of one's autonomy; and the vulnerability of persons with addiction is largely in our own hands.

Beyond these specific applications to UK drugs policy, our argument has broader implications for any jurisdiction reconsidering its policies with regard to the public health consequences of problematic drug use. It can be tempting, in considering the debate between harm reduction and prevalence reduction, to see harm reduction as a policy of last resort, to be adopted only when prevalence reduction proves to be too difficult or too costly. On this view, adopting harm reduction requires us to accept the ongoing autonomy-sapping vulnerability of addiction as the lamentable but necessary cost of preventing a public health crisis of death, disease, homelessness and crime.

But if our argument is correct, this is a mistake. The choice between harm reduction and prevalence reduction should not be cast as a difficult trade-off between public health and autonomy, one where the overriding interest in preventing death and disease forces us to accept a loss of autonomy and an increase in regrettable vulnerability. Instead, we should recognise the ways in which harm reduction can enhance autonomy without ending addiction. A concern for the autonomy of drug users is no reason to prefer prevalence reduction or abstinence-based approaches to drug policy. We do not need to choose between autonomy and public health, and we do not need to see harm reduction as a second-best strategy, worth adopting only when normatively better strategies have failed. To the extent that addiction impairs autonomy and increases vulnerability, this is often an avoidable result of misguided drugs policies, rather than an inevitable consequence of drug dependence. The way to protect autonomy and reduce vulnerability is therefore not to double-down on abstinence-based prohibitionist policies, but rather to expand the harm reduction approach through the full range of drug policies.