Model melee: understanding models of addiction

KEYWORDS:

• Behavior
• addictive
• substance-related disorders
• brain diseases
• bioethics
• psychiatry

Introduction

Over 138 million Americans report illicit drug use in their lifetimes [1]. The number may be surprising, but the rationale for use is not: drugs can provide pleasure, pain relief, cognitive enhancement, and new experiences, among other functions [2]. Where the rationale becomes unclear is addiction, in which people continue to seek and use drugs (licit or illicit) seemingly against their own interests, as costs appear to outweigh benefits [3]. To develop strategies to help people suffering from addiction – of whom there are more than 20 million in the US alone [1]—theories of addiction that explain this puzzle are essential. The brain disease model of addiction (BDMA) and the choice model of addiction do so in different ways, and each implies different avenues of research and treatment. Although the BDMA is dominant, the choice model better captures the evidence.

Three models

The most widespread model today is the BDMA, which characterizes addiction as a chronic and relapsing neurobiological disease characterized by compulsive drug seeking and use despite negative consequences [4]. This is a straightforward explanation of why an addicted individual would continue to use drugs: in addiction, drug use is no longer a voluntary behavior. Brain structure and function changes in the mesolimbic reward system directly cause compulsive drug seeking and use [4]. There is no single authoritative definition of compulsion in addiction research, but in the case of the BDMA, it generally refers to the drug seeking and use not being subject to voluntary control [3,5]. According to the BDMA, drug use in addicted people is not an irrational choice because it is not a choice at all.

The choice model, however, rejects that voluntary control is bypassed in addiction [6]. Even in addiction, it maintains that drug use is a choice. It instead solves the puzzle of addiction through appeal to value: addicted people use drugs because the drugs have higher reinforcement value than available non-drug alternatives, even if the cost-benefit ratio seems to greatly favor these alternatives to outside observers. Possible mechanisms of this discrepancy will be discussed later.

Both models share a denial of the older moral model, which holds that drug use is a choice and condemns this choice as morally wrong. It blames people with addiction as unwilling to stop their quest for pleasure that is costly to themselves and others [6,7]. The BDMA avoids moral condemnation by rejecting that drug use in addiction is a choice. If drug use is not voluntary choice, addicted people are intuitively not responsible for it and thus cannot be morally blamed. While the choice and moral models agree that addicted drug use remains voluntary, the choice model does not morally condemn it [6]. If drug use is not wrong, addicted people cannot be morally blamed for it.

Animal model evidence

Of course, moral implications of the models do not define their accuracy – evidence does. Animal models present one source of evidence from experiments that may be impractical or unethical to perform in humans. Early animal research seemed to provide evidence for the BDMA (at least in rats). These studies showed that rats will escalate cocaine consumption as well as effort and risk for cocaine seeking and use following extended self-administration [8]. In fact, rats would self-administer cocaine and other drugs over even food and water, sometimes until death. Such extreme findings might be interpreted to indicate that addiction in rats is a disease of compulsion that destroys control over drug behavior. However, these experimental designs were flawed. The rats involved did not have valuable choices available other than drug self-administration [8]. Addiction in humans, by contrast, theoretically involves drug seeking and use in place of a range of other, seemingly more beneficial, behaviors.

Some recent animal models have produced much different results. For example, a 2018 forced-choice study, in which rats had a mutually exclusive choice between methamphetamine/heroin administration or social reward, provides evidence for the choice model in rats. All rats chose social reward over drug self-administration, including rats meeting criteria for addiction in an established rat model based on the Diagnostic and Statistical Manual of Mental Disorders [9] (A simplified summary of this model’s criteria is whether the rats will persevere in drug seeking even when there is no longer reward for doing so, whether they are motivated to self-administer drugs even when it requires increasingly more time and energy, and whether they will maintain drug seeking despite punishment by foot-shock [10]). A preference for the drug reward could only be induced through delay or punishment of the social reward. That no rats chose drugs over the alternative reward means all of them retained choice and could not possibly have had a brain disease of compulsion. As even rats who are addicted (according to a rat model) maintain choice, this supports the choice model of addiction. Addicted rats chose to abstain from drugs because the social reward was more valuable to them than the drug reward. Other studies in rats and nonhuman primates have similarly found that many animals will prefer food or sweetened water to cocaine or heroin [8].

Of course, there are many reasons to be skeptical about the generalizability of animal models to people. First, the determinants of drug choice in laboratory animal studies are multifaceted, including a variety of factors such as the drug dose, the schedule and temporal parameters of reinforcement, drug mixtures, punishment, and other behavioral-economic considerations [11]. Thus, it can be difficult to make conclusions about the effects of the drugs themselves. Moreover, human drug addiction does not naturally occur in forced-choice scenarios. For example, drug use for some people may allow or even encourage social interaction rather than prevent it. Also, humans and other animals may value rewards differently. Venniro points out that human addiction persists despite available social interactions [9]. He thinks that this difference from the experimental rats comes not from the mutual exclusivity of the choice, but from the fact that humans evaluate social interaction based on meaningful participation in society, a reward less concrete and immediate than mere interaction with a companion. People who do not expect to have this ‘stake in conventional life’ may not value social reward as much compared to drugs [9]. Furthermore, human mental functioning is different from other animals in ways that may be relevant to addiction. For example, only humans have autobiographical conceptions of their experiences, life narratives, and future goals [12]. However, if anything, this should give humans more resources than other animals to control their own behavior, so if other animals maintain choice in addiction, it is logical that humans would, too.

Evidence on human choice in addiction

Thankfully, we do not need to rely solely on animal models. There is also evidence that choice is maintained in addicted humans. One important piece of evidence is that sometimes addicted people do choose abstinence. Most people who become addicted to illegal drugs are no longer addicted by age thirty [13,p.31]. Though addiction is (by definition) chronic and relapsing, this statistic would not be possible if addicted people (or at least, if all addicted people) had a brain disease of compulsion, as they voluntarily choose to quit.

More evidence that choice is maintained in addicted humans comes from the success of contingency management treatments, in which incentives are offered for desired behaviors like abstinence. For example, therapeutic workplace interventions, in which addicted people are hired for a job, have been successful [14]. The catch is that the participants must provide drug-free urine samples or take addiction medications to access the workplace or get their full pay. This acts like the mutually exclusive choice in the rat studies, negatively reinforcing drug use and positively reinforcing abstinence through earned payment and giving addicted people part of the stake in conventional life conceived by Venniro. Though contingency management treatments are not successful in all people, and addiction may remain chronic and relapsing even after the treatment, if the BDMA were true – and dysfunctional brain changes overrode voluntary choice to make addicted people use drugs – these interventions could not be as successful as they are. Instead, as the choice model predicts, voluntary control over drug use is retained even in addicted people, and they can and often will choose other behaviors that are more valuable to them than drug use.

Yet, there are first-person reports from addicted people who say that their urges to use drugs feel irresistible [3]. This denial of choice is consistent with the BDMA. Such feelings may represent a conflation of the great difficulty of abstaining with true compulsion, but may also represent genuine cases of a brain disease of compulsion.

Evidence for the mechanisms of addiction predicted by the models

The BDMA and choice model each offer mechanisms that could underly the puzzling continued use of drugs in addiction. The BDMA predicts that pathological brain changes circumvent choice. There is evidence that brain changes do occur in addiction, such as ‘mesolimbic hyper-reactivity to drug cues and drug-imagery’ [15]. These changes may sometimes be pathological and circumvent choice, indicating cases of addiction consistent with the BDMA. However, just because changes are observed in addiction does not make them pathological (for example, all learning changes the structure and function of the brain, but these changes are not pathological – to claim that a change is pathological requires an account of normal brain function [3]) nor imply that the symptoms central to addiction are caused by them. It is also plausible that these changes could count as pathological but make choosing to abstain merely difficult rather than impossible. In this case, there would be a brain disease, but not a brain disease of compulsion, in the sense of the loss of free will or the bypassing of voluntary control. Thus, challenging the BDMA does not necessarily require denial of changes in brain structure and function, but denial that these changes override volitional behavior.

The choice model denies compulsion and explains the choice of addicted people to continue using drugs by the relative value of that choice appearing greater than that of non-drug alternatives for the addicted person compared to the outside observer. One mechanism of this discordance is anomalous decision-making causing increased expected value or decreased expected costs of drug use [16]. There is evidence for such anomalies: compared to non-addicted people, addicted individuals perform worse on tests for some decision-making tasks, indicating atypicality in various aspects of decision-making, such as tolerance of risk [17]. These differences in decision-making processes are not necessarily dysfunctional but conform with the idea that different judgment of costs and benefits contributes to addicted people making drug use decisions that seem puzzling to outside observers, as predicted by the choice model.

Another reason that continued drug use would seem valuable to addicted people while seeming detrimental to outside observers is that drug use may have value to addicted people that is hidden, or at least not obvious, to non-addicted observers [16]. One example of many is that identifying as an ‘addict’ may itself have value. For instance, it can produce social reward and community belonging among others who share the identity. This effect is visible in Bourgois and Schonberg’s fieldwork in a community of homeless people addicted to heroin in their book Righteous Dopefiend. In this community, individuals care for and support each other, and none could survive without this cooperation [18]. By becoming abstinent and removing their addict identities, individuals would lose this social interaction and community support for which they may have no alternative. Indeed, Hanna Pickard points out that, ‘for people who lack a genuine alternative sense of self and social identity, recovery represents an existential threat’ [16]. Thus, for them, continued drug use provides an identity whose value is real but unobvious to those without it, justifying the choice (likely along with numerous other factors) in a manner consistent with the choice model.

Room for both models?

As alluded to earlier, the BDMA and choice model of addiction are not necessarily mutually exclusive. The available evidence all but rules out the idea that all (or even most) cases of addiction involve a brain disease of compulsion; however, it is also conceivable that there is heterogeneity in addiction and that the BDMA does explain some cases. These cases could include those who fail to recover spontaneously or respond to contingency management. It is also important to emphasize that the choice model does not deny all involvement of neurobiological processes in addictive behaviors and learning, but rather acknowledges that individuals’ choices can be driven by the fact that drugs can have immense value for them without the need to appeal to compulsion.

Potential harms of the models

One potential concern about the choice model is that, even if it rejects the moral condemnation and blame of the moral model, it may still cause stigma, reducing access to healthcare and treatment. Indeed, the BDMA has been advocated for as a means to combat the stigma of addiction [4].

First, it is important to establish that the downstream social effects of the models are unrelated to their empirical truth. It is possible that the BDMA is entirely incorrect and yet that labeling addiction as a brain disease has the positive effect of eliminating harmful stigma. However, whether this effect is observed in practice is a separate empirical question.

The evidence suggests that the answer is that this is mostly not observed. For example, the BDMA may be effective in decreasing attributions of individual blame for friends, family, and colleagues of people addicted to drugs but has not done so for the general public [2,19]. In fact, the BDMA increases social rejection and ostracization of addicted people by the general public as well as perceptions of their dangerousness and unpredictability [19,20]. Furthermore, biogenetic models such as the BDMA are associated with less empathy from clinicians than psychosocial models and have negative effects on addicted people themselves, such as pessimism and helplessness about recovery and a diminished sense of agency [19]. Therefore, the evidence that the BDMA challenges stigma – and that the choice model exacerbates it – is varied at best.

Conclusion

Addiction is a puzzle, but not a game: solving it is of serious importance to care for the countless people suffering from addiction. The BDMA may explain some cases, but the evidence points toward the choice model solving most of the puzzle. Animal models and human data both suggest that most, if not all, people maintain choice in drug use despite addiction. Brain changes predicted by the BDMA do exist, but they have yet to be proven pathological or explanatory of addiction, necessary conditions for the BDMA’s validity. While this is also not definitive proof, there is evidence supporting the existence of anomalous addicted decision-making and hidden value of drug use, as predicted by the choice model.

The debate between these models is not merely a theoretical exercise. It has significant implications for hospital practice and addiction treatment. For example, the degree of faith placed in each can affect the limited research funding available for addiction treatment. Specifically, a sole focus on the BDMA manifests as focusing addiction treatment research and funding on pharmacotherapies – at the expense of treatment strategies like contingency management and others that help shift the balance of choice by giving those suffering from addiction part of the ‘stake in conventional life’ described by Venniro [2,9].

The evidence for the choice model suggests a need for a paradigm shift toward prioritizing comprehensive care strategies that include psychological support and environmental modifications designed to empower patients to choose valuable non-drug alternatives, and indeed to create these meaningful alternatives. This must occur not only in research for the development of treatment strategies, but also in hospital practice, where healthcare teams must be able to assess individual patients’ value assessments and how best to guide them toward recovery. By acknowledging the role of choice, healthcare providers might adopt more collaborative approaches to care, engaging patients in meaningful discussions about their treatment options, values, and goals. This could foster a sense of agency and responsibility in patients that may be hindered by a focus on the BDMA. If serious progress in addiction treatment is to be made, the evidence on the role of choice must be taken seriously to further elucidate the determinants of choice in addiction and how these insights can be translated into practice, ultimately leading to more nuanced and effective approaches to treatment.

Declaration of financial/other relationships

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. Peer reviewers on this manuscript have received an honorarium from IPGM for their review work but have no other relevant financial relationships to disclose.

Additional information

Funding

This paper was not funded.