Background
Advances in biobehavioral science occurring over the past several decades have made significant headway in elucidating mechanisms that undergird addictive behavior. This large body of research suggests that addiction is best regarded as a cycle of compulsive substance use subserved by dysregulation in neural circuitry governing motivation and hedonic experience, habit behavior, and executive function [
1]. Though findings from the basic science of addiction have yielded novel treatment targets that may inform the development of promising pharmacotherapies, the behavioral treatment development process often lags behind the ever-accelerating pace of mechanistic discovery. In that regard, the mainstays of behavioral addictions treatment, cognitive-behavioral therapy and motivational interviewing, were developed decades ago and prior to the current understanding of addiction as informed by neuroscience. Yet, to the extent that behavioral therapies target dysregulated neurocognitive processes underlying addiction, they may hold promise as effective treatments for persons suffering from addictive disorders.
Contemporary developments in addiction neuroscience have been paralleled by increasing interest in the age-old mental training practice of mindfulness meditation as a potential treatment for addictive behavior. This interest was sparked by the successful integration of mindfulness techniques into secularlized behavioral intervention programs, including Mindfulness-Based Stress Reduction (MBSR) [
2] and Mindfulness-Based Cognitive Therapy (MBCT) [
3]. Standardized mindfulness training programs were originally focused on reducing emotional distress, and indeed, for psychiatric disorders and symptoms mindfulness-based interventions (MBIs) have been shown through meta-analysis to be efficacious and comparable to other active, head-to-head treatments [
4]. More recently, MBIs like Mindfulness-Based Relapse Prevention (MBRP) [
5] and Mindfulness-Oriented Recovery Enhancement (MORE) [
6] have been tailored to directly to address the mechanisms that undergird addiction. A growing body of controlled research studies demonstrates that MBIs may produce significant clinical benefits for users of a panoply of addictive substances, including alcohol, cocaine, nicotine, and opioids. The aims of this report were to operationalize the construct of mindfulness with respect to therapeutic processes that mediate its potential efficacy; review the current state of research on mindfulness as a treatment for addiction; and to envision the next wave of research in this emerging and important field. With regard to setting a future research agenda here we highlight issues related to: research rigor and reproducibility; treatment optimization based on mechanistic discoveries; the sequencing of MBIs in multimodal treatment packages; the need to consider dose–response relationships; the translation and dissemination of MBIs into standard, community-based addiction treatment settings; and the possibility of construing mindfulness as an integral component of a recovery-oriented lifestyle rather than a time-limited treatment.
Current state of the field: a review of clinical outcomes of mindfulness-based treatments for addiction
A considerable body of findings has amassed supporting the capacity of MBIs to reduce substance use and attenuate factors promoting substance use, such as craving and stress. Over the past decade, multiple systematic reviews have been conducted to identify the effects of MBIs on addictive behaviors, and have found accumulated evidence for the positive effects of MBIs [
61‐
63]. More recently, a meta-analysis focused on the broad clinical efficacy of MBIs for a range pf psychiatric disorders conducted subgroup analyses to examine the effects of MBIs on addiction/smoking and found MBIs to be superior to active control conditions and comparable to other evidence-based treatments [
4]. In the only published meta-analysis solely focused on MBIs for substance misuse, Li, Howard, Garland, McGovern, and Lazar [
64] identified 34 randomized controlled trials differing in terms of the types of MBI and comparison groups contrasted, sample demographics, and measures of outcomes and other constructs. Despite the notable methodological heterogeneity of these investigations, the authors concluded that “virtually all studies found that mindfulness treatments were associated with superior treatment outcomes at posttreatment and follow-up assessments compared to comparison conditions” (p. 69). Effects (Cohen’s d/odds ratios) ranged from moderate-to-large across the synthesized effect sizes computed for studies within the substance use (
d = 0.33, − 0.49 to 0.17, p < 0.05), cigarette smoking (OR = 1.76, 0.99–3.15, p = 0.056), craving (
d = 0.68, − 1.11 to − 0.025, p < 0.01), and stress (
d = 1.12, − 2.24 to –0.01, p < 0.05) domains.
With regard to secondary or mechanistic outcomes, as expected, MBIs produced significant increases on the Five Factor Mindfulness Questionnaires in all eight studies that used this measure (
d = 0.62, − 0.02 to 1.26, p = 0.057). In individual studies, MBIs produced a host of other significant salutary effects including increases in emotion regulation [
41,
54], substance-related self-efficacy [
65,
66], and positive emotions [
33], as well as decreases in attentional bias [
52,
66,
67], addictive automaticity [
66], dysphoric affect [
40,
66], and pain severity and related functional interference in patients with chronic pain [
41]. Several studies reported positive associations between the degree to which participants engaged in mindfulness homework exercises and changes in cigarette, marijuana, and alcohol use posttreatment (e.g., [
68‐
70]).
Of the 34 RCTs reviewed in this meta-analysis, ten used treatment-as-usual comparison groups, whereas two used inert comparison groups, sixteen employed an alternative psychotherapeutic treatment (typically matched to the MBI group vis-à-vis intensity, duration, and format), and six examined brief mindfulness treatments compared to alternative therapies in laboratory settings. Twenty-eight of the reports presented the first published findings from the related study and six reports presented results of secondary analyses. Any given study could contribute findings only once to meta-analyses conducted within outcome domains. The adequacy of randomization was examined in all studies and analysis of covariance and linear mixed modeling were often used to control for any remaining pretreatment differences. Nearly half of the studies had samples sizes less than fifty. Many studies had high attrition rates at posttreatment and subsequent follow-ups. Most of the 34 studies reviewed relied extensively on self-report measures of substance use and other constructs. All RCTs examined were single-site studies. The most common methodological limitations were failure to interview collateral informants regarding study participants’ substance use behaviors at posttreatment and follow-up and to employ posttreatment and follow-up interviewers who were blind to participants’ treatment assignments. Fewer than half of the RCTs employed objective verification of participants’ self-reported substance use, such as urinanalysis.
Subgroup analyses within outcome domains indicated that MORE treatment was associated with larger effects than other MBIs for substance use, craving, stress, and mindfulness measures [
64]. Studies comprised entirely of men also reported larger effects for MBIs compared to studies with samples comprised only of women or those with mixed gender samples across measures of craving, stress, and mindfulness.
Li et al. [
64] also reported findings from a random effects meta-regression analysis examining effects of MBI type, primary type of substance misused, study sample size, sample age and gender distributions, type of comparison group, treatment dosage in hours, and study methodological rigor on effect sizes by domain. Results indicated that studies with samples of only men experienced larger reductions in levels of craving and stress, and significantly larger increases in levels of mindfulness, compared to studies with samples comprised only of women or studies with samples comprised of women and men. Although the authors did not include a formal search for “gray literature” related to MBI treatment of substance misuse, they noted that funnel plots and Egger’s test analyses suggested that their findings were not likely due to publication bias.
Randomized controlled trials suggest that MBIs are a promising treatment for substance misuse and exert their effects via increases in levels of mindfulness across a wide array of substance-misusing behaviors and clinical populations. Future research should employ larger samples, longitudinal designs with follow-up periods of at least 1-year, manualized interventions with treatment fidelity assessment, intent-to-treat analyses, and probability sampling designs allowing generalizability to specific clinical and general populations.
Laying out a research agenda
Research rigor and reproducibility
MBIs are promising treatments for addiction. Results from rigorous, full-scale RCTs indicate that MBIs can produce short and long-term reductions in craving and addictive behavior. At this juncture in the development of the field, additional Stage III and IV clinical trials (for a review of the NIH Stage Model, see [
71]) are needed to replicate these promising findings via gold-standard research design features including the use of active control conditions, detailed fidelity monitoring procedures, and triangulation of self-reported outcomes with biochemical verification of drug use and blinded clinical evaluations. With additional replications of positive clinical outcomes, MBIs could rightfully be considered empirically-supported therapies for addictive behaviors. Conversely, replication failures could indicate the need to “return to the drawing board” and engage in treatment development research to optimize the next generation of MBIs as interventions for addiction. Thus, in the lifespan of this nascent field, it is now an opportune moment to answer definitively the question “Are MBIs efficacious and comparatively effective treatments for addiction?”
Assuming an affirmative answer to the aforementioned question, studies should then aim to address research questions pertaining to
mediation (“How do MBIs improve addiction-related outcomes?”) and
moderation (“For whom do MBIs work most optimally to improve addiction-related outcomes?”). As discussed in “
Mindfulness as a means of targeting mechanisms of addiction” section, a corpus of research has begun to amass on the mediators of MBI effects on addiction. In contrast, there is very little research on moderators of MBIs. The only study of MBI moderators for addiction outcomes is a secondary analysis of data from two RCTs of MBRP, which found that patients with greater substance use disorder severity and more affective symptoms received significantly greater benefit from mindfulness training than patients with low levels of substance use and affective symptoms [
72].
A number of additional research questions remain unanswered. Here we lay out an agenda for the next wave of research in the field.
Elucidating the neurobiological mechanisms of mindfulness as a treatment for addiction
Little is known about the neurobiological mechanisms of mindfulness as a treatment for addiction. Though various conceptual models have been advanced [
22,
23,
73], few tests of these specific neural hypotheses have been conducted. Adequately powered, randomized fMRI studies are needed to test basic mechanistic assumptions long held in the field. For instance, do MBIs decrease addictive behavior by strengthening inhibitory control via activation of top-down neural circuitry? Do MBIs decrease addictive behavior by reducing activation of bottom-up neural circuitry to drug cues? Similarly, functional neuroimaging methods are needed to test novel hypotheses, such as the
restructuring reward hypothesis (“Do MBIs restructure the relative responsiveness to drug and natural rewards by increasing functional connectivity between top-down and bottom-up neural circuits?”). Furthermore, molecular neuroimaging (e.g., positron emission tomography; PET) is needed to understand effects of MBIs on neurotransmitters and neuropeptides implicated in addictive behavior like dopamine, endogenous opioids, γ-aminobutyric acid (GABA), and endocannabinoids. Finally, dynamic effects of mindfulness practice on addictive responses are unknown, and could be elucidated through functional neuroimaging techniques with high temporal resolution like electroencephalography (EEG) or magnetoencephalography (MEG). Such methods could answer other pertinent questions. For instance, does the acute state of mindfulness attenuate initial attentional orienting to drug cues? Or, does mindfulness facilitate attentional disengagement and recovery from drug cue-exposure? These questions can be answered by investigating how mindfulness training influences the time course of neural responses to drug cues.
Although understanding treatment mechanism is not necessary to establish a given treatment modality as an empirically supported intervention, understanding the mechanisms of mindfulness can inform the refinement of MBIs to yield larger effect sizes and produce additional therapeutic benefits. A case in point is MORE, which was refined based on mechanistic discoveries. Following the first trial of MORE, it was found that mindfulness reduces pain severity by fostering a shift from affective to sensory processing of pain as innocuous sensory information [
74]. As a result of this discovery, when MORE was optimized as a treatment for prescription opioid misuse among chronic pain patients, the intervention was modified to include a “mindfulness of pain” technique that involved using mindfulness to deconstruct pain into its sensorial subcomponents and disentangle sensation from its affective overlay. Similarly, evidence that increasing physiological responsiveness to natural rewards via mindful savoring predicts decreased prescription opioid misuse [
75] and craving [
30] has led to an enriched emphasis on mindful savoring practice in the MORE intervention. It is possible that these intervention refinements may account for the changes in brain reward circuitry function observed among smokers treated with MORE [
34]. As another example, recent investigation of the role of the posterior cingulate cortex in meditation experience has implicated this brain region as a target for neurofeedback interventions to potentiate the efficacy of MBIs [
76], and indeed, trials of such neurofeedback-enhanced MBIs are underway (e.g., NCT02413177).
Sequencing of mindfulness as a part of multimodal treatment packages
It is not known whether MBIs are most efficacious as standalone treatments or as a part of a more comprehensive treatment package. In many inpatient addictions treatment programs, clients receive multiple behavioral interventions (e.g., motivational enhancement therapy, cognitive-behavioral therapy, dialectical behavior therapy) during the same 30-day time frame. Further, optimal treatment sequencing has not been studied. For instance, would MBIs be more efficacious following several sessions of motivational interviewing? Given that MBIs involve mindfulness practice, and regular practice requires motivation, introducing several sessions of motivational interviewing before initiating a course of mindfulness training might increase practice engagement and thereby boost clinical outcomes. Conversely, mindfulness training might potentiate motivational enhancement therapy by increasing interoceptive awareness of adverse consequences of addictive behavior on bodily health. In a similar vein, mindfulness training might increase adherence to medication-assisted therapy (MAT) by increasing awareness of how medication adherence allays the dysphoria associated with craving and thereby potentially improves quality of life. In turn, MAT might improve adherence to MBIs by attenuating distracting withdrawal symptoms and decreasing obsessive thinking about obtaining the next drug dose, thereby freeing cognitive and motivational resources to devote to learning mindfulness skills. Psychopharmacological interventions, cognitive training via computer- or smartphone-deployed technology, neurofeedback, and neurostimulation (via transcranial magnetic stimulation or transcranial direct current stimulation) administered prior to initiating a course of MBI might also improve cognitive function to facilitate learning of mindfulness techniques, and thereby improve MBI outcomes.
Sequential, multiple assignment, randomized trials (SMART) could be used to assess the efficacy of dynamic treatment regimens, including those that are individually tailored based on decision rules that dictate how the type or dosing of treatment should change based on the specific clinical needs of the patient [
77]. For instance, MBI non-responders might need a supplementary course of motivational enhancement therapy, computerized cognitive remediation, or booster sessions (see “The Need for Dose/Response Research” below) to enhance outcomes. Finally, given that many MBIs are multimodal in nature and combine various mindfulness meditation practices and psychoeducational modules, studies that employ the multiphase optimization strategy (MOST) could also be used to examine the independent and additive effects of various MBI treatment components on addictive behaviors [
78]. The MOST research process could allow for resource-intensive and complex MBIs to be pared down to their most efficacious elements to maximize efficacy and efficiency by eliminating techniques that do not confer therapeutic benefits and augmenting those that do.
The need for dose–response research
In pharmacological research, it is imperative to examine dose–response relationships to identify the optimal therapeutic dose. Dose–response curves can help to identify the dose needed to achieve a satisfactory clinical outcome while minimizing the side-effect profile of the drug. Although MBIs delivered in clinical settings appear to have few adverse effects [
79], the costs and time required to deliver complex behavioral treatments like MBIs necessitate dose–response considerations to identify the minimal therapeutic dose. Null effects of MBIs observed in Stage II or III clinical trials might very well be qualified by extent of mindfulness practice, and thus mindfulness practice engagement should be tested as a treatment outcome moderator. Furthermore, responder analyses might reveal that individuals classified as non-responders are those who do not meet the minimal therapeutic dose of mindfulness skill practice whereas individuals classified as responders are those who surpass this minimal therapeutic dose of practice.
Given meta-analytic findings that extent of mindfulness practice is significantly associated with treatment outcomes [
80], different doses of mindfulness practice might produce different therapeutic effect sizes or different durations of therapeutic effects for addicted populations. Most MBIs for addictive disorders (e.g., MBRP and MORE) are approximately 2 months in length given that they were modeled on the canonical 8-week structure of MBSR [
81]. However, due to their clinical complexity, individuals with substance use disorders are typically excluded from participating in MBSR. Although MBIs like MORE and MBRP have produced significant reductions in addictive behaviors [
64], it is plausible that to achieve full remission from moderate-to-severe substance use disorder, patients might require additional weekly treatment sessions beyond the standard 8-weeks of treatment. Moreover, following a full course of a multi-week MBI, periodic booster sessions might be needed to extend treatment benefits for the long-term. Such booster sessions could come in the form of mindfulness practice sessions (with or without group process and psychoeducational content) conducted via in-person or telemedicine formats, and their additive efficacy could be tested with SMART research designs.
The challenge of dissemination/implementation
One of the greatest challenges confronting the movement towards evidence-based practice in addictions treatment is the research-to-practice gap: that is, empirically-supported therapies with proven efficacy as revealed by Stage II randomized clinical trials are often not successfully translated into effective clinical interventions in standard addiction practice settings [
82]. Successful transfer of research to practice involves programmatic change in the form of activities including exposure, adoption, implementation, and practice of new empirically-supported approaches [
83]. These activities are especially complicated in the context of MBI implementation, insofar as many common MBIs require intensive instructor training. For example, the MBSR certification process costs more than $10,000 and requires approximately 3 years to complete depending on how long it takes a prospective instructor to meet the requirements, which include personal practice and participation in multi-day meditation retreats, didactic and experiential workshop training, experience leading multiple MBSR groups, and clinical supervision [
84]. Further complicating this issue, individuals without clinical licensure can be certified in MBSR, yet most addictions treatment settings require staff to be licensed healthcare professionals. In contrast, other MBIs like MORE require clinical licensure but entail a much briefer and less costly training process. It remains an open question for future research as to how much clinical training, supervision, and personal practice experience is required for effective implementation of MBIs in clinical settings. Moreover, it is not known which training formats are most effective (in person, online, role play, virtual reality, etc.) in disseminating MBIs. Issues around treatment fidelity are also crucial to successful implementation of MBIs in clinical practice. However, few fidelity measures have been validated for MBIs for addiction (for a notable exception, see
85), treatment fidelity research is time intensive, and little is known about empirical relations between clinician training format, therapist adherence/competence, and MBI treatment outcomes. Similarly, the acceptability of MBIs may influence their implementation in clinical practice settings. Factors influencing the acceptability of MBIs for the treatment of addiction are poorly understood. For instance, it is plausible that patients who initially experience mindfulness meditation as rewarding (i.e., alleviating psychological distress and generating positive sensations and emotions) or who are positively reinforced by the therapist for engaging in meditation practice may be most likely to continue to practice mindfulness skills. In contrast, patients who experience an exacerbation of aversive thoughts and feelings during meditation or who receive neutral responses from the therapist might be most likely to drop out from an MBI. Moreover, non-specific factors like therapeutic alliance, and allegiance might drive MBI acceptability, adherence, and outcome in a similar fashion to other behavioral therapies. Strategic attention to such factors might in fact boost the uptake and clinical efficacy of MBIs.
In outlining issues pertaining to advancing the clinical science of MBIs, Dimidjian and Segal highlight the tension between the need to make MBIs disseminable in the context of real-world resource constraints and complex client populations while not allowing outcomes to suffer as MBIs are scaled up in the translation to community treatment settings [
86]. This is indeed a challenge, as MBIs with demonstrated efficacy in Stage II trials may fail to show effectiveness in Stage III and IV trials when delivered by community clinicians. Yet, work now needs to be done to understand the feasibility, acceptability, and impact of delivering MBIs in addiction treatment settings.
Mindfulness as a relapse prevention strategy versus mindfulness as a vehicle for recovery
Finally, it is unknown whether mindfulness might best ameliorate addiction through participation in time-limited interventions or if mindfulness should be used daily as part of a wellness lifestyle. With regard to the latter, shifting from an addiction-oriented lifestyle to adoption of a wellness lifestyle is conceptualized as integral to the recovery model [
87]. In this vein, studies should examine mindfulness not only as a technique in circumscribed interventions to prevent addiction relapse but also examine mindfulness as a long-term, sustainable health behavior that promotes addiction recovery. Pursuit of a healthy lifestyle is not something that is finalized over the course of an 8-week intervention; to the contrary, maintenance of physical health requires ongoing, regular exercise and nutritious dietary choices on a daily basis that do not exceed the caloric needs of the individual. Why then should mindfulness practice be any different? As a point of consideration, 12-Step programs encourage participation in regular meetings for the entirety of one’s life. Similarly, mindfulness might need to be practiced daily or nearly every day on an ongoing basis to achieve durable therapeutic effects and maintain addiction recovery, especially in view of the chronicity of addictive disorders.
From a neurobiological perspective, increasing grey matter density, strengthening of white matter tracts, synaptic remodeling, and other neuroplastic modifications to brain structure and function needed to undo the pathophysiology of addiction might require recurrent mindfulness practice for the long-term. From a psychological perspective, long-term mindfulness practice may be needed to induce self-referential plasticity and facilitate flexible reconfiguration of the self-schema in relation to the world [
88] so as to restructure reward processes away from valuation of drug reward and towards valuation of personally meaningful pursuits and relationships [
23,
29]. This latter process is consistent with the ancient soteriological intention of mindfulness as a means of reducing craving by gaining insight into the true nature of the self as impermanent and interdependent [
89]—paralleling Bateson’s classical cybernetic model of addiction recovery [
90].
Conclusion
The study of mindfulness as a treatment for stress and chronic pain is more than 30 years old, and researchers have investigated mindfulness as a treatment for depression for more than two decades, yet it is only in the past 10 years that research on MBIs for addiction has proliferated. This is a young scientific field, and more research is needed to elucidate the clinical outcomes and mechanisms of this promising new treatment approach for addictive disorders. One recent meta-analysis [
64] indicates that MBIs produce statistically significant effects on craving (pooled Cohen’s
d = 0.68) and substance misuse (pooled Cohen’s
d = 0.33), suggesting that MBIs may be efficacious treatments for addiction. Overall, a number of RCTs with active control conditions have been conducted in the past decade—a sign that the methodological rigor of this field is increasing. However, with several notable exceptions (e.g., [
40,
91,
92]), few studies of MBIs for addiction have had large enough sample sizes to ensure the robustness and reproducibility of clinical outcomes. Moreover, few long-term follow-ups have been conducted to assess the durability of observed treatment effects. In addition, as indicated earlier, little is known about mediators and moderators of MBIs for addiction, although understanding how and for whom MBIs work is crucial to the overall evolution of this therapeutic approach. Lastly, research is needed to situate MBIs into treatment sequences with high external validity that adaptively address the needs of responders and non-responders in a way that can be realistically implemented in community-based treatment settings. Thus, the nascent field of mindfulness treatment for addictive behaviors remains open to rigorous, scientific exploration and in need of innovative research questions and methodologies.
Coming full circle, MBIs are some of the newest additions to the armamentarium of addictions treatment. It is perhaps no coincidence that the rise of MBIs has been co-incident with advances in the neuroscience of substance use disorders. In recognizing that addiction is, in large part, mediated by cognitive and behavioral automaticity propelled by alterations to hedonic regulatory systems in the brain, this perennial form of human suffering may be especially tractable to treatment approaches like mindfulness that enhance top-down conscious control over bottom-up automatic habits and motivational drives. Insofar as the original purpose of many mindfulness meditation practices was to extinguish craving by revealing the “middle way” between attachment to pleasure and aversion to pain, MBIs may ultimately provide a skillful means of liberating the individual from the push and pull of hedonic dysregulation underlying addiction.