RESEARCH DOSSIER
Contingency Management and Community Reinforcement in Addiction
Abstract. Contingency management (CM) and the community reinforcement approach (CRA) are behaviorally grounded interventions that modify the consequences and opportunity structures governing substance use and recovery. This monograph critically reconstructs their historical development, mechanisms, operational protocols, clinical trials, systematic reviews, durability, ethical constraints, and implementation across substance-use treatment systems. Its method is a structured critical narrative synthesis of identifiable primary experiments, randomized trials, comparative effectiveness studies, quantitative research syntheses, clinical guidelines, and government implementation documents available through November 2026; it is neither a preregistered systematic review nor an original meta-analysis. The evidence supports CM as an effective means of improving objectively verified substance-use and engagement outcomes during treatment, with especially substantial support for stimulant-use disorder. CRA reorganizes accessible social, occupational, recreational, and relationship-based reinforcement so that sustained non-use becomes more rewarding and practicable. Combining them addresses complementary timescales: immediate verified consequences and enduring ecological alternatives. Follow-up effects vary by protocol, population, measurement, and outcome; mechanistic mediation, long-term generalization, equity, and scalable fidelity remain important questions. A defensible treatment ecology joins participant choice, dependable measurement, attainable rewards, accessible recovery resources, and supportive continuity of care.
Keywords: contingency management; community reinforcement approach; addiction psychology; stimulant use disorder; cocaine use disorder; substance use disorder; operant conditioning; voucher reinforcement; prize-based incentives; community reinforcement and family training; behavioral economics; drug testing; treatment retention; abstinence; implementation science; health equity; reinforcement schedules; long-term recovery.
Introduction: The Research Object
Addiction treatment provides an unusually informative laboratory for examining how consequences alter behavior under conditions of competing immediate and delayed rewards. Drug consumption can produce rapid pharmacological consequences, can be cued by familiar environments, and may persist even when a person clearly recognizes serious longer-term harm. Treatment cannot assume that insight alone reverses the distribution of available reinforcement. The psychological research object is the way deliberately arranged contingencies and naturally occurring community rewards change the probability, stability, and social conditions of substance-related behavior.
Contingency management is a clinical arrangement in which a predefined, independently verifiable target behavior produces a timely, valued consequence. Community reinforcement is a comprehensive treatment strategy that increases access to rewarding activities, relationships, work, and practical capacities compatible with recovery. They can be studied separately and together. The distinction matters because an isolated voucher-for-negative-test program does not by itself deliver vocational or relationship support, while a community reinforcement intervention can function without formal monetary vouchers. The combined clinical strategy links short-horizon learning with a more sustainable reconfiguration of everyday incentives.
This dossier examines treatment mechanisms, causal claims established by trials, alternative explanations, reward design, implementation, and research boundaries. It follows the historical literature from alcohol-focused community reinforcement to cocaine voucher trials, multisite prize studies, contemporary meta-analyses, and public-sector program implementation. A central source is the ASAM/AAAP stimulant use disorder guideline (2024) American Society of Addiction Medicine & American Academy of Addiction Psychiatry (2024), which recommends CM as a primary component for stimulant-use disorder treatment and identifies CRA among the preferred accompanying psychosocial therapies. This clinical recommendation applies to a specified condition and has stated evidence grades; it cannot be generalized unchanged to every behavioral or substance-use disorder.
Definitions and Fundamental Distinctions
A contingency specifies the relation between an observable action, the conditions in which that action occurs, and the consequences that follow. Positive reinforcement denotes an increase in subsequent behavior attributable to the presentation of a consequence. In clinical CM the operational target is usually a negative biological test, attendance, or medication adherence, rather than a vague judgment about motivation or character. The nominal reward is only a reinforcer when it actually influences the relevant behavior in a particular context. Its magnitude, delay, predictability, participant value, and competing costs are therefore substantive components of treatment rather than incidental administrative details.
CRA is a broad behavioral treatment whose functional hypothesis is that substance-using and substance-free patterns compete for reinforcement within a social environment. Treatment may include functional analysis of high-risk circumstances, communication and problem-solving skills, access to rewarding sober activities, employment assistance, relationship planning, and relapse-oriented coping. The classic formulation of this ecological principle appears in Hunt and Azrin (1973) Hunt (1973) and its subsequent development in Azrin (1976) Azrin (1976). A systematic account is provided by Smith, Meyers, and Miller (2001) Smith (2001).
Community reinforcement and family training (CRAFT) is a related but independently specified intervention for concerned significant others, particularly to encourage treatment engagement by someone reluctant to seek services. Its principal participant is initially the family member or partner, and its proximal endpoint may be treatment entry. CRAFT cannot simply be counted as another trial of CRA in directly treated patients. A therapeutic alliance, motivational interviewing, cognitive behavioral therapy, pharmacotherapy, and recovery capital are adjacent concepts whose explanatory levels differ. They may interact with CM and CRA, but each requires separate operationalization and evidence assessment.
Historical Origins: From Alcohol Rehabilitation to Behavioral Ecology
Hunt and Azrin's 1973 community-reinforcement study Hunt (1973) approached problematic drinking through rearrangement of everyday incentives linked to work, family, and social life. The intervention thereby treated community contingencies as elements of the therapeutic mechanism. The experiment belongs to an era when behavior modification was being adapted beyond laboratory learning tasks to complex clinical patterns. Its historical importance is the explicit proposal that the social value of a non-drinking lifestyle could be deliberately increased.
In 1976 Azrin Azrin (1976) reported modifications to the original program, incorporating added social and adherence procedures, including features tied to disulfiram. The historical program should be interpreted in its own therapeutic context. Some older components depended on treatment assumptions and clinical practices that require modern safety and regulatory reassessment. The general behavioral hypothesis is distinct from the endorsement of every historical technique.
As treatment programs professionalized, CRA became a recognizable multicomponent, manual-guided intervention rather than a loose exhortation to build a better life. Meyers and Smith's 1995 clinical guide Meyers (1995) codified a structured practical model. Later work extended community-reinforcement concepts to cocaine and opioid use, homeless populations, and concerned relatives. The historical sequence illuminates a recurrent methodological problem: efficacy estimates from comprehensive packages do not automatically identify which component was necessary or sufficient to produce an observed improvement.
The Cocaine Voucher Experiments
The influential Higgins et al. (1993) randomized study Higgins (1993) assigned 38 outpatient cocaine-dependent patients to a multicomponent behavioral treatment or drug-abuse counseling. The behavioral condition combined CRA-based counseling with cocaine-negative urine-contingent incentives, and thus compared treatment packages rather than isolating every component. In the published abstract, treatment completion and sustained objective abstinence favored the behavioral package. Small sample size, the bundled design, and differences between comparison conditions limit component-level attribution.
The clearer additive experiment is Higgins et al. (1994) Higgins (1994), in which 40 adults were randomized to a community-reinforcement-based treatment with or without voucher incentives contingent on cocaine-negative urine during the initial 12 weeks. Holding the broader behavioral platform more nearly constant made the incremental effect of incentives more interpretable. The conceptual contribution is substantial: an objective event could trigger a reliable alternative consequence within an existing counseling program, thereby changing the immediate choice environment without requiring a new theory of the patient's personality.
These early experiments shaped subsequent CM designs around escalating reward value for consecutive target achievements, rapid verification, and transparent resetting when a qualifying sample was absent or positive. Their lessons have been tested in larger treatment settings, but they do not define a universally optimal monetary level or schedule. The original design and population determine the permissible causal inference, and laboratory or specialist-clinic control cannot be assumed to persist unchanged after program-wide dissemination.
Operant Mechanisms and Competing Reinforcement
The proximate mechanism of CM is response–consequence learning. A reliably delivered incentive creates a shorter and more salient pathway from an abstinence-related act to a favorable consequence. Drug rewards are often experienced closer to the moment of use than the health, employment, financial, or relational gains associated with abstinence. CM compensates for this temporal asymmetry by bringing a compatible reward forward. In a clinical trial, randomization can establish that a CM package changed outcomes; it does not by itself establish that every participant experienced the same learning process.
Reward magnitude matters insofar as it changes the net value of the target action against competing alternatives. Reinforcement immediacy matters because long intervals introduce other behavior and contextual events between action and outcome. An escalating schedule may reward continuity more strongly than isolated achievements, while a restart rule can make lapses costly in foregone opportunity. Such design concepts should be viewed as hypotheses about behavior under different contingencies and validated for the target population rather than treated as universal numerical formulas.
Community reinforcement expands the field of available consequences: social participation can produce belonging; employment can provide financial stability and identity; structured leisure can offer enjoyment; skill acquisition can improve confidence and agency. These activities become therapeutically relevant when they are accessible and rewarding to the participant, with outcomes measurable beyond mere enrollment. The research problem concerns substitution and complementarity: CM can help establish a period of non-use in which alternative activities become viable, while CRA can build the continuing sources of reinforcement that persist after vouchers stop.
Behavioral Economics and Delayed Outcomes
Behavioral economic analysis asks how choices vary with reinforcer price, effort, delay, availability, and substitution. Substance use can persist when the alternative activities require money, travel, scheduling, safety, or social access that are unavailable. A nominally attractive reward may lack practical value if redeeming it is difficult; by contrast, transportation or grocery support may enable both immediate participation and subsequent recovery-related activities. This reinforces the case for assessing subjective value rather than equating face value with psychological impact.
Temporal discounting provides a useful explanatory framework for why certain immediate incentives can affect treatment engagement even when the person endorses the long-term benefits of change. The framework remains a behavioral model; evidence of elevated discounting in a group does not mean every individual has a fixed deficient capacity for future-oriented choice. Environmental scarcity, stress, institutional schedules, and the volatility of available resources can make short-term consequences rationally salient.
The policy implication is that incentive magnitude alone is an incomplete measure of intervention quality. The relevant unit is a functioning contingency with a verified behavior, prompt consequence, meaningful reward, reachable treatment encounters, and credible ongoing opportunities. A CM program that requires expensive transportation while offering small incentives might inadvertently penalize people with the greatest access barriers. That problem is analytically separable from the effectiveness of the core reinforcement principle.
Community Reinforcement as a Treatment System
CRA begins with a functional account of when substance use occurs and what it accomplishes for a particular person. The same observable pattern may be maintained through positive social reward, relief from withdrawal or distress, avoidance of conflict, or habitual responses to familiar settings. Treatment can then arrange alternative social and behavioral pathways and strengthen competencies needed to use them. This individualized formulation coexists with manualization: the intervention can have reproducible procedures while adapting goals to a participant's actual social environment.
Smith, Meyers, and Delaney (1998) Smith (1998) randomized homeless alcohol-dependent individuals to CRA or standard treatment at a day shelter. The trial illustrates the importance of examining interventions within settings where housing and other fundamental resources constrain access to reinforcement. The therapeutic target extends beyond alcohol consumption to the conditions under which improved functioning becomes possible. Findings from this population require careful generalization to other housing and treatment contexts.
Roozen et al. (2004) Roozen (2004) systematically reviewed CRA studies addressing alcohol, cocaine, and opioid addiction. Their work is valuable for organizing the historical evidence across substances, although the underlying programs vary in therapeutic package, intensity, setting, and co-intervention. The review supplies neither an identical universal dose of CRA nor proof that each module has a constant causal effect. Program descriptions, treatment fidelity, participant preferences, and meaningful functional outcomes are essential for replication.
Community Reinforcement and Family Training
CRAFT translates principles of contingent responding and alternative reinforcement into a structured intervention for concerned significant others. Its objective often includes increasing the likelihood that a person with problematic substance use enters treatment while improving the relative's ability to respond safely and effectively. It requires clear distinctions between encouraging voluntary treatment, setting boundaries, protecting oneself from violence, and taking responsibility for another adult's decisions.
In the Miller, Meyers, and Tonigan (1999) trial Miller (1999), 130 concerned significant others were randomized to CRAFT, Al-Anon facilitation, or a Johnson Institute intervention. The abstract reports treatment engagement of initially reluctant problem drinkers of 64%, 13%, and 30%, respectively, across these assigned approaches. Engagement was the decisive outcome. The result should not be relabeled as a direct patient-level abstinence effect or a demonstration that every family relationship benefits.
The Meyers et al. (1998) drug-use study Meyers (1998) likewise involved concerned significant others and an initial phase aiming to engage relatives in treatment. It illustrates the importance of defining the unit of analysis: family-member outcomes, the identified person's treatment entry, and subsequent substance-use outcomes are different endpoints. CRAFT should remain visible in this dossier because it operationalizes the community side of reinforcement, but it is not a substitute for directly treating an acute substance-use disorder.
Target Behaviors, Reinforcers, and Verification
A sound CM protocol defines its target in advance, including what counts as success, how often a participant can demonstrate it, who verifies it, and what consequence follows. Rapid biological assays may confirm recent non-use of a target substance, while attendance records or observed medication administration verify different behaviors. An abstinence incentive cannot logically be evaluated with attendance alone, nor can a negative test be treated as proof of lifelong recovery. The 2024 U.S. HHS/ASPE report U (2024) distinguishes objectively verifiable clinical events and describes both voucher and prize-based formats.
Urine or breath testing introduces its own measurement system, with detection windows that depend on substance, sampling frequency, assay characteristics, and physiology. A positive or missed test should be handled according to a transparent therapeutic protocol and medical circumstances, including legitimate prescribed medication, access barriers, and assay error. Testing procedures require privacy protections, consent, appropriate confirmation policy, and proportionate responses. Scientifically, ambiguous missing data should never silently become evidence of the same event as verified substance use.
An incentive must be offered on predictable terms. A voucher ordinarily has a specified redeemable value; a prize-based arrangement offers chances of obtaining items of differing values under published probabilities. A variable prize draw can be less expensive than guaranteed vouchers, but expected reward value and administrative burden must be evaluated. A credible quality audit verifies behavioral contingency, delivery timing, participant receipt, inventory, handling of missing tests, and fidelity across staff members and sites.
Voucher and Prize-Based Randomized Trials
As contingency management moved beyond specialist university clinics, investigators tested whether the basic contingency could be implemented inside ordinary outpatient programs. Petry et al. (2005) Petry (2005) randomized 415 cocaine or methamphetamine users in eight community outpatient psychosocial treatment programs to standard care or standard care with 12 weeks of abstinence-contingent prize opportunities. The multicenter design improved ecological relevance over earlier single-site experiments. The primary inference concerns the addition of the prize protocol to established care at those facilities, not the efficacy of every imagined prize schedule or its applicability to all substances.
In Petry, Alessi, Marx, Austin, and Tardif (2005) Petry (2005), 142 cocaine- or heroin-dependent outpatients were randomized among standard treatment, voucher CM, and prize CM. Both incentive conditions improved retention and documented abstinence durations relative to standard treatment, without a statistically significant difference between the two incentive methods in that sample. Their longer follow-up did not show clear between-group abstinence differences, although duration of continuous abstinence during treatment predicted later outcomes. That pattern motivates careful separation of active-treatment effects, post-treatment group differences, and prognostic associations.
Prize draws and fixed vouchers differ in their reinforcement distributions and in participant expectations, but both require documented contingent delivery. Benishek et al. (2014) Benishek (2014) synthesized prize-based randomized studies and reported an average end-of-treatment effect favoring CM, followed by reduced and uncertain longer-term group differences. The literature supports both formats as valid intervention families while leaving local choice dependent on evidence-based protocols, cost, patient acceptability, fidelity, and relevant funding rules.
Meta-Analyses and the Comparative Evidence Hierarchy
The Prendergast et al. (2006) meta-analysis Prendergast (2006) pooled 47 treatment-versus-control comparisons from studies published between 1970 and 2002 and reported a positive average treatment effect of d = 0.42 using its fixed-effects model. It documented variation across substances, study periods, and researcher involvement. The historical value of the synthesis lies as much in identifying moderators and design heterogeneity as in its pooled estimate: a number averaged across heterogeneous treatments is not a patient-specific predicted outcome.
Dutra et al. (2008) Dutra (2008) compared broad classes of psychosocial interventions for substance-use disorders. Such meta-analysis can establish an intervention's position within a field, but direct causal ranking of interventions remains difficult when trials differ in recruitment, intensity, background treatment, outcome definitions, and comparison strength. An intervention with the largest mean standardized effect in one synthesis is not automatically the best practical intervention for every clinical population.
A later Pfund et al. synthesis (2024 print publication; online 2022) Pfund (2024) reviewed five CM meta-analyses encompassing 84 studies and about 11,000 participants. Using its stated quality criteria, the authors characterized two component meta-analyses as moderate quality and three as low or critically low. For the moderate-quality syntheses, the summarized post-treatment abstinence effect was d = 0.54 with a 95% confidence interval of 0.43 to 0.64, while the pooled follow-up effect was d = 0.08 with interval 0.00 to 0.16. The latter value cautions against describing durable efficacy as a uniform large effect. Reviews of reviews also involve overlapping primary trials and cannot be read as 84 newly independent experiments.
Network Meta-Analysis and the CM–CRA Combination
De Crescenzo et al. (2018) De Crescenzo (2018) conducted a systematic review and network meta-analysis of psychosocial treatments for adults with cocaine or amphetamine addiction. The review compared structured therapies, including combinations, on abstinence and acceptability measures. Its authors highlighted CM combined with CRA as an important comparative reference for future trials. The result suggests that immediate reinforcement and broader environmental restructuring may complement one another, although network estimates depend partly on indirect comparison.
Network meta-analysis assumes sufficient comparability of trials across treatment networks. When the distribution of illness severity, concurrent medication, care setting, or intensity of the control differs across nodes, apparent treatment rankings may represent differences in context. Confidence in a combined intervention depends on the availability and quality of direct comparisons, estimates of inconsistency, and precise operational descriptions of what was delivered. The phrase CM plus CRA is insufficient without specifying the voucher contingency, counseling components, duration, contact hours, and external resources.
The scientific question is whether combining interventions produces an added effect beyond either component. A factorial experiment with CM present or absent and CRA present or absent offers a clean conceptual test, subject to feasibility and ethical constraints. More commonly, observed outcomes arise from multicomponent packages, with possible synergies that cannot be directly identified. A credible synthesis therefore presents the combined approach as a well-supported practical option in relevant stimulant contexts while treating the independent contribution of each element as an empirical question when a study has not isolated it.
Stimulant Use Disorders and the Clinical Evidence
Cocaine and methamphetamine use disorders are central to the modern CM literature because many controlled trials involve objective stimulant-use endpoints. The ASAM/AAAP guideline (2024) American Society of Addiction Medicine & American Academy of Addiction Psychiatry (2024) identifies CM as the current standard of care for stimulant-use disorder and gives a strong recommendation with high certainty for using CM as a primary component of the treatment plan in conjunction with psychosocial treatments. It lists CRA as a conditional complementary intervention with low certainty. Those certainty labels describe different bodies of evidence and should appear together, rather than being collapsed into one undifferentiated endorsement.
The WHO evidence recommendation updated in 2023 World Health Organization (2023) supports offering CM and cognitive behavioral therapy for psychostimulant dependence, with explicitly stated evidence certainty. These recommendations guide interpretation of a large and heterogeneous literature; they do not substitute for local diagnosis, safety assessment, or individualized clinical decision-making.
The biological verification of cocaine or methamphetamine abstinence is a relatively direct measure of the targeted behavior within a defined detection window. Yet clinical recovery additionally includes reduced overdose risk, physical health, psychosocial functioning, and continuing care. A person may improve clinically without immediately meeting an abstinence-oriented incentive criterion; a successful series of negative tests alone does not establish global functioning. Future trials should evaluate both the target behavior and broader, participant-relevant outcomes.
Opioid Use Disorders and Medication Integration
In opioid use disorder, evidence-based medications such as methadone and buprenorphine retain their own clinical roles. CM is properly studied as an adjunct to established opioid-use-disorder treatment, including problems such as concurrent stimulant use, attendance, and medication adherence. Reinforcing visits or drug-negative tests cannot substitute for evidence-based medication management or overdose prevention. A program that inadvertently makes essential medication contingent on incentive performance would cross a clinically important boundary.
Bolívar et al. (2021) Bolívar (2021) synthesized 74 randomized trials with 10,444 adults receiving medication for opioid use disorder; 60 studies contributed to quantitative analyses. Effects at treatment end favored CM for stimulant abstinence, polysubstance outcomes, illicit opioid abstinence, cigarette smoking, therapy attendance, and medication adherence. Reported standardized effect sizes varied by target: stimulant abstinence d = 0.70 (95% CI 0.49–0.92); illicit opioid abstinence d = 0.58 (0.30–0.86); therapy attendance d = 0.43 (0.22–0.65). These estimates came from different sets of trials and are not interchangeable measures of a single generalized recovery outcome.
The authors reported substantial between-study heterogeneity in several pooled categories, meaning that context and design deserve attention. The intervention's practical appeal is its capacity to address co-occurring behavioral treatment problems while medication addresses opioid dependence through distinct pharmacological mechanisms. The combined clinical system should safeguard continuity of prescribed treatment, confidentiality, overdose prevention and equitable access independently of any positive reinforcement schedule.
A 2026 systematic review and meta-analysis by Jegede and colleagues Jegede (2026) specifically addressed concurrent stimulant–opioid use rather than aggregating all outcomes across polysubstance samples. It identified 26 randomized trials including 2,356 participants, with trials primarily involving cocaine and heroin. In the trials contributing to each pooled endpoint, CM increased the longest continuous period of simultaneous stimulant-and-opioid abstinence by an average 1.42 weeks (95% CI 0.77–2.08) and increased the odds of combined stimulant/opioid-negative urine samples (OR 2.46, 95% CI 1.96–3.10). Overall treatment retention did not differ significantly. Critically, the included trials did not report fentanyl-specific exposure or outcomes. These data strengthen the evidence for simultaneous abstinence in the studied combinations while delimiting direct inference about contemporary fentanyl and methamphetamine co-use.
Alcohol, Cannabis, and Tobacco: Transfer Boundaries
Alcohol-focused CRA has a longer history than many modern drug-testing incentive programs. The original community-reinforcement studies situated drinking in the social economy of family, employment and recreation. Breath testing and attendance incentives are possible CM targets in alcohol services, but evidence from stimulant-only trials cannot be transplanted without checking the alcohol-specific studies, event frequency, baseline care, and feasibility of verification.
Cannabis-use research has examined behavioral interventions and sometimes CM in combination with cognitive or motivational components. The boundary condition is important because a favorable effect for cannabis-treatment engagement may differ from an effect on sustained biologically verified abstinence or long-term functioning. Participants' developmental stage, co-occurring conditions, frequency of visits, and the detection window for cannabis metabolites further complicate outcome interpretation. A program should not represent results from mixed treatment packages as a standalone CM or CRA effect.
Tobacco cessation has also been studied using contingent reinforcement, including during opioid medication treatment, as analyzed by Bolívar et al. (2021) Bolívar (2021). The magnitude and endpoints of tobacco studies should be kept separate from illicit-drug trial estimates. The synthesis favors diagnosis-specific protocol selection and careful reporting over treating substance-use disorders as a behaviorally uniform category.
Outcome Measurement and Endpoint Validity
Clinical outcome studies can measure point-prevalence abstinence, proportion of biologically negative samples, longest continuous abstinence, retention, session attendance, medication adherence, self-reported frequency, quality of life, and adverse events. These endpoints capture different aspects of recovery. Longest continuous abstinence rewards sustained sequences, while the proportion of submitted negative samples is sensitive to visit availability and missing-test rules. A trial reporting higher treatment attendance can demonstrate improved engagement even if its drug-use endpoint remains uncertain.
Randomized intention-to-treat comparisons are most informative when the full assigned sample is analyzed and attrition is transparently recorded. Biological verification reduces some reporting biases but introduces detection-window and missed-visit problems; self-report provides broader temporal and phenomenological information but may be affected by recall or social desirability. Agreement between measurement types is often more informative than privileging one mechanically.
The primary estimand should be stated before interpreting the result. Is a trial estimating the effect of being offered CM, receiving an adequate dose, attending enough sessions, or achieving early abstinence? Conditioning analyses on observed success can select participants whose prognosis was already better, producing a misleading mechanistic story. Missing data, censoring, differential access, and treatment switching all require explicit handling.
Long-Term Abstinence, Generalization, and Follow-Up
The durability question is central to the relationship between CM and CRA. CM creates contingent rewards during an intervention window; the sustainable benefit after rewards cease requires learning, new routines, altered expectations, or environmental conditions that persist. CRA directly targets longer-lived activities and social reinforcers, but durable benefit from a particular CRA component must be demonstrated in studies that measure that component and its consequences.
Ginley et al. (2021) Ginley (2021) meta-analyzed 23 randomized CM trials with objective illicit-drug-use indicators obtained up to one year after incentives ended. The pooled odds ratio for abstinence at follow-up was 1.22 (95% CI 1.01–1.44), with I-squared about 36.7%. Longer active treatment duration emerged as a significant moderator of follow-up outcome in their analysis. These findings support a modest average lasting benefit in the reviewed sample of trials, without establishing that every program or individual sustains abstinence.
A different Pfund et al. (2024) Pfund (2024) synthesis produced a comparatively small aggregated follow-up standardized mean effect. These results need not be treated as a simple factual contradiction: they differ in inclusion criteria, source overlap, outcomes, effect metrics, and synthesis strategy. They jointly warn against replacing follow-up evidence with either a categorical claim that effects vanish immediately or an assurance that short-term success persists automatically.
Treatment Dose, Timing, and Reinforcement Schedules
A CM protocol has a behavioral dose measured not only by calendar weeks but by the number of opportunities for qualifying behavior, testing frequency, delay to reinforcement, accessible reward value, and the continuity of participant engagement. Two nominally 12-week programs can expose participants to different numbers of verified opportunities and very different amounts of reinforcement. The dose can be delivered imperfectly through delayed staff action, missed appointments, stock shortages, or unclear reset rules.
Schedule comparisons must also account for participant characteristics, baseline use severity, and the burden of testing. An escalating schedule can make continuity more valuable and missed opportunities costly, but its effect may be attenuated when barriers unrelated to drug use generate missed visits. A program should explicitly distinguish a positive test from an uncollected test and document accommodations without quietly changing the target. Fairness and scientific interpretability both depend on this distinction.
The decision to end incentives creates a transition problem. Rapid withdrawal of external rewards may uncover unmet social and occupational needs; extending incentives indefinitely raises cost and administrative questions. Trials comparing tapering, extended duration, or coordinated transition into naturally reinforced community activities are important for identifying how treatment gains can be maintained with fewer externally financed consequences.
Retention, Therapeutic Alliance, and Participant Experience
CM can improve participation by making the purpose and consequences of appointments concrete. However, incentives interact with the therapeutic relationship. A clinician can offer transparent, respectful reinforcement as one part of collaborative care or use the identical mechanism in a manner that feels coercive and punitive. Participant perceptions, trust, language access, and privacy are therefore legitimate variables in the effectiveness of a real program.
Treatment retention is both an outcome and a pathway to additional care. Someone who attends more frequently may receive more psychosocial counseling, medication monitoring, infection prevention, and crisis support. Analyses should examine whether improved drug-use outcomes remain when treatment exposure differs, while recognizing that simple statistical control for post-randomization attendance can introduce bias. Programs should specify whether retention is the intended incentivized behavior or a hypothesized downstream effect.
Qualitative evidence about participant experience is important but should be interpreted in relation to its sampling frame. An interview series drawn only from completers could overstate acceptability; interviews of people who declined or left treatment may reveal privacy, financial, cultural or logistical obstacles. Therapeutic acceptability is measurable through informed consent, uptake, continued participation, respectful delivery, and participant-reported benefit, rather than inferred from the clinical investigator's favorable description.
The Clinical Program as a Complex Intervention
An implemented CM–CRA program is a system of people, verification methods, appointments, reward inventory, social services, staffing, and funding. Scientific fidelity includes delivering the agreed contingency consistently, confirming the correctness of results, preserving records, training clinicians, and auditing the participant's actual reward access. CRA fidelity includes competent functional analysis and meaningful provision of alternative reinforcement, rather than merely assigning home activities.
The Petry et al. (2005) multisite trial Petry (2005) and the HHS/ASPE implementation report (2024) U (2024) provide different kinds of evidence: controlled effectiveness in real programs, and synthesis of service-system procedures and safeguards. Implementation research must preserve that distinction between experimentally estimated clinical efficacy and administrative reports about program feasibility.
The most important generalizable operational lesson is the specificity of the intervention. A clinic that offers occasional discretionary prizes to clients it considers deserving has implemented a different causal arrangement from a protocol awarding prompt, documented incentives for an agreed event. Likewise, a clinic calling routine counseling CRA without structured assessment of daily reinforcers creates uncertainty about treatment identity. Reliable terminology is indispensable to meaningful comparison.
Implementation Science, Staff Training, and Fidelity
The scientific credibility of an intervention depends on the way it is enacted. A protocol must define the behavior eligible for reinforcement, the testing method, the acceptable interval between confirmation and reward, the reward inventory, and the clinician's response to a missed or disputed result. For community reinforcement, it must also identify the functional assessment, participant-selected rewarding alternatives, skills practice, and the real service pathways that help make these alternatives possible. Supervision can monitor these elements without assuming that identical scripts fit every cultural and socioeconomic context.
Kirby, Benishek, Dugosh, and Kerwin (2006) Kirby (2006) documented addiction-treatment providers' beliefs and objections about CM. Provider attitudes are part of the implementation environment; skepticism may reflect incomplete understanding of learning principles, concern over money, distrust of clients, or frustration with restrictive funding. Training that simply announces the effectiveness of incentives is unlikely to address all such concerns. A scientifically oriented implementation program combines education with demonstrations of contingency fidelity, respectful language, realistic budgets, and transparent evaluation.
The ASAM/AAAP guideline (2024) American Society of Addiction Medicine & American Academy of Addiction Psychiatry (2024) identifies staff preparation, reliable testing, sufficient incentive magnitude, and frequent clinical contact as essential operational resources. The U.S. Department of Health and Human Services' 2024 CM report U (2024) likewise considers service quality, access and program integrity. A small incentive combined with slow redemption and inconsistent verification may amount to a poorly delivered version of the intervention, so an ineffectual local implementation cannot automatically falsify outcomes from adequately delivered experimental protocols.
Economic Evaluation and Value of Incentives
The apparent cost of CM is often described as the total face value of rewards, but an economic evaluation needs a wider perspective. It must include staff time, training, laboratory tests, verification systems, technology fees, administrative oversight, lost appointments, participant travel, and the costs of managing funds. It should also consider the health and social outcomes the program changes: use days, treatment continuity, emergency care, occupational participation, justice-system involvement, and quality of life. The timeframe and payer perspective determine which costs and benefits are counted.
A fixed budget can support different reinforcement schedules, participant numbers, or durations. Comparing only the maximum value a patient could theoretically earn obscures average redemptions and variation in accessibility. Moreover, the reward is a transfer of purchasing power to a participant, while the resources used to deliver and audit it represent additional social and administrative costs. A sound cost-effectiveness analysis must specify which perspective it adopts and avoid conflating those terms.
The practical question is not whether every program produces net savings. It is whether a specified evidence-supported contingency, in a specified population, achieves a valued clinical outcome at an acceptable incremental resource cost relative to feasible alternatives. Independent prospective economic data and transparent sensitivity analyses are especially useful for comparing voucher versus prize-based systems, contact frequencies, and remote versus in-person verification.
United States Financing, Compliance, and Service Integrity
American implementation has historically encountered concerns about federal funding restrictions, allowable reward values, fraud prevention, billing rules, and confusion between a therapeutic incentive and an improper inducement. The policy context evolves independently of the clinical evidence. A SAMHSA funding notice issued in January 2025 Substance Abuse and Mental Health Services Administration (2025) allowed qualifying grantees under programs authorizing CM to provide motivational incentives up to $750 per patient annually subject to particular safeguards, replacing the earlier $75 limit. The January 2025 funding change initially applied to State Opioid Response and Tribal Opioid Response grants under updated terms. This is a conditional grant rule, not a generally applicable payment standard or a guaranteed reimbursement entitlement. Eligible items and vouchers are distinguished from impermissible cash payments.
Accordingly, a service needs its current grant terms, payer requirements, fraud-and-abuse analysis, audit trail, and participant eligibility rules before implementation. No historical national ceiling should be treated as perpetual across all federal programs, Medicaid authorities, private insurers, or future regulatory changes. Clinical effectiveness data cannot alone authorize a particular financial mechanism. A compliant program also needs to protect against reward diversion, discriminatory allocation, inaccurate test recording, or incentives for behaviors that staff cannot verify.
California's Recovery Incentives Program California Department of Health Care Services (n.d.) illustrates the development of a public-sector stimulant-use CM benefit within Medicaid-linked delivery arrangements. Statewide implementation may test the effectiveness of training and supervision systems at scale, but aggregate enrollment is not automatically equivalent to randomized evidence of abstinence effectiveness. Implementation reporting should distinguish program uptake, individual outcomes, provider fidelity, and changes in population-level harms.
Equity, Poverty, Access, and Treatment Justice
A treatment that pays rewards only after frequent clinic visits may unintentionally distribute benefits according to transportation access, work flexibility, disability accommodation, childcare, and local laboratory availability. These are characteristics of the clinical environment, not indicators of a participant's moral commitment. Equity analysis therefore asks who can practically qualify for rewards, who is enrolled, who drops out, and whether financial or administrative friction affects the estimated treatment effect.
Incentive value may also differ with household resources and immediate material need. An identical voucher can be highly meaningful to one participant and of little value to another. Allowing a carefully governed, equivalent-value choice among relevant goods or services may increase acceptability without abandoning a transparent schedule. An intervention that assumes a stable telephone, banking access, identity document, or private address will miss some of the people who face substantial health burdens.
Cultural adaptation and disability access are further design questions requiring empirical evaluation. Written rules should be available in comprehensible language, testing should protect privacy, redemption should be accessible, and participant appeals should be possible when administrative errors affect reward delivery. Program-level disparities in enrollment, qualifying events, payouts, and discontinuation should be audited as actual measurable outcomes rather than inferred from benevolent intentions.
Remote Verification and Digital Contingency Management
Remote services can reduce travel time and make reinforcement more immediate. The scientific task is to verify behavior with sufficient reliability while preserving privacy and minimizing surveillance burdens. Digital approaches may use video-observed sample collection, remotely connected monitoring devices, secure laboratory data transfer, or appointment verification. Each mechanism changes the relation among accuracy, convenience, cost and trust, and should therefore be evaluated on its own merits rather than grouped under the vague label digital care.
The ASAM/AAAP guideline American Society of Addiction Medicine & American Academy of Addiction Psychiatry (2024) allows evidence-based digital behavioral interventions as adjunctive components for stimulant-use disorder and recommends considering telemedicine when in-person access is difficult. Such guidance does not establish a positive outcome for every commercial application. Credible evaluation requires comparative trials, clearly defined data handling, independent fidelity audits, and analysis of who is excluded by technology requirements.
Artificial systems can support scheduling, protocol checks, missing-data review, quality assurance, and accessibility tools in the Artificial Era. These are proposed operational roles in a human-governed clinical system. An algorithm that decides reward eligibility on the basis of opaque predictions introduces new sources of error and inequity. Future human–artificial clinical partnerships should be tested for reproducibility, transparency, participant appeal, and effects on actual treatment outcomes before autonomous decisions are normalized.
Integration With Other Psychological and Medical Care
CM is compatible with cognitive behavioral therapy when incentives reinforce measurable treatment-linked actions and cognitive behavioral sessions address coping, learning, and relapse risk. CRA addresses the accessibility and attractiveness of drug-free activities and social contexts, while motivational interviewing can explore ambivalence and patient preferences. The ASAM/AAAP recommendations American Society of Addiction Medicine & American Academy of Addiction Psychiatry (2024) name CBT and CRA as preferred partners of CM, with distinctly graded evidence. A combined program's increased contact hours and clinical attention can themselves influence outcomes, so comparison designs should address these differences.
Opioid use disorder care illustrates the need for interdisciplinary coordination. CM may improve attendance or co-occurring stimulant abstinence among persons already receiving medications for opioid use disorder, as in the Bolívar et al. (2021) evidence synthesis Bolívar (2021). Medication continuation, overdose-risk education, infection treatment, and emergency support should remain clinically available according to their own indications. The CM mechanism is an addition to integrated treatment, not a prerequisite for basic healthcare.
Other substance-use disorders have different measurement windows, medication options, and patterns of harm. Clinical teams should distinguish the reward target from the patient's complete care plan and avoid using abstinence test results to suppress necessary symptom management. Treatment design is strongest when personal goals, evidence, medical safety, and objective program integrity are aligned through consent and regular reassessment.
Safety, Consent, and the Difference Between Reinforcement and Punishment
A clinically appropriate incentive is delivered under prospective, knowable conditions. Withholding an unearned optional reward is analytically different from taking away a person's housing, medication, essential food, or legal rights. Programs should keep this distinction explicit. Punitive sanctions and unpredictable loss of necessary services can increase shame, destabilize treatment, and create incentives to conceal use; they are not the beneficial reinforcement mechanism tested in CM trials.
Biological testing raises confidentiality, false-positive, detection-window, and surveillance concerns. Decisions should be based on validated methods, written protocols, individualized clinical context, and confirmatory procedures when warranted. A disputed result should have a documented pathway for review. Special populations, including adolescents, pregnant persons, and people under criminal-justice supervision, may face additional coercion or disclosure risks. Consent and legitimate alternatives matter even when objective abstinence data are scientifically useful.
Attention to safety includes avoiding stigmatizing attributions and honoring patient-defined outcomes. A return to substance use may prompt renewed assessment, prevention planning, and continued care. It should not be treated as proof of personal failure. The empirical claim that reinforcement can modify a target behavior does not erase the ethical requirement for voluntary, fair, clinically proportionate treatment.
Causal Identification and Methodological Problems
Randomization is the strongest available design for estimating the effect of being offered a specified CM package compared with a defined alternative, but it does not automatically isolate all mechanisms. If both reward opportunities and contact frequency change, the estimate belongs to the whole package. A dismantling design could vary reward contingency while matching contact and expected benefit, and an adequately powered factorial trial could evaluate CM and CRA main effects and interactions. Such designs are needed to identify additive and synergistic mechanisms.
The biological assay is often treated as an objective endpoint, but the probability of observing a sample itself depends on transportation, engagement, scheduling, and the treatment's effect on attendance. A missing specimen can reflect active use, missed transport, illness, or departure from care. Statistical analyses should state which estimand they pursue, whether missing samples are classified as positive, and how alternative assumptions affect conclusions. Point-prevalence abstinence, consecutive abstinence, frequency reduction and psychosocial functioning require different measurement logic.
A meta-analysis must consider trial overlap, comparators, sample differences, effect metric, risk of bias, selective outcomes, funding, and publication bias. Prendergast et al. (2006) Prendergast (2006), De Crescenzo et al. (2018) De Crescenzo (2018), Ginley et al. (2021) Ginley (2021), and Pfund et al. (2024) Pfund (2024) answer related but distinct questions. Their effect estimates should be read in relation to each review's included participants, outcomes, follow-up points, and design. The current monograph is an interpretive critical synthesis, without a new pooled estimate or a claim to PRISMA-compliant exhaustive retrieval.
Disputed Explanations and Theoretical Limitations
One disputed question is whether external reinforcement undermines intrinsic motivation. The empirical literature on addiction CM should be judged through observed clinical outcomes rather than an assumed universal motivational law. A reward may facilitate entry into treatment and make early abstinence more achievable, while the availability of meaningful longer-term activities determines whether alternatives become self-sustaining. The relevant experiments compare specified schedules and assess functioning after the reward period.
A second dispute concerns the durability of effects. Short-term trials often find more robust gains than some follow-up syntheses. Yet Ginley et al. (2021) Ginley (2021) found a modest average long-term advantage in trials with objective follow-up measures; the review reported study-quality and missing-data limitations. It is consequently more precise to expect heterogeneity and to investigate duration and context than to assert either permanent success or complete immediate extinction of benefit.
A third dispute concerns what counts as improvement. Full abstinence from one target drug is a legitimate clinical outcome for some patients, but reductions in use, avoidance of overdose, better functioning, treatment retention, safer environments, and voluntary engagement may also matter. A CM program reinforcing a single biological endpoint must be evaluated on that endpoint and then separately on broader quality of life and health. Neither endpoint type can stand in for all others.
Research Agenda: From Incentives to Sustained Community Reinforcement
A high-value research program would compare adequately specified CM, CRA, combined CM–CRA, and matched alternatives in pragmatic factorial trials while measuring treatment fidelity and participant-selected outcomes. It would prospectively define both in-treatment and post-treatment endpoints, include objective testing with transparent missing-data sensitivity analyses, and evaluate meaningful life functioning and adverse effects. Recruitment should reach groups historically excluded from specialist experimental programs, including populations with unstable housing, rural access barriers and multiple co-occurring conditions.
A second program would examine mechanisms: Does early biologically verified abstinence enable participation in alternative community reinforcement? Which employment, social, leisure and family opportunities predict durable change? Are measured changes in access to alternatives causal mediators or merely indicators of improved prognosis? Longitudinal and experimental manipulation of community supports could distinguish these possibilities more convincingly than correlations alone.
A third program would evaluate implementation at scale under public funding and digital delivery. It would compare reward size, timeliness, schedule, testing burden, participant choice and fairness, with independent cost-effectiveness and equity analyses. For human–artificial systems, the relevant research questions concern whether transparent automation reduces delay and error without increasing exclusion or intrusive surveillance. These questions place technology in a defined scientific role without mistaking a proposed engineering application for established clinical efficacy.
Conclusion
Contingency management and community reinforcement organize recovery as a change in the consequences and opportunities available to a person. Their scientific history begins with reinforcement-based restructuring of alcohol-related social environments and continues through randomized cocaine voucher studies, multisite stimulant trials, meta-analyses, and the contemporary implementation of structured incentive programs. The interventions operate at complementary levels. CM supplies a reliable and timely consequence after an agreed, verified behavior. CRA develops social, vocational, interpersonal, and recreational opportunities that can sustain a more rewarding pattern of life.
The strongest contemporary clinical evidence for CM concerns stimulant-use disorder and supports its use as a primary psychosocial component, with the comparative certainty for CRA as a companion intervention more limited. Combinations of CM and CRA have shown favorable results in network comparisons, but inference about their interaction requires close inspection of trial design and comparator. Follow-up studies suggest that some benefits can persist after incentives end, with effects that vary across protocols and samples. Long-term psychosocial functioning and the causal role of natural reinforcement remain important research objectives.
The scientific synthesis establishes practical criteria for accountable implementation: a real contingency, timely verifiable delivery, adequately valued rewards, treatment accessibility, safeguards against punishment and exclusion, reliable evaluation, and meaningful recovery alternatives. Within the Artificial Era, digital assistance and human–artificial coordination can be studied as tools for making these criteria more consistently achievable. Clinical authority remains accountable to verifiable evidence, informed participant choice, and the outcomes actually measured.
References
- American Society of Addiction Medicine & American Academy of Addiction Psychiatry. (2024). The ASAM/AAAP clinical practice guideline on the management of stimulant use disorder. Journal of Addiction Medicine, 18(1), 1–56. https://doi.org/10.1097/ADM.0000000000001299
- Azrin, N. H. (1976). Improvements in the community-reinforcement approach to alcoholism. Behaviour Research and Therapy, 14(5), 339–348. https://doi.org/10.1016/0005-7967(76)90021-8
- Benishek, L. A., Dugosh, K. L., Kirby, K. C., Matejkowski, J., Clements, N. T., Seymour, B. L., & Festinger, D. S. (2014). Prize-based contingency management for the treatment of substance abusers: A meta-analysis. Addiction, 109(9), 1426–1436. https://doi.org/10.1111/add.12589
- Bolívar, H. A., Klemperer, E. M., Coleman, S. R. M., DeSarno, M., Skelly, J. M., & Higgins, S. T. (2021). Contingency management for patients receiving medication for opioid use disorder: A systematic review and meta-analysis. JAMA Psychiatry, 78(10), 1092–1102. https://doi.org/10.1001/jamapsychiatry.2021.1969
- California Department of Health Care Services. (n.d.). Recovery Incentives Program: California's contingency management benefit. https://www.dhcs.ca.gov/recovery-incentives-program-californias-contingency-management-benefit/
- De Crescenzo, F., Ciabattini, M., D'Alò, G. L., De Giorgi, R., Del Giovane, C., Cassar, C., Janiri, L., Clark, N., Ostacher, M. J., & Cipriani, A. (2018). Comparative efficacy and acceptability of psychosocial interventions for individuals with cocaine and amphetamine addiction: A systematic review and network meta-analysis. PLOS Medicine, 15(12), e1002715. https://doi.org/10.1371/journal.pmed.1002715
- Dutra, L., Stathopoulou, G., Basden, S. L., Leyro, T. M., Powers, M. B., & Otto, M. W. (2008). A meta-analytic review of psychosocial interventions for substance use disorders. American Journal of Psychiatry, 165(2), 179–187. https://doi.org/10.1176/appi.ajp.2007.06111851
- Ginley, M. K., Pfund, R. A., Rash, C. J., & Zajac, K. (2021). Long-term efficacy of contingency management treatment based on objective indicators of abstinence from illicit substance use up to 1 year following treatment: A meta-analysis. Journal of Consulting and Clinical Psychology, 89(1), 58–71. https://doi.org/10.1037/ccp0000552
- Higgins, S. T., Budney, A. J., Bickel, W. K., Foerg, F. E., Donham, R., & Badger, G. J. (1994). Incentives improve outcome in outpatient behavioral treatment of cocaine dependence. Archives of General Psychiatry, 51(7), 568–576. https://doi.org/10.1001/archpsyc.1994.03950070060011
- Higgins, S. T., Budney, A. J., Bickel, W. K., Hughes, J. R., Foerg, F., & Badger, G. (1993). Achieving cocaine abstinence with a behavioral approach. American Journal of Psychiatry, 150(5), 763–769. https://doi.org/10.1176/ajp.150.5.763
- Hunt, G. M., & Azrin, N. H. (1973). A community-reinforcement approach to alcoholism. Behaviour Research and Therapy, 11(1), 91–104. https://doi.org/10.1016/0005-7967(73)90072-7
- Jegede, O. O., Oliva, H. N. P., Prudente, T. P., Rash, C. J., Muvvala, S. B., & Angarita, G. A. (2026). Contingency management for stimulant-opioid co-use: A systematic review and meta-analysis. Journal of Addiction Medicine. Advance online publication. https://doi.org/10.1097/ADM.0000000000001708
- Kirby, K. C., Benishek, L. A., Dugosh, K. L., & Kerwin, M. E. (2006). Substance abuse treatment providers' beliefs and objections regarding contingency management: Implications for dissemination. Drug and Alcohol Dependence, 85(1), 19–27. https://doi.org/10.1016/j.drugalcdep.2006.03.010
- Meyers, R. J., & Smith, J. E. (1995). Clinical guide to alcohol treatment: The community reinforcement approach. Guilford Press. https://www.guilford.com/books/Clinical-Guide-to-Alcohol-Treatment/Meyers-Smith/9780898628579/summary
- Meyers, R. J., Miller, W. R., Hill, D. E., & Tonigan, J. S. (1998). Community reinforcement and family training (CRAFT): Engaging unmotivated drug users in treatment. Journal of Substance Abuse, 10(3), 291–308. https://doi.org/10.1016/S0899-3289(99)00003-6
- Miller, W. R., Meyers, R. J., & Tonigan, J. S. (1999). Engaging the unmotivated in treatment for alcohol problems: A comparison of three strategies for intervention through family members. Journal of Consulting and Clinical Psychology, 67(5), 688–697. https://doi.org/10.1037/0022-006X.67.5.688
- Petry, N. M., Alessi, S. M., Marx, J., Austin, M., & Tardif, M. (2005). Vouchers versus prizes: Contingency management treatment of substance abusers in community settings. Journal of Consulting and Clinical Psychology, 73(6), 1005–1014. https://doi.org/10.1037/0022-006X.73.6.1005
- Petry, N. M., Peirce, J. M., Stitzer, M. L., Blaine, J., Roll, J. M., Cohen, A., Obert, J., Killeen, T., Saladin, M. E., Cowell, M., Kirby, K. C., Sterling, R., Royer-Malvestuto, C., Hamilton, J., Booth, R. E., Macdonald, M., Liebert, M., Rader, L., Burns, R.,. Li, R. (2005). Effect of prize-based incentives on outcomes in stimulant abusers in outpatient psychosocial treatment programs: A National Drug Abuse Treatment Clinical Trials Network study. Archives of General Psychiatry, 62(10), 1148–1156. https://doi.org/10.1001/archpsyc.62.10.1148
- Pfund, R. A., Ginley, M. K., Boness, C. L., Rash, C. J., Zajac, K., & Witkiewitz, K. (2024). Contingency management for drug use disorders: Meta-analysis and application of Tolin's criteria. Clinical Psychology: Science and Practice, 31(2), 136–150. https://doi.org/10.1037/cps0000121
- Prendergast, M., Podus, D., Finney, J., Greenwell, L., & Roll, J. (2006). Contingency management for treatment of substance use disorders: A meta-analysis. Addiction, 101(11), 1546–1560. https://doi.org/10.1111/j.1360-0443.2006.01581.x
- Roozen, H. G., Boulogne, J. J., van Tulder, M. W., van den Brink, W., De Jong, C. A. J., & Kerkhof, A. J. F. M. (2004). A systematic review of the effectiveness of the community reinforcement approach in alcohol, cocaine and opioid addiction. Drug and Alcohol Dependence, 74(1), 1–13. https://doi.org/10.1016/j.drugalcdep.2003.12.006
- Smith, J. E., Meyers, R. J., & Delaney, H. D. (1998). The community reinforcement approach with homeless alcohol-dependent individuals. Journal of Consulting and Clinical Psychology, 66(3), 541–548. https://doi.org/10.1037/0022-006X.66.3.541
- Smith, J. E., Meyers, R. J., & Miller, W. R. (2001). The community reinforcement approach to the treatment of substance use disorders. American Journal on Addictions, 10(Suppl. 1), S51–S59. https://doi.org/10.1080/10550490150504137
- Substance Abuse and Mental Health Services Administration. (2025). Using SAMHSA funds to implement evidence-based contingency management services. https://www.samhsa.gov/resource/ebp/using-samhsa-funds-implement-evidence-based-contingency-management-services
- U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. (2024). Contingency management for the treatment of substance use disorders: Enhancing access, quality, and program integrity for an evidence-based intervention. https://www.ncbi.nlm.nih.gov/books/NBK606622/
- World Health Organization. (2023). Psychosocial support for the management of psychostimulant use disorders. https://www.who.int/teams/mental-health-and-substance-use/treatment-care/mental-health-gap-action-programme/evidence-centre/drug-use-disorders/psychosocial-support-for-the-management-of-psychostimulant-use-disorders/