Behavioral Treatment of Drug and Alcohol Abuse

Maxine L. Stitzer and Stephen T. Higgins

Drug abuse treatments generally derive from a conceptual and theoretical definition of the problems to be treated and of their underlying causes. Two primary schools of thought have been advanced to explain the clinical characteristics of substance dependence and have led to different treatment approaches. These contrasting views have been reviewed and discussed more fully by Marlatt and Gordon (43). The "disease model" postulates that substance-dependent individuals have a biological abnormality (possibly genetic) that predisposes them to seek and use chemical substances to excess. In this view, drug-seeking behavior is outside the control of the afflicted individual. The treatment approach that has derived from the disease model emphasizes admitting powerlessness over drugs or alcohol, accepting total abstinence as the goal of change, and adopting the norms and values of a new social group, the Alcoholics Anonymous (AA) self-help group, in order to achieve and sustain abstinence (20). The disease model is highly regarded by clinicians, but it has not been extensively researched and therefore currently lacks strong empirical support.

The "learning model," which flows mainly from an academic research tradition, rests on recent scientific evidence that drugs act as primary reinforcers directly on brain reward systems and that orderly learning processes, both classical and operant, underlie the acquisition and maintenance of drug self-administration behavior. Treatments developed under a learning model acknowledge the chronic relapsing nature of dependence, as well as the important role of the physical and social environment in promoting relapse versus abstinence. Furthermore, these treatments call for development of new behavioral strategies on the part of the substance abuser, as well as reorganization of the physical and social environment, to counteract relapse tendencies engendered by learned drug associations. Consistent with its roots in academia, the behavioral approach has received considerable evaluation, much of which supports the effectiveness of this general approach to treatment. The purpose of the present chapter is to review recent developments in learning-based approaches to treatment of heroin, cocaine, and alcohol dependence (see Adaptive Processes Regulating Tolerance to Behavioral Effects of Drugs and Animal Models of Drug Addiction).

Methadone is an important modality for opioid abusers, with close to 100,000 people presently in treatment within the United States (56). Methadone treatment has demonstrated efficacy in suppressing heroin use and associated criminal activity (5, 31), but does not directly address the full range of problems that drug abusers bring to treatment. Polysubstance abuse is common among methadone patients, with illicit use of cocaine (12, 24, 38) replacing benzodiazepines (33, 71) as the most serious and prevalent secondary substance of abuse. Nevertheless, methadone brings polydrug-abusing patients into a therapeutic setting and provides the context for contingency management procedures that are well-suited to addressing the clinical problem of illicit drug supplementation, and that have been shown effective for improving treatment outcome. Methadone treatment also provides an excellent model for the integration of behavioral and pharmacological treatments. The use of contingency management procedures in methadone programs demonstrates how behavioral and pharmacological treatments, which address different aspects of the clinical problems of drug abusers, can be successfully integrated to produce outcomes that are better than those obtained with either therapy alone.

A number of potential rewards and punishments can be identified within the context of a daily methadone dispensing clinic for use in contingency management procedures. The methadone take-home privilege, in which an extra daily dose of methadone is dispensed to the patient for ingestion on the following day, offers a convenient and valued (by clients) incentive for use in abstinence reinforcement protocols (68) and is one of the most potent positive reinforcers available within the context of routine clinic operation. An early study by Stitzer et al. (70), as well as a more recent study by Iguchi et al. (34), examined take-home incentives in methadone patients who chronically supplemented with benzodiazepines. When take-home privileges could be earned for providing drug-free urines, temporary abstinence was observed in about 50% of study patients during the contingent take-home intervention lasting 12–20 weeks. Magura et al. (40) found that 1-month contracting for contingent take-home privileges resulted in 34% of their polydrug-abusing subjects achieving abstinence, whereas Milby et al. (51) found a similar percentage of clients responding to a take-home incentive program with increased numbers of consecutive drug-free urines, as required by the contingent intervention. Most of these studies focused on selected groups of identified polydrug abusers and used within-subject designs to evaluate effectiveness of contingent take-home programs for improving treatment outcomes.

A more recent controlled clinical trial was conducted by Stitzer et al. (72) to examine take-home incentive effects in 54 newly admitted methadone maintenance patients randomly assigned to receive take-home privileges under contingent or noncontingent conditions. In the contingent condition, the first take-home privilege was available after a relatively short (2-week) period of demonstrated abstinence from supplemental drugs, in an attempt to provide an immediate reward for positive behavior change. The conditional probability that a patient would improve on drug use was 2.5 times greater for the contingent than for the noncontingent study condition, whereas the probability of worsening on the drug use measure was two times greater for the noncontingent than for the contingent group. Overall, 32% of contingent patients achieved sustained periods of abstinence during the intervention (mean 9.4 weeks; range 5–15 weeks). The beneficial effect of contingent take-home delivery was replicated within the group of noncontingent patients who switched to the contingent intervention after their 6-month evaluation in the main study (partial crossover design). In this case, 28% improved substantially and achieved, on average, 15.5 drug-free weeks. In both the main study and the partial crossover, as well as in a recent analysis of the characteristics of take-home earners in clinical practice (37), lower rates of drug-positive urines early in treatment predicted improvement under the contingent take-home program, but patients using cocaine and benzodiazepines were equally likely to respond.

A study by McCaul et al. (47) examined effects of an incentive package on relapse to opiate drug use during a 90-day ambulatory methadone detoxification. Patients selected for the study had submitted {ewc MVIMG, MVIMAGE,!lesseq.bmp}50% opiate-positive urines during a baseline period and were randomly assigned to a control or contingent intervention condition. In the latter, opiate-free urines resulted in $10 cash and a take-home day, whereas opiate-positive urines resulted in increased counseling contact, urine sample collection, and data questionnaire requirements for patients. The contingent procedure was shown to be effective for promoting sustained periods of opiate abstinence and delaying relapse. Reductions in drug use during ambulatory detoxification were also reported in a controlled trial by Hall et al. (21) using small amounts of money ($4 to $10 per sample) as the reward for drug-free urines.

Methadone dose changes have also been used in contingent arrangements and have been shown to be effective for suppressing supplemental drug use. Impressive suppression of opiate use during ambulatory methadone detoxification was achieved in a study by Higgins et al. (30) under conditions where the contingent incentive for opiate-free urines was the opportunity to increase the methadone dose by up to 20 mg, an opportunity that remained active only so long as urines were opiate-free. Suppression of opiate use was shown to be specifically related to the contingency between increased methadone dose and drug use, because the benefits were not apparent in a group that could receive noncontingent increases in their methadone dose. Another study by Stitzer et al. (69) extended evaluation of dose change incentives by showing that decreases in polydrug abuse could be demonstrated during methadone maintenance both when methadone dose was increased above original maintenance levels as a result of drug-free urines and when dose was decreased below original maintenance levels as a consequence of drug-positive urines.

Structured contingency programs appear to hold considerable promise as a means to begin making inroads into the chronic supplemental drug use of methadone patients. The studies reviewed demonstrate that incentives, including methadone take-home privilege, money, and methadone dose changes, when used in a contingent fashion, can delay relapse during detoxification and promote improvements in supplemental drug use during treatment. However, the effectiveness of these contingent incentive programs generally appears limited to patients with less severe polydrug abuse. Future research needs to focus on ways to optimize the utility and cost-effectiveness of incentives that are readily available in the context of methadone clinic operation, to further characterize patients who do and do not respond to treatment, and to develop and evaluate more potent reinforcers for their ability to influence the drug use of more severely dependent polydrug abusers.

Another commonly used intervention that has received research evaluation is the contingent availability of further methadone treatment. Treatment termination contracting provides a means of formulating specific behavioral improvement objectives for poorly performing methadone patients, with the consequence of noncompliance being dose reduction and termination from treatment. Liebson et al. (39), in an elegant early study, showed that treatment outcome could be dramatically improved for severely alcoholic methadone patients by using the threat of treatment termination to motivate participation in a monitored disulfiram program at the clinic. McCarthy and Borders (46) showed that structured treatment involving the threat of termination for failure to meet specified standards of drug-free urine submissions could improve outcomes on measures of opiate drug use during treatment. Two other studies using pre- versus post-intervention evaluation designs have demonstrated the effectiveness of contingent treatment termination approaches. Dolan et al. (14) showed that 50% of study patients (N = 21), all with histories of demonstrated failure at controlling illicit drug supplementation, improved and submitted drug-free urines during a 30-day treatment contracting period. Similarly, Saxon et al. (66) have recently shown that 40% of a VA methadone maintenance population with ongoing supplemental drug use during treatment responded to termination contracts by slowing or stopping illicit drug use. In both the Dolan and Saxon studies (15, 66), patients most likely to succeed had been in treatment a longer time, submitted relatively fewer drug-positive samples prior to contract initiation, and were less likely to be abusing cocaine or an opiate–cocaine combination than other classes of illicit drugs. Thus, although treatment termination contracting is clearly effective with some proportion of methadone patients, these interventions pose an ethical dilemma, because they inevitably result in voluntary or involuntary treatment termination for those patients who will not or cannot conform to the demands of the contingent treatment contract (34, 69).

Overall, the studies reviewed indicate that both positive and negative incentives can be used in contingent arrangements to promote at least temporary periods of abstinence from supplemental drug use among methadone patients. Both approaches are effective with a subgroup of patients (generally 30–50% of the total treatment sample), who are typically the less severely dependent polydrug abusers. However, positive incentives have the advantage of keeping patients in treatment, whereas negative incentives, particularly those involving methadone dose decrease or threat of treatment termination, result in treatment dropout. One interesting issue that deserves additional research is whether there are any patient characteristics associated with response to positive versus negative incentive procedures. A second issue is whether multiple incentives can be utilized together to increase the potency of contingency management procedures for improving outcomes during methadone treatment.

Ever since its inception in the late 1960s, drug abuse counseling has been a part of methadone treatment. Only recently, however, has research begun to define the elements of drug abuse counseling and to assess the effectiveness of these counseling services. In a recent landmark study, McLellan et al. (50) have specifically investigated treatment outcomes for methadone patients as a function of the amount and type of counseling treatment offered. Male intravenous-opiate-abusing veterans (N = 92) were randomly assigned to receive methadone with minimal, standard, or enhanced counseling services. The minimal counseling group received perfunctory contact with therapists focused on satisfying routine requests for information transfer to outside agencies, requests for methadone dose change, and enforcement of program rules. These patients could receive take-home privileges based on employment and independent of urine test results. Standard and enhanced care patients met routinely with counselors to discuss their drug use and life adjustment issues. Treatment for these groups included contingencies targeted on drug use; take-home medication privileges could be earned by employed clients who were also drug-free. Furthermore, the number of counseling sessions per week was increased for those submitting drug-positive urines.

A primary outcome examined during the 24-week study was the percent of patients meeting predetermined criteria for treatment failure. The criteria included unremitting illicit drug use and/or multiple emergency situations requiring immediate health care. Sixty-nine percent of patients receiving methadone with minimal counseling met these criteria and were transferred to standard care within the first 12 weeks of the study. In contrast, only 41% of the standard counseling patients and 19% of the enhanced counseling patients met treatment failure criteria during the course of the study. Furthermore, 90–100% of standard and enhanced care patients were able to sustain 8 weeks of abstinence from opiates and cocaine during the study as compared with only 30% of clients in the minimal services condition. These findings clearly support the utility of structured drug abuse counseling services that include contingencies targeted on objective evidence of drug use versus abstinence.

The McLellan et al. (50) study discussed above also provided a contrast between standard counseling and counseling enhanced by psychiatric, family, and employment counseling resources. Although not all services were provided in the amounts intended, the groups did differ significantly in the planned direction on amounts of family therapy and psychiatric (but not employment) services received. Enhanced care patients submitted significantly fewer opiate- (but not cocaine)-positive urines during the treatment episode. Furthermore, outcome measures of drug use, psychiatric status, and social functioning generally favored the enhanced care over the standard care group in data gathered at the end of treatment.

Similar outcomes were shown in an earlier related study by this same group (75, 78), in which methadone-maintained volunteers (N = 110) were randomly assigned to receive 6 months of drug abuse counseling alone or counseling supplemented with psychotherapy (supportive–expressive or cognitive behavioral). Outcome assessments, both during therapy and after therapy ended, showed generally better performance for the psychotherapy-plus-counseling than for counseling-alone subjects, with psychotherapy subjects requesting fewer methadone dose increases and fewer psychiatric medications during treatment. It should be noted that in this study, psychotherapy patients attended twice the number of sessions as did counseling-only patients, so that extra attention rather than any specific content of the psychotherapy cannot be ruled out as a mechanism for the effects. In further analysis of the data from this study, the investigators found that differential benefits of the added psychotherapy were most apparent in patients with highest prestudy levels of psychiatric symptoms, as measured by the Addiction Severity Index (76). In contrast, patients diagnosed with antisocial personality disorder did not benefit from added psychotherapy treatment (77).

Taken together, these findings support the conclusion that treatment outcomes for a subgroup of methadone patients may be improved by providing extra therapy (independent of contingency management interventions) and suggest that it may be possible to target ancillary treatments to patients who can best benefit from these treatments. Future research needs to determine how added services and psychotherapy can best be integrated with contingency management techniques to enhance the broadest array of treatment outcomes.

A specialized form of behavior therapy that could be especially beneficial to drug abusers involves training in the social skills necessary to function effectively as a nonabuser. Programs have been developed to train drug abusers in relapse prevention skills (26, 44) and in skills necessary to seek and obtain employment (22, 23, 61). Relapse prevention skills programs are designed to bolster drug-free support systems and to teach drug abusers the skills needed to handle problematic situations that might lead to relapse. Hawkins et al. (25, 26) delivered relapse prevention skills training to therapeutic community patients before they left residential treatment. The program effectively enhanced interpersonal and problem-solving skills, as assessed in post-treatment role-play tests (26). However, there was little evidence of skills training effects on post-treatment drug use or relapse rates (25). McAuliffe and colleagues (44, 45) evaluated an aftercare treatment package called Recovery Training and Self-Help, which combined skills training with participation in self-help groups after formal treatment ended. The package was administered to opiate abusers drawn from a variety of program types (methadone, drug-free, detoxification, and residential treatment) both in the United States and China. Outcome evaluation during a 1-year follow-up found significantly more good outcomes (abstinent or using opiates less than once a month) among subjects in the experimental group than among control group subjects, with a 15–17% improvement in abstinence rates. In each group, however, there was no assessment of treatment impact on nonopiate illicit drug use, including cocaine. The diversity of study populations makes it difficult to draw conclusions about impact on any particular subgroup.

Hall et al. (22, 23) developed and evaluated a Job Seekers Workshop for drug abusers. The program was shown to effectively teach the desired skills (including preparing applications and interviewing) and to result in higher employment rates for experimental than for control subjects. However, there was no evaluation of impact on drug use.

Overall, research on skills training for heroin abusers is at an early stage. However, this approach appears promising for teaching skills needed to adopt a drug-free lifestyle (e.g., job-seeking), as well as for relapse prevention. Future research needs to consider integration and timing of effective treatment elements—including contingency management, psychotherapy, and skills training—to best achieve positive long-term outcomes.

Important recent advances have been made in learning-based interventions for outpatient treatment of cocaine abuse. Across several controlled trials, learning-based interventions have resulted in clinically significant reductions in cocaine use in dependent patients. These reports have focused almost exclusively on during-treatment abstinence and relatively short follow-up periods. These projects were initiated in the midst of the U.S. cocaine epidemic, and the pressing clinical challenge was to identify interventions that could effectively retain patients in treatment and establish even an initial period of cocaine abstinence. Longer-term outcome (1 year after treatment entry) has been assessed in recent trials (7, 27); now that an empirical base has been established and investigators are more familiar with cocaine abuse, other trials are planned.

In a series of controlled trials conducted in the same clinic by Higgins et al. (27, 28, 29), an outpatient treatment combining CRA and contingency-management procedures was demonstrated to be effective in retaining patients in treatment and establishing clinically significant periods of cocaine abstinence. CRA attempts to eliminate drug use by systematically altering naturalistic contingencies so that reinforcement density is relatively high when the subject is abstinent and low during and immediately following use. Systematic interventions are used to improve marital/family relations, vocation, social, and recreational activities. The common goal across these different treatment components is to enrich the quality of the cocaine user's life when sober and to have him or her experience a time-out from those enriched circumstances when drug use occurs. The primary contingency-management procedure used in this treatment is one in which patients earn vouchers exchangeable for retail items contingent on documentation via urinalysis testing that they have recently abstained from cocaine. The voucher system is in effect for weeks 1–12 of treatment, whereas a $1.00 state lottery ticket is awarded for each cocaine-negative urinalysis test during treatment weeks 13–24. The value of the vouchers increases with each consecutive cocaine-negative specimen delivered, and cocaine-positive specimens reset the value of vouchers back to their initial level. Those who are continuously abstinent (all cocaine-negative urine tests) could earn the equivalent of $997.50 during weeks 1–12 and $24 during weeks 13–24. This translates to an average maximum earning of $6.08 per day for those who are continuously abstinent from cocaine; in practice, the average earning has been approximately $3.50 per day.

Two of the trials examining the efficacy of this treatment compared it against standard outpatient drug counseling based on the disease model of drug dependence and the 12 steps of recovery (28, 29). The first trial was limited to 12 weeks, whereas the second trial was 24 weeks. Both treatments were delivered by experts in the respective approaches during twice-weekly sessions during weeks 1–12 of treatment and, in the randomized trial, once-weekly sessions during weeks 13–24. Patients were assigned to the two treatments as consecutive admissions in the first trial and randomly in the second trial. In both trials, the behavioral treatment retained patients significantly longer and documented significantly longer periods of continuous cocaine abstinence than did standard counseling. For example, in the randomized trial, 58% of patients assigned to the behavioral treatment completed 24 weeks versus 11% of those assigned to standard counseling. Furthermore, 68% and 42% of patients in the behavioral group achieved 8 and 16 weeks of documented, continuous cocaine abstinence versus 11% and 5% of those in the counseling group.

In a third trial conducted with this treatment, patients were randomly assigned to receive the behavioral treatment with or without the voucher program (27). Treatment was 24 weeks in duration and the voucher versus no-voucher difference was in effect during weeks 1–12 only. Both treatment groups were treated the same after week 12. Both treatment groups were followed for an additional 6 months after treatment termination so as to cover a 1-year period from the point of treatment entry. This study was the first in a series of studies planned to dismantle this multicomponent treatment to determine which components actively contribute to outcome. Vouchers significantly improved treatment retention and cocaine abstinence. Seventy-five percent of patients in the group with vouchers completed 24 weeks of treatment versus 40% in the group without vouchers, and average duration of continuous cocaine abstinence were 11.7 ± 2.0 weeks in the former versus 6.0 ± 1.5 in the latter. At the end of the 24-week treatment period, significant decreases from pretreatment scores were observed in both treatment groups on the Addiction Severity Index (ASI) family/social and alcohol scales, with no differences between the groups. Both groups also decreased on the ASI drug scale, but the magnitude of change was significantly greater in the voucher group than in the no-voucher group. Only the voucher group showed a significant improvement on the ASI psychiatric scale. Those significant pretreatment to post-treatment changes and between-group differences on the ASI remained stable through follow-ups conducted at 9 and 12 months after treatment entry.

Because these studies were conducted in a clinic located in Burlington, Vermont with almost exclusively Caucasian patients, questions could be raised about the generality of these findings to inner-city, minority cocaine abusers. A recent, well-controlled study conducted with cocaine-abusing methadone maintenance patients in a clinic located in Baltimore, Maryland extended the generality of the voucher program to inner-city, primarily minority abusers (67). During a 12-week study, patients in the experimental group (N = 19) received vouchers exchangeable for retail items contingent on cocaine-negative urinalysis tests. A matched control group (N = 18) received the vouchers according to a schedule that was yoked to the experimental group and not contingent on urinalysis results. Both groups received a standard form of outpatient drug and alcohol abuse counseling. Cocaine use was substantially reduced in the experimental group, but remained relatively unchanged in the control group. For example, 47% and 42% of patients in the experimental group achieved 4 and 8 weeks of continuous cocaine abstinence, whereas only a single control subject was abstinent for 4 weeks and none were abstinent for 8 weeks.

Overall, there are four controlled trials supporting the efficacy of the community reinforcement approach to outpatient treatment of cocaine abuse. Future studies will focus on (a) which components of this multicomponent treatment other than the voucher program actively contribute to positive outcomes, (b) longer-term outcomes, and (c) the generality of this treatment to other settings and populations. The voucher program thus far appears to have generality to a fairly broad array of patients and settings. Whether that is true for other components is unknown. Application of this approach to pregnant abusers, in whom even short-term abstinence is very important, appears to be an especially promising avenue for future research.

Another treatment demonstrated in controlled clinical trials to be efficacious as an outpatient intervention for cocaine dependence is relapse prevention. This is a cognitive-behavioral treatment that teaches patients to recognize high-risk situations for drug use, to implement alternative coping strategies when confronted with high-risk events, and to apply strategies to prevent a full-blown relapse should an episode of drug use occur (43). In a randomized trial conducted by Carroll et al. (8), 42 cocaine-dependent patients were assigned to relapse prevention treatment or interpersonal psychotherapy, which teaches strategies for improving social and interpersonal problems. Both treatments were delivered by professional therapists during weekly sessions. Outcomes achieved with relapse prevention during the 12-week trial were significantly better than those obtained with interpersonal psychotherapy: 67% of those assigned to relapse prevention versus 38% assigned to interpersonal psychotherapy completed treatment, and 57% of those in relapse prevention versus 33% in interpersonal psychotherapy achieved 3 or more weeks of continuous cocaine abstinence.

In a subsequent randomized trial conducted by the same group, relapse prevention and case management were compared in a two-by-two design in which patients also received either desipramine or placebo (9). One hundred thirty-nine patients were randomized to one of four treatment groups (relapse prevention plus desipramine, relapse prevention plus placebo, case management plus desipramine, case management plus placebo); data analyses were based on 110 patients who received at least two sessions of their respective treatments. Case management was designed to provide a nonspecific therapeutic relationship and an opportunity to monitor patients' clinical status. Both treatments were delivered in weekly therapy sessions during the 12 weeks of treatment. All treatment groups improved from pretreatment to post-treatment on measures of cocaine use and the ASI drug, alcohol, family/social, and psychiatric scales; however, there were no significant main effects for psychosocial (relapse prevention versus case management) or drug treatment (desipramine versus placebo). An interesting retrospective analysis suggested outcome differences as a function of whether patients reported using high (>4.5 g) or low (1–2.5 g) amounts of cocaine per week at pretreatment. With relapse prevention, high-use patients were retained for a significantly greater mean number of sessions than were low-use patients (8.6 versus 6.0). For clinical management, there was a nonsignificant trend in the opposite direction; that is, mean number of sessions with high-severity patients was less than with low-severity patients (6.1 versus 8.0). Comparable, nonsignificant trends in the same directions were noted when continuous cocaine abstinence in the two treatments was analyzed as a function of severity of cocaine use. In a follow-up of patients during the year after treatment, patients who received relapse prevention achieved significantly higher levels of cocaine abstinence at 12-month follow-up than did patients who received case management (7).

Overall, evidence exists supporting the efficacy of relapse prevention as an outpatient treatment for cocaine dependence, but further studies will be needed to determine its reliability in producing positive during-treatment and longer-term outcomes. That is, the initial trial by Carroll et al. (8) supports its efficacy during treatment, but follow-up data were not reported. The subsequent trial (7, 9) supports its efficacy at 1-year follow-up but not during the treatment period. Obviously, both areas are necessary to recover from cocaine dependence; and if relapse prevention is demonstrated to reliably facilitate one or both, it would be an important contribution to cocaine abuse treatment research.

In addition to CRA and relapse prevention, four other learning-based treatment approaches appear promising based on preliminary results, including active cue exposure (11), coping skills training (63), chemical aversion therapy (16), and neurobehavioral treatment (62). The first three in this list are designed to serve as adjuncts to more comprehensive treatments, whereas the fourth is a comprehensive treatment.

Active cue-exposure teaches patients to engage in coping behaviors (e.g., relaxation) when confronted with environmental stimuli that elicit a conditioned drug response (e.g., craving) or that otherwise have previously set the occasion for drug use. Coping skills training is similar to active-cue exposure, but is designed to teach specific drug refusal and social skills deemed important for accessing alternatives to drug use and for coping with events that place the patient at high risk for drug use. Chemical aversion therapy is designed to establish an aversion to cocaine use by repeatedly pairing use in the clinic with a strong chemically induced nausea. To approximate cocaine use during therapy sessions, patients use an inactive placebo substance that closely resembles cocaine in appearance. Neurobehavioral treatment emphasizes many of the elements described above in relapse prevention and coping skills training in an attempt to provide the user with the skills necessary to abstain from cocaine use and avoid relapse. The prefix neuro is included to note special attention in the treatment process to difficulties likely to arise due to putative neurobiological changes that accompany initial and sustained abstinence from cocaine following chronic use. Each of these treatments address important areas of concern or interest in the quest to develop empirically based and effective treatments for cocaine abuse. Additionally, each is currently being evaluated in one or more controlled clinical trials, and thus their efficacy should be known in the near future.

Obviously, much remains to be learned about treatment of cocaine abuse, but, considered together, the aforementioned treatments represent a promising start. That optimism is bolstered by the fact that most of the treatments that appear promising with cocaine abuse have been previously demonstrated to be efficacious in the treatment of other forms of substance abuse.

The literature on learning-based treatments for alcoholism and problem drinking has been developing over the past two to three decades. The phrase problem drinking is used here to refer to a continuum of drinking-induced problems ranging from mild disruptions in functions to severe alcohol dependence. Learning-based treatments for problem drinking include an array of empirically based and effective treatments. Some are comprehensive treatments, whereas others are designed as treatment adjuncts. No one of these treatments can be touted as the superior intervention. It is true of alcoholism treatment in general that not all patients uniformly respond positively to any particular treatment. Problem drinking is a complex condition with multiple determinants, and no one intervention should be expected to be effective with everyone. Importantly, this perspective has spurred a large-scale research initiative aimed at systematically matching patients to optimal treatments (53). Matching is a plausible concept with some empirical support, but its practical utility awaits thorough experimental evaluation. A multisite trial (Project Match) sponsored by the National Institute on Alcohol Abuse and Alcoholism is currently underway to examine the merits of this intriguing concept.

The specific treatments described below are interventions demonstrated to be efficacious in controlled clinical trials conducted with problem drinkers enrolled in treatment. Other promising primary and secondary prevention strategies aimed at reducing alcohol-related harm in those exhibiting early signs or risk factors for developing serious alcohol-related problems are not covered below because they fall outside the treatment focus of this chapter (e.g., see refs. 3 and 4).

In behavioral self-control training, patients are taught to monitor their drinking, set ingestion limits, use strategies to control their alcohol intake, reward successes in achieving goals, analyze and learn from failed efforts, and develop alternative coping skills for achieving some of the benefits previously derived from drinking. Treatment is usually conducted in small groups during approximately eight weekly 1 to 1.5-hr sessions, with periodic follow-ups. The treatment can be used to achieve outcomes of controlled drinking or total abstinence.

Behavioral self-control training is based on a position that drinking patterns, be they light, moderate, or excessive, are determined, at least in part, by one's learning history and current environmental circumstances. Hence, those with excessive or otherwise problematic drinking patterns should be able to acquire a moderate, problem-free drinking style given the requisite skills and alterations in pertinent aspects of their drinking environment. A relatively extensive research literature provides empirical support for that position, with the important qualification that a goal of controlled, asymptomatic drinking is contraindicated in severely dependent alcoholics (19, 55).

A series of experimental studies reported over a 10-year period indicated that approximately 20–70% of clinical samples could learn to drink moderately and that those effects could be sustained for up to 2 years (52, 65, 73, 74). A recent longer-term study conducted by Miller et al. (55) with 140 patients meeting criteria for alcohol abuse and/or dependence revealed more modest, but nevertheless clinically important, outcomes. Follow-up conducted 3.5–8 years post treatment with 99 of the originally treated patients revealed that 14% were asymptomatic drinkers, 23% were totally abstinent, 22% were clinically improved but still impaired drinkers, 35% were unremitted problem drinkers, and 5% were deceased. Asymptomatic drinking was most likely to be achieved by those without severe alcohol dependence or without a family history of alcoholism.

An important clinical trial in this area by Sanchez et al. (65) demonstrated that among less-dependent problem drinkers, outcome was not influenced by treatment goals of asymptomatic drinking versus abstinence. Subjects were randomly assigned to receive behavioral self-control training with treatment goals of moderate drinking or total abstinence. Approximately 75% of subjects in both treatment groups were asymptomatic drinkers or abstinent at 2-year follow-up with no significant differences in outcome resulting from the different treatment goals.

Studies comparing outcomes when behavioral self-control training was delivered in individual versus group formats or by therapist- versus client-guided use of a treatment manual consistently have demonstrated substantial clinical improvements with no significant differences related to the mode of treatment delivery (52, 54).

An important caveat is that none of the aforementioned results quantify the direct contributions of behavioral self-control training to the observed outcomes. Such an assessment can only be gleaned from clinical trials comparing behavioral self-control training versus a no-treatment or standard-treatment control group, but relatively few such studies have been reported. One published trial conducted with a sample of 60 drunk drivers demonstrated a superior outcome with behavioral self-control training compared to drunk driver education or no treatment control (6). The controlled drinking group showed a 52% reduction from pretreatment baseline in amount of drinking reported at 12-month follow-up, compared to a 28% reduction and 14% increase in drinking in the standard education and untreated control groups, respectively.

Behavioral self-control training appears to be a promising approach for curtailing heavy drinking and its related problems. More thorough quantification of the contribution of the therapy to long-term outcomes would be an issue to address in future studies.

As was noted above, CRA is a multicomponent treatment developed within an operant conceptual framework, with a basic goal of decreasing substance use by systematically altering naturalistic contingencies so that reinforcement density is relatively high when the subject is abstinent and low during and immediately following use. In CRA studies with alcoholics, an alcohol-free social club that resembled a bar in atmosphere was sometimes made available to patients for socializing, parties, and so on. Patients had to be sober to attend. CRA was initially developed and tested in alcoholics residing in a state hospital, which often includes a treatment-recalcitrant population. The treatment for alcoholics is typically 6–8 weeks in duration with periodic follow-up sessions and a treatment goal of total abstinence.

Four controlled studies have all supported the efficacy of this intervention. In the seminal study by Hunt and Azrin (32), 16 males admitted to a state hospital for alcoholism were divided into matched pairs and randomly assigned to receive CRA plus standard hospital care or the standard care alone. Following discharge from the hospital, CRA patients received a tapered schedule of counseling sessions beginning on a once-weekly basis during the first month and then a once-monthly basis over the next several months. During the 6-month follow-up period, patients who received CRA reported approximately 6- to 14-fold less time drinking, unemployed, away from their families, or institutionalized than did control patients.

CRA was subsequently refined to include monitored disulfiram therapy, some additional crisis counseling after hospital discharge, and a "buddy" system wherein individuals in the alcoholic's neighborhood were available to give assistance with practical issues such as repairing cars, and so on. In a second study by Azrin (1), CRA was compared to a standard inpatient hospital program. Twenty matched pairs of alcoholic males were randomly assigned to receive CRA or the standard hospital program. During the 6 months after discharge, the CRA group spent 3- to 28-fold less time drinking, unemployed, or away from home compared to controls. The CRA group spent no time in an institutional setting (hospital, jail), whereas controls spent 45% of their time in such settings during the 6-month follow-up. During the 2 years following discharge, the CRA group spent 90% or more time abstinent; comparable data were not reported for controls. Compared to the prior CRA treatment package, this refined treatment involved less counseling time and resulted in greater abstinence levels.

A subsequent study by Mallams et al. (41) examined the effects of adding the social club to a standard regimen of outpatient counseling for alcoholism. Forty alcoholics were randomly assigned to receive systematic encouragement to attend the social club described above or to a control group that was informed about the existence of the club but received no encouragement to attend. A great deal of effort also was made to socially integrate the experimental subjects with other club members when they did attend. Similar efforts apparently were not made with controls. Subjects were assessed at intake and 3 months later. There were no differences between the two groups in subject characteristics at intake. Mean attendance at the social club during the 3 months was 2.47 ± 2.43 for the experimental group and 0.13 ± 0.5 for the control group. The experimental group showed significant improvements from intake to 3-month follow-up on measures of quantity–frequency of drinking, behavioral impairment, and time spent in heavy-drinking situations, whereas the control group did not improve significantly on any of those measures.

The fourth study by this group (2) was designed to assess the contribution of disulfiram and other aspects of CRA to outcome. Forty-three alcoholic outpatients were randomly assigned to receive (a) traditional treatment and traditional disulfiram therapy, (b) traditional therapy plus monitored disulfiram therapy, or (c) monitored disulfiram therapy plus the other components of CRA. Outcome was best with the full treatment of CRA plus monitored disulfiram, intermediate with monitored disulfiram alone, and poorest with the traditional treatment and disulfiram. Outcomes with the full CRA treatment were comparable to those reported in earlier studies, providing a third replication. When the results were analyzed according to patients' marital status, a potentially important interaction was noted. Married patients did equally well with the full CRA treatment or monitored disulfiram; it was only the single subjects who needed the complete CRA treatment to achieve abstinence.

The outcomes achieved with CRA equal or exceed the results of any controlled treatment-outcome study in the alcoholism literature. Its efficacy with severely dependent alcoholics makes it an important addition to learning-based treatments for problem drinking. Notable limitations are that all of the published studies supporting the efficacy of this approach were conducted by a single group of investigators in a rural area of Illinois. Whether these results can be replicated by other investigators in other settings remains an important question. A replication effort is currently underway in Albuquerque, New Mexico. As was described above, CRA has been adapted as an effective outpatient treatment for cocaine dependence (28, 29).

Involving spouses and family in alcoholism treatment is supported by at least three lines of reasoning: (a) family members may engage in behaviors that initiate or reinforce drinking; (b) family members may be able to acquire skills that promote decreased drinking; and (c) family members are an important potential source of alternative reinforcement once drinking stops.

Three recent well-controlled studies by O'Farrell et al. have assessed the effects of behavioral marital therapy on treatment outcomes. One study (60), that included a 2-year follow-up (59) involved 36 couples, in which the husbands had recently begun individual alcoholism treatment and received a prescription for disulfiram. Couples were randomly assigned to a no-marital therapy control group, a behavioral couples group, or an interactional couples group. Couples in the behavioral group signed a contract regarding disulfiram compliance, and they received counseling to increase positive family activities and improve communication. Couples in the interactional group primarily shared feelings about their relationship during therapy sessions. Behavioral couples therapy produced better outcomes on marital adjustment ratings than did other therapies, but there were no significant differences in abstinence among the three groups. For example, percent of days abstinent during the 10-week treatment, as reported by subjects, was 99.4%, 82.7%, and 90.6% for the behavioral, interactional, and control therapies, respectively.

In a similar study by another research group, McCrady et al. (48, 49) randomly assigned 45 patients to one of three groups: (a) minimal spouse involvement, which consisted of individual counseling with the spouse present for support; (b) alcohol-focused spouse involvement, in which spouses learned specific therapeutic skills such as how to reinforce abstinence and decrease behaviors that might occasion drinking; or (c) alcohol behavioral marital therapy, which involved all of the above elements plus skills training on how to improve other aspects of the marriage. There were no significant differences between the three groups in overall abstinence levels either during or after treatment. However, time trend analysis revealed that while the percentage of days abstinent decreased steadily over 18 months in the minimal and alcohol-focused spouse involvement groups, this trend was reversed in the behavioral marital therapy group during the second half of the follow-up period. Those between-group differences were statistically significant, and based on graphic display they appeared to represent approximately 10–15% more abstinent days in the behavioral marital group during the second half of the follow-up period. Several measures of marital and personal adjustment also indicated better outcomes during follow-up with behavioral marital therapy.

A more recent study by the O'Farrell group (58) has added insight regarding the potential importance of post-treatment relapse prevention maintenance sessions. Fifty-nine couples, defined by the inclusion of an alcoholic husband, were randomly assigned to receive or not receive 15 maintenance sessions following completion of 5 months of weekly behavioral marital therapy that included a contract for disulfiram compliance. Abstinence improved significantly from pretreatment levels in both groups during the follow-up period, but couples who received the maintenance sessions reported significantly greater abstinence and greater use of the disulfiram contract than did those who did not receive the extra sessions. Improvements in marital adjustment outcomes during follow-up as compared with pretreatment also tended to favor the group that received extra sessions.

Two studies have specifically examined the use of behavioral procedures involving spouses to increase disulfiram compliance, with one reporting positive (2) and the other negative (36) outcomes. The studies differed along numerous dimensions, making it difficult to speculate on what might account for the different outcomes.

Considered together, the results obtained using behavioral marital therapy in alcoholism treatment give grounds for cautious optimism. Although the magnitude of improvement in abstinence outcomes has not been particularly impressive, the relatively small number of studies and some between-study discrepancies underscore the importance of further research. Studies are needed that address the apparent discrepancies concerning the efficacy of using spouses to improve disulfiram compliance. Recent findings about the importance of longer-term therapy should be pursued. Finally, it is possible that the issue of overriding importance is whether or not a spouse is involved in treatment, rather than the particular brand of treatment delivered. Studies that compare couples and individual therapy would be useful in this regard.

Problem drinkers often report that they use alcohol to cope with unpleasant or stressful events, and such events have been reported to influence relapse (42). For these and other reasons, behavior therapists have examined whether social and problem-solving skills training would improve outcomes of problem drinkers. A series of clinical trials have clearly demonstrated that skills training can be an efficacious adjunct treatment for problem drinkers. The majority of trials have examined coping skills as an adjunct to inpatient treatment, and they have focused on assertiveness and related social skills as well as on general problem-solving skills. Positive outcomes have been reported both when patients were specifically selected because they exhibited certain skill deficits (18) and when training was done with general alcoholic samples (17).

In an important, well-controlled study on this topic, Chaney et al. (10) randomly assigned 40 inpatient, male alcoholics to either (a) an eight-session skills-training group focused on drinking-related problem-solving or (b) a discussion control condition in which similar topics were discussed but no specific training was provided. During a 1-year follow-up period, the skills group as compared to the combined control groups (which did not differ from each other) reported, on average, fourfold fewer drinks taken, sixfold fewer days drunk (11 versus 64 days during the 12-month follow-up), and a ninefold reduction in duration of drinking episodes (averaging 5 versus 44 days).

Similarly positive outcomes have been reported in clinical trials in which assertiveness rather than problem-solving skills was the focus of the intervention. In one study by Oei and Jackson (57), 32 alcoholics residing in an inpatient program were selected for participation based on low scores on an assertiveness scale. Patients were matched on several relevant characteristics and assigned to one of four treatment groups: (a) social skills training; (b) cognitive restructuring; (c) a combination of social skills training and cognitive restructuring; or (d) a control group consisting of traditional supportive therapy. All of the experimental groups had better outcomes than the control group on measures of assertiveness and drinking during a 1-year follow-up, but the combined group fared best and the social skills group had poorest outcomes. During the week preceding the 12-month follow-up, for example, mean ethanol ingestion levels in the control, social skills, cognitive restructuring, and combined groups were 34, 17, 11, and 5 ounces, respectively.

In one of the few negative trials on skills training, training in problem-solving strategies was compared to covert sensitization and a discussion control in chronic alcoholics residing in a half-way house (64). There were no outcome differences across the three treatment groups at 6-, 12-, or 18-month follow-up. During treatment, subjects assigned to the skills-training group could recite the strategies they were taught, but they showed marked decrements in their ability to do so at follow-up. Whether those decrements account for the failure to discern positive effects at follow-up is unclear, but seems plausible.

One recent study in the coping-skills literature by Kadden et al. (35) has illustrated the potential importance of patient-treatment matching for achieving positive outcomes. Ninety-six male and female problem drinkers who recently completed an inpatient treatment were randomly assigned to receive coping skills or interactional therapy during aftercare. Subjects in both treatment groups improved on measures of drinking and social stability, and there were no significant differences between the treatments. The investigators then looked post hoc for interactions between treatment group and patient characteristics. Coping-skills therapy was more effective for patients higher in psychiatric severity (measured by Addiction Severity Index psychological scale) and sociopathy (measured by California Psychological Inventory Socialization Scale), whereas interactional therapy was more effective for patients with lower psychopathology. Patients who scored higher on neuropsychological impairment (derived from several scales) did better with interactional therapy. The same treatment–outcome interactions were still evident at 2-year follow-up (13). Of course, retrospective analyses of this type must be treated cautiously pending replication in prospective trials, but they underscore the importance of anticipating that particular subgroups of problem drinkers may respond differently to skills training as well as any other treatment for problem drinking.

Overall, this literature is quite positive in demonstrating that skills training can produce significant and enduring improvements in outcome. More needs to be learned regarding the efficacy of different types of skills training, whether outcomes could be further improved through greater systematic matching of the interventions to particular skill deficits, and specification of any boundary conditions on the type of patient who benefits from skills training.

Significant advances have recently been made with regard to the behavioral treatment of heroin, cocaine, and alcohol abusers. The beneficial effects of including contingency-based drug abuse counseling in methadone treatment have been demonstrated. In both methadone and outpatient cocaine treatment, contingent reinforcement interventions have been shown effective for promoting sustained periods of abstinence during treatment. Some of the most impressive findings have come from an outpatient treatment for cocaine abusers that involves abstinence reinforcement (vouchers that can be traded for retail items in the community) offered in the context of an aggressive behavioral treatment program that attempts to enhance sources of nondrug reinforcement in the environment (CRA). In general, however, influencing environmental factors that may hold the key to abstinence versus relapse continues to be a challenge to behavior therapists. Learning-based relapse-prevention and skills-training therapies have shown some efficacy for teaching drug abusers skills that can help to insulate them from the insidious influence of social and other environmental stimulus factors that promote drug use. Also, involvement of significant others during treatment (BMT for alcoholism, CRA for cocaine) appears to be a promising way to exert a positive influence on the immediate environment of the abuser. More research is needed in all these promising areas to systematically assess the dose and elements of behavioral treatment that are effective, to compare potency (effect sizes) of different types of treatment when delivered to similar populations, and to examine hypotheses related to patient–treatment matching.

published 2000