By Timothy J. O??TFarrell, John P. Allen, and Raye Z. Litten


Alcoholism is a major public health problem in the United States (National Institute on Alcohol Abuse and Alcoholism 1990) and throughout much of the world (World Health Organization 1977). It is the most prevalent psychiatric disorder in the United States (Helzer et al. 1991, pp. 81-115). Development of effective treatments for alcoholism, including the use of medications that reduce drinking, has been an important public health goal for many years (National Academy of Sciences 1990).

Disulfiram (Antabuse) is a medication that inhibits metabolism of acetaldehyde, a toxic breakdown product of alcohol, and produces unpleasant symptoms (flushing, headache, nausea, vomiting, dizziness, light-headedness, tachycardia) if a person consumes alcohol. The patient who stops taking disulfiram can experience the disulfiramalcohol reaction for up to 14 days (and generally for at least 4 to 5 days) after discontinuing ingestion of disulfiram. The rationale for disulfiram in treating alcoholism is that most alcoholics taking disulfiram will not drink for fear of getting sick. Disulfiram thus prevents impulsive drinking in response to acute craving or stressors.

Of course disulfiram as a deterrent to drinking among alcoholics, like all medications, is only effective for as long as the patient complies with taking the disulfiram. Therefore, this chapter will review briefly the use of disulfiram in treating alcoholism with a special emphasis on the use of behavioral contracts between the alcoholic and a concerned significant other to maintain compliance with disulfiram.

Classic Study of Disulfiram by Fuller and Colleagues 1986

In their classic large-scale clinical trial, Fuller and colleagues (1986) randomly assigned male alcoholics to one of three treatment conditions: (1) 250 mg of disulfiram (N = 202); (2) 1 mg of disulfiram (N = 204), a control for the threat of the disulfiram-alcohol reaction; or (3) no disulfiram (N = 199), a control for the counseling that all participants received. All alcoholics in this study were scheduled to receive weekly outpatient counseling for 6 months followed by biweekly counseling  sessions for the final 6 months. Followup interviews were scheduled to occur bimonthly during the study year.

Study participants on average were middle aged (42 years mean age), ethnically diverse (54 percent white), moderately socially stable (54 percent married, 70 percent employed) chronic alcoholics (12 years of alcohol abuse). The Fuller and colleagues (1986) study produced four major results. First, a customary therapeutic dose (250 mg) of disulfiram was not superior to placebo (1 mg) or no disulfiram in producing abstinence, reducing time to first drink, or improving social or employment status among male alcoholic patients.

Second, failure of patients to take the drug as prescribed may have rendered disulfiram ineffective. In fact, only 20 percent of the subjects were judged ??ogood compliers??? as evidenced by consistent urinalysis results positive for riboflavin, the biochemical marker of disulfiram employed. Third, better compliance was related to more abstinence. A higher percentage of compliant than noncompliant patients was continuously abstinent for the 1 -year followup period (43 percent for the compliant versus 8 percent for the noncompliant, p < .001). Fourth, patient acceptance of disulfiram was low.

Only 38 percent of the 1,618 alcoholics who met study criteria and were medically cleared to take disulfiram agreed to enter the study and possibly take disulfiram. Most study refusals were due to patients??T reluctance to take disulfiram. The Fuller and colleagues (1986) study showed that disulfiram was not effective because of serious problems with patient acceptance and compliance. However, abstinence was observed among patients who took the medication consistently. Further, the low patient acceptance may have been related to the fact that the disulfiram was not an integral part of the alcoholism counseling used.

Findings such as these underscore the importance of techniques to increase compliance with disulfiram. After an overview of such techniques, the use of behavioral contracts to increase compliance and to make disulfiram an integral part of psychosocial treatments for alcoholism will be considered in detail.

Overview of Techniques to Increase Compliance With Disulfiram

Varied strategies to enhance disulfiram compliance have been devised. This overview will examine implants, patient instructional sets, and incentives, each of which are covered in depth elsewhere (Allen and Litten 1992).  Disulfiram Implants. The most potent guarantee for disulfiram compliance would seem to be physical implant to release disulfiram into the bloodstream at a consistent rate and at a level sufficient to cause an adverse physical reaction should the patient drink.

Disulfiram implants have been available for 30 years, and more recent investigations on implants have been reasonably well controlled (e.g., Johnsen et al. 1990; Wilson et al. 1984). Such studies have identified serious limitations, however, to disulfiram implants. First, most such depots fail to release adequate levels of disulfiram. Second, the bolus of the implant under the skin has led to adverse effects of infections and rejection after the implant surgical procedure. Third, controlled studies have not found superior outcomes for alcoholics treated with currently available implants.

These results led Allen and Litten (1992) to conclude that disulfiram implants, for biochemical reasons per se, have been largely ineffective. Patient Instructions. An alternative strategy to enhance attractiveness of taking disulfiram is modification of patient instructions and expectations for the medication. Two different approaches have shown promising results in initial investigations.

Duckert and Johnsen (1987) allowed patients a choice of methods for using disulfiram ranging from the conventional long-term use to prevent drinking to infrequent, periodic use for specific reasons chosen by the patient (e.g., to prevent drinking in a high-risk situation). Kofoed (1987) significantly increased disulfiram compliance among alcoholic outpatients by informing them and their case managers of results of carbon disulfide breathalyzer tests taken at each counseling session to corroborate extent of recent disulfiram use.

Tangible Incentives.

Other studies have considered the effectiveness of tangible incentives for taking disulfiram. Most of the incentives that have been tried relate in some way to circumstances and conditions specific to the patient group of interest. Among these have been (1) less restrictive probation for individuals charged with alcohol-related offenses (Boume et al. 1966; Brewer and Smith 1983); (2) methadone contingency for methadone patients with alcohol problems (Liebson et al. 1973, 1978); (3) money returned from a security deposit made by alcoholics initiating outpatient counseling (Bigelow et al. 1976); (4) job security for industrial workers referred by their employers for drinking-related job problems (Robichaud et al. 1979); (5) continued affiliation with the treatment program in which the patient is currently enrolled rather than discharge from the program or referral to another clinic (Bickel et al. 1989; Sereny et al. 1986); and (6) more frequent clinic visits as a form of psychological  incentive among socially isolated, inner-city alcoholics (Gerrein et al. 1973).

While each of these studies suffers from specific methodological limitations, the findings uniformly suggest better disulfiram compliance and more favorable clinical outcomes for those who received an incentive for taking disulfiram. In such studies, direct observation of the patient taking disulfiram by a court or clinic staff person, often referred to as ??osupervised??? disulfiram, was typically part of the procedure to assess compliance so that the incentive could be provided.

Observed or supervised disulfiram without specified incentives is the basis for disulfiram contracts and other compliance enhancement procedures-the subject of the remainder of this chapter.


Description and History of Disulfiram Contracts

??oBehavioral contracting??? is generally done with both the client and a significant other, usually the spouse, in the client??Ts living environment. Interestingly, behavioral agreements typically do not explicitly stipulate tangible consequences for taking or refusing disulfiram, although they do specify social reinforcers, e.g., expression of appreciation by the spouse when the alcoholic takes disulfiram.

These agreements also require that both the patient and the significant other formally and publicly commit themselves to observation of disulfiram use. Figure 1 provides a sample disulfiram contract taken from O??TFarrell and Bayog (1986), who describe the clinical procedures involved in some detail including methods for dealing with common resistances and problems encountered.

Two slightly different versions of the disulfiram contract appeared in the literature at about the same time. The Behavioral Marital Therapy (BMT) version, first described by Miller and Hersen (1975) and Miller (1976) provides for observed disulfiram with mutual thanking by alcoholic and spouse plus a commitment to refrain from discussions (except during BMT sessions) about the alcoholic??Ts drinking (see item 3 of the contract in figure 1). 


FIGURE 1. Sample disulfiram contract used in behavioral marital therapy with alcoholics. [See source article.]

Like all disulfiram contracts, the BMT version seeks to maintain disulfiram ingestion and abstinence. The BMT version also seeks to restructure the couple??Ts relationship to reduce their conflicts about past drinking or the likelihood of future drinking and to decrease the spouse??Ts anxiety, distrust, and need to control the alcoholic.

The BMT version tries to deal with these presumed relationship dynamics of the early  sobriety period in order to increase support for abstinence and reduce the risk of relapse (O??TFarrell 19933). The Community Reinforcement Approach (CRA) version derives from Azrin??Ts (I 976) attempt to augment the effectiveness of his CRA approach to alcoholism treatment (Hunt and Azrin 1973) by adding a disulfiram component to it.

The CRA version of the disulfiram contract is identical to the BMT version except that CRA does not include item 3 restricting discussions about drinking. Studies of each of these two approaches are considered next.

Behavioral Marital Therapy Studies of Disulfiram

Miller and Hersen (1975) reported a case in which a disulfiram contract and BMT were used to promote abstinence and reduce marital conflict. The 49-year-old factory worker husband, whose 10-year history of alcoholism was characterized by many arrests, car accidents, and marital problems, had been consuming a pint to a fifth of vodka daily prior to his admission to a hospital for alcoholism treatment.

The wife had decided to divorce him if he did not stop drinking. Pretreatment assessment revealed very little positive communication and extensive negative comments by the wife about drinking (e.g., blaming for past drinking, threats about future drinking). Treatment consisted of BMT sessions to increase constructive communication and a contract specifying daily disulfiram intake by the husband and cessation of discussion about drinking by the wife. (The contract was quite similar to the model in figure 1.)

Treatment started with weekly sessions in the hospital with the husband visiting home each weekend and continued on an outpatient basis biweekly for 3 months and monthly thereafter for 3 months. Ninemonth followup revealed that the husband had remained abstinent and was still taking disulfiram daily. In addition, the couple was communicating more constructively and going out together regularly. The wife had stopped mentioning the past and was generally more pleasant.

The couple had handled several problems quite well. The positive BMT results from the Miller and Hersen (1975) case report, along with similar affirming results from other early case reports and uncontrolled studies of BMT, led to controlled studies of BMT.

The Counseling for Alcoholics??T Marriages (CALM) Project studies of BMT included disulfiram contracts as part of the Project CALM BMT program (O??TFarrell 1993a, pp. 170-209).  In an initial Project CALM study (known as the CALM-l study), O??TFarrell and colleagues (O??TFarrell et al. 1985, 1992) investigated the effect of adding BMT couples group treatment with a disulfiram contract to individually oriented outpatient treatment of married male alcohol abusers.

Thirty-six couples, in which the husband had recently begun individual alcohol counseling that included a disulfiram prescription, were randomly assigned to (1) 10 weekly sessions of a BMT (behavioral rehearsal of communication skills and marital agreements) couples group plus a disulfiram contract; (2) 10 weeks of an interactional (largely verbal interaction and sharing of feelings) couples group without a disulfiram contract; or (3) a no-marital-treatment control group.

Results at the end of treatment (O??TFarrell et al. 1985) showed that adding BMT plus a disulfiram contract to individual alcoholism counseling produced significant improvements in marital and drinking adjustment that were superior to outcomes of individual counseling alone and to individual counseling plus interactional couples therapy.

Results during the 2 years after treatment (O??TFarrell et al. 1992) showed that alcoholics and their wives who received the additional BMT remained significantly improved on marital and drinking adjustment throughout the 2 years. Although BMT continued to appear superior to individual counseling alone on marital adjustment throughout much of the 2-year followup, the strength and the consistency of findings favoring BMT diminished as time after treatment lengthened.

In terms of drinking outcomes during the 2 years after treatment, the addition of BMT no longer produced better results than did interactional couples therapy or individual treatment alone. The specific contribution of the disulfiram contract to the results observed in the CALM-l study cannot be determined. The disulfiram contract was part of the BMT program. The extent of patients??T use of disulfiram was not measured.

Still, it seems likely that during treatment the disulfiram contract may have contributed importantly to the lower rate of drinking and drinking-related problems observed in the BMT couples as compared with the other couples who did not use the disulfiram contract. Perhaps after treatment ended, use of the disulfiram contract decreased since BMT no longer produced less drinking than the other treatments.

The CALM-l study results suggested the need for a study of treatment methods to maintain the use of the disulfiram contract and the gains produced by BMT, especially for drinking and related behaviors. Results of CALM-l produced CALM-2, a study to evaluate the usefulness of couples relapse prevention (RP) sessions for maintaining changes in marital and drinking adjustment produced by short-term BMT.  Continued use of the disulfiram contract, especially for individuals suffering more severe drinking problems, was one of the goals of the RP sessions.

In this study, after participating weekly for 5 months in a BMT couples program, 59 couples with an alcohol-abusing husband were assigned randomly to receive or not receive 15 additional couples RP sessions over the next 12 months. Outcome measures were collected before and after BMT and at quarterly intervals for the 2-1/2 years after BMT. The CALM-2 investigation produced three major findings (Cutter et al. 1993; O??TFarrell et al. 1993).

First, results for the entire sample showed the additional RP sessions produced better outcomes during and for the 6 to 12 months after the end of RP. Specifically, alcohol abusers who received RP after BMT had more days abstinent and used the disulfiram contract more than those who received BMT alone. The superior RP drinking outcomes continued through 18 months followup (i.e., 6 months after the end of RP).

Couples who received the additional RP also maintained improved marriages longer (through 24 months followup) than did their counterparts who received BMT only (through 12 months followup). Second, for alcoholics with more severe marital and drinking problems, RP produced better marital and drinking outcomes throughout the 30-month followup period. Specifically, alcoholics with more severe alcohol problems at study entry used the disulfiram contract more (see figure 2) and showed a less steep decline in use of the disulfiram contract (see figure 3) throughout the 30 months after BMT if they received the additional RP than if they did not.

Further, alcoholics with more severe marital problems at study entry experienced better marital adjustment and more days abstinent and maintained relatively stable levels of abstinence if they received the additional RP, while their counterparts who did not receive RP had poorer marital adjustment and fewer abstinent days and showed a steep decline in abstinent days in the 30 months after BMT.

Third, greater use of the disulfiram contract was associated with more days abstinent and more positive marital adjustment test scores after BMT for all subjects irrespective of the amount of aftercare received.

To summarize, two Project CALM BMT studies of disulfiram contracts have been completed. The CALM-l study showed that adding BMT plus a disulfiram contract to individual alcoholism counseling led to better short-term drinking and marital outcomes than a disulfiram prescription alone accompanied by either an alternative form of couples counseling or   individual counseling alone.

FIGURE 2. Two-way alcohol problem severity by treatment interaction for use of disulfiram (Antabuse) contract in Cutter and colleagues (1993) CALM-2 study of BMT and disulfiram contract.

Superior BMT drinking results did not persist through the 2-year followup period, possibly because many couples discontinued their disulfiram contract after treatment ended. The CALM-2 study indicated that adding couples RP sessions in the year after BMT enhanced use of the disulfiram contract and yielded better marital and drinking outcomes than BMT alone. These better RP outcomes persisted for 18 to 24 months after BMT for the entire sample and throughout the entire 30 months followup after BMT for those with more severe marital and drinking problems. Thus the Project CALM studies suggest that disulfiram contracts used with BMT are associated with less drinking and greater disulfiram compliance. However, the specific contribution of disulfiram contracts to the multifaceted BMT treatment package remains to be investigated. 

FIGURE 3. Three-way alcohol problem severity by treatment by time interaction for use of disulfram (Antabuse) contract in Cutter and colleagues (1993) CALM-2 study of BMT and disulfiram contract.

Community Reinforcement Approach Studies of Disulfiram Contracts

Azrin developed a CRA to treating alcoholics that was based on operant conditioning principles. CRA rearranged community reinforcers such as the job, the family, and the social relations of the alcoholic such that drinking produced a ??otime-out??? from a high density of reinforcement.

The results from the initial CRA study convincingly demonstrated that the eight alcoholics who received this CRA counseling drank less, worked more, and spent more time with their families and out of institutions than did a matched control group of eight alcoholics who did not receive these procedures (Hunt and Azrin 1973). A

zrin (1976) attempted to augment the effectiveness of CRA by adding a disulfiram component to it. Subjects were 18 inpatient alcoholic males randomly assigned to CRA with a disulfiram contract or to a matched control group receiving standard alcoholism treatment with general advice to take disulfiram. Several procedures, all of which have become standard in both CRA and BMT studies of the disulfiram contract, were employed with the experimental group to heighten disulfiram compliance.

These included (1) instructing the client and significant other on the rationale and benefits of disulfiram as a ??ochemical time-delay device??? to avoid impulsive drinking and its consequences, (2) encouraging the alcoholic to personally request that the other person monitor the use of disulfiram, (3) establishing specific links between taking the medication and recurrent daily activities, (4) monitoring disulfiram administration by the significant other or the counselor, and (5) referring the client to a physician supportive of disulfiram.

A formal contract in which the client agreed to take disulfiram (similar to the one in figure 1 except that item 3 was omitted) was signed with the counselor. In the Azrin (1976) study, 6-month self-report followup measures demonstrated that the CRA with the disulfiram contract was substantially more effective in reducing the number of drinking days than standard treatment with general advice to take disulfiram. Additional followup for 2 years of CRA subjects (followup for the control-group subjects was limited to 6 months) showed continued positive outcomes for CRA subjects on number of days drinking, percent time employed, percent time institutionalized posttreatment, and amount of time spent with the family.

The authors also concluded that the CRA with the disulfiram contract was preferable to the previous version of CRA without disulfiram in that the newer strategy reduced drinking and the amount of  CRA counseling time needed. While these results suggest benefits of contracting for disulfiram and of employing a range of enhancement techniques, unfortunately it is not possible to isolate the effects of the disulfiram contract from the remaining elements of CRA itself. A subsequent CRA study (Azrin et al. 1982) more explicitly evaluated the benefits of disulfiram contracting and compliance aids.

Outpatients in a rural community alcoholism clinic who did not suffer co-morbid drug dependence or psychosis were encouraged to take disulfiram and were referred to the agency physician and a nearby pharmacy to obtain the medication.

Following the first session, 43 subjects were randomly assigned to one of three treatment conditions: 1. Traditional treatment plus a prescription for disulfiram without special disulfiram contract procedures; 2. Traditional treatment with disulfiram contract procedures similar to those employed in the Azrin (1976) study cited above; or 3. CRA including disulfiram contracting.

Six-month followup in the Azrin and colleagues (1982) study demonstrated that patients in the three conditions differed on number of days on which disulfiram was taken, days drinking, days intoxicated, and average amount of ethanol consumed per drinking episode.

There were two major findings: First, patients receiving CRA and a disulfiram contract performed best; those in traditional therapy without disulfiram contract fared worst; those in traditional therapy with a disulfiram contract responded at a level intermediate between the other two groups.

Figure 4 suggests that, while the groups differed throughout the followup period, disulfiram use declined appreciably by the second month for the traditional therapy group without the disulfiram contract and decreased quite rapidly thereafter with no disulfiram being taken after 3 months.

The clients in the two groups given the disulfiram contract were taking disulfiram about 90 percent of the time initially and showed less of a decrease over time remaining with two-thirds or more days taking disulfiram on average through 6 months followup. Second, the authors found that married or cohabiting clients assigned to the disulfiram contract and traditional treatment performed about as well on the four outcome measures as they did with CRA plus the disulfiram contract. 

FIGURE 4. Mean number of days on which disulfiram (Antabuse) was taken during each month (30 days) of the 6 months of followup. Disulfiram was given in the usual manner in the ??otraditional??? group whereas adherence was socially motivated for the ??oDisulfiram Assurance??? group. The ??obehavior therapy??? group received community-oriented reinforcement therapy in addition to the disulfiram assurance program. (Reprinted from Azrin et al. 1982, p. 109, by permission of Pergamon Press.)

NOTE: Reprinted from Journal of Behavioral Therapy and Experimental Psychiatry, Vol. 13, No. 2, Azrin, N.H.; Sisson, R.W.; Meyers, R.; and Godley, M. Alcoholism treatment by disulfiram and community reinforcement therapy, pp. 105-112 (1982), with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, 0X5 1GB, UK.

Single clients, however, achieved additional gain from CRA plus disulfiram contract over traditional therapy and disulfiram contract. Table 1 from the Azrin and colleagues (1982) study illustrates this finding. 

TABLE 1. Mean number of days abstinent during the 6th month (30 days) of followup (N = 43). Singles Couples Traditional counseling with disulfiram prescription 6.75 17.40 Traditional counseling with disulfiram contract 8.00 30.00 CRA with disulfiram contract 28.30 30.00 Reprinted from Azrin et al. 1982, p. 110. Adapted by permission from Pergamon Press.

NOTE: Reprinted from Journal of Behavioral Therapy and Experimental Psychiatry, Vol. 13, No. 2, Azrin, N.H.; Sisson, R.W.; Meyers, R.; and Godley, M. Alcoholism treatment by disulfiram and community reinforcement therapy, Volume 13, pp. 105-112 (1982), with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington 0X5 1GB, UK.