The demonstration that animals would self-administer many drugs of abuse led to a major reformulation of the conceptual framework of the problem of drug addiction. Today, it is well accepted that drugs of abuse serve as positive reinforcers for the behavior leading to their acquisition and self-administration. Physical dependence is not a necessary condition for laboratory animals to voluntarily self-administer most of the same drugs that humans abuse. Further, it is also recognized that, for drugs that produce physical dependence, drug self-administration may also be maintained by the reduction of the aversive withdrawal state, i.e., by negative reinforcement. The conceptualization of drugs of abuse within the framework of the principles of behavior analysis pointed to antecedent variables and modulating events that had been shown to be of importance with behaviors maintained by more commonly used reinforcers such as food, water or sex, and therefore might be of importance in controlling drug-seeking behaviors. Studies of such antecedent variables as deprivation, satiation as well as schedules of reinforcement have shown that drug reinforcers follow the same general rules as other positive reinforcers. More recently, human laboratory research has confirmed and extended the relevance of a behavior analytic approach to understanding the etiology, prevention and treatment of substance abuse. The clinical studies represented at this symposia on contingency management (CM) represent an outgrowth of this research and an example of how basic research can lead to important clinical applications. CM, which is defined as the systematic reinforcement of therapeutically desired behaviors and/or the withholding of reinforcement or punishment of undesired behaviors has been shown in numerous clinical trials to be an effective strategy in the treatment of alcohol and other drug use disorders. Recent developments in contingency management treatment, including research focus on feasibility, effectiveness and cost-effectiveness as well as applicability to special sub-populations, make a symposium on the topic especially timely.

Dr. Stitzer discussed the history of contingency management in drug abuse treatment. Dr. Higgins addressed contingent incentives in the treatment of cocaine and other substance abuse disorders, and Dr. Silverman discussed a therapeutic workplace for the treatment of heroin and cocaine addiction.

Symposium Chair: Susan M. Stine, MD, PhD, Wayne State University School of Medicine, Detroit, MI

Discussant: Charles R. Schuster, PhD, Wayne State University School of Medicine, Detroit, MI

Drugs are a source of potent and immediate biological reinforcement. As such, they act somewhat like a magnet, drawing in drug-experienced individuals back to the behavior of drug seeking and drug self-administration despite current intentions to abstain or stop the use of drugs.

Clinically, this plays out in a process that is termed ambivalence. Drug abusers walk the fence. They are being pulled in two directions. In the one direction, the potent and immediate biological reinforcement offered by drugs is always drawing them back and is underlying the tendency for this to be a chronic, relapsing disorder. On the other side, there are reasons to renounce drug use and to try to abstain.

One of the challenges faced in designing treatments for drug abuse is to enhance the motivation for abstinence and to try to change the balance between the draw back to drug use because of those reinforcing effects versus the positive effects that could be gained from abstinence.

Several approaches have been taken historically to this issue of enhancing motivation to abstain from drugs. One of the earliest approaches, which derives from the disease model of drug abuse and is adhered to by some of the 12-step philosophy groups, is that in fact there is very little that anyone can do to address motivation to change. This must come to the addicted individual by experiencing the full impact of the natural consequences of drug use. Nothing is going to change until that whole process has unfolded.

Within the treatment field, however, there have been some attempts to take the kernel of this idea and try to enhance motivation through techniques of verbal persuasion that provide feedback to the addicted individual about these negative problems and consequences of drug use. Within this approach, two specific approaches have been tried. One of the traditional ways of approaching this idea of feedback is a confrontational method. This has been traditionally used in the drug abuse field, where the negative consequences and drug-related problems are fed back in a sort of blunt and reality-based way in an effort to get the addicted individual to see the light and acknowledge that renouncing drug use would be the only real solution to these problems. Having the same goal, but coming at it in a different way, is motivational enhancement therapy, which takes a much gentler approach to feeding back the drug-related problems and trying to get individuals to understand the link between their problems and their drug use and to come to their own conclusion with the insight that renouncing drug use would be a good idea.

A third technique is derived from principles of behavioral psychology. In this technique, positive reinforcement is used to try to increase motivation for change. The generic term for this is contingency management, and the rationale behind it is to try to make abstinence a more attractive option by introducing positive reinforcement into the environment of the drug user and to try to counteract the allure of that potent intermediate drug reinforcement.

This field was started back in the 1940s and 1950s by B.F. Skinner. Skinner was the father of operant psychology or behavioral psychology. He was working in a laboratory setting with rats and pigeons and showed that rats would learn to press a lever if the consequence was the delivery of food. From Skinner’s early work, a great deal of science and research has gone into a fuller understanding of the principles of behavior, but the fundamental concept is that behavior interacts with the environment in a very important way to produce consequences, and the consequences of behavior feed back to change the frequency of behaviors. In a nutshell, the basic principles are that behavior will increase if the consequence is a positively rewarding one, such as the case with the biological reinforcement produced by drugs.

Behavior will also increase if the consequence is avoidance of some negative event. With drug use, drug seeking and drug taking is instrumental in avoiding or escaping from withdrawal. On the other side of the coin, there are some interventions or methods that can be used to decrease behavior — punishment or the direct delivery of an aversive event as a consequence of the behavior will suppress or reduce it. Also, behavior will decrease if rewards are consistently delivered for the absence of a particular behavior.

A tremendous amount of work has been done regarding the science of behavior and delineating the principles of behavior. Reinforcers and punishers are not defined by their inherent properties, but need to be empirically and individually determined. Even something like money, which is a universal reinforcer for most human beings, does not work in all instances. Someone who is wealthy might not be influenced by the offer of a monetary incentive.

A great deal of work has gone into understanding how patterns and rates of behavior are influenced by the schedules on which reinforcement is delivered. Finally, it is important to understand some of the features that are needed in order to identify an effective reinforcer, such as the saliency, the desirability to the individual, the fact that the reinforcer must be able to be repeatedly administered, and that immediate reinforcement is better than delayed.

Token economy is a concept that grew out of the basic principles of behavior and is applicable to shaping positive behaviors in many different environments.

Two of the basic things that are needed to apply contingency management are a measurable target behavior and a rewarding consequence. It is important to link the behavior very tightly to the reward because giving things away for free does not change behavior.

Vouchers can have positive therapeutic benefits during and after treatment. There is an assertion in the field that vouchers only work while they are in place. This is not always true. One population with whom research suggests that it's not true is treatment seekers — people who are asking for help to change their drug-using behavior. Among treatment seekers, incentives can actually show positive effects beyond the time that they are in place.

Dr. Higgins asks cocaine-dependent individuals to stay in treatment for 1 year in an outpatient setting. In the initial 12 weeks, which is when the vouchers take place, they are coming for twice-weekly Community Reinforcement Approach (CRA) counseling and then for three times-weekly urinalysis testing. In the second half of the treatment, the vouchers are no longer in place. At this point, counseling is done once weekly, and urinalysis is done twice weekly. In the second 6 months, he recommends aftercare rather than abrupt discontinuation from the clinic.

The voucher program is a 12-week intervention. When it is used in cocaine-dependent individuals, in weeks 1 through 12, it is built around a Monday, Wednesday and Friday urine toxicology-monitoring schedule. That schedule should allow little opportunity for undetected drug use. It is very important in contingency management interventions that you have an objective measure of the target behavior. The value of the vouchers increases with each consecutive negative specimen. Positive specimens or failure to provide a specimen resets values back to the initial level. All purchases are made by staff, so there is never any exchange of cash, because, for many cocaine users, cash is actually a cue for cocaine use. The total possible earnings in a 12-week intervention with these cocaine-dependent outpatients are almost $1,000.

CRA is a fairly intensive behavior therapy intervention that emphasizes lifestyle changes that are incompatible with drug abuse. CRA addresses a wide spectrum of needs that cocaine abusers present with and are efficacious aside from the vouchers.

CRA plus vouchers was compared to drug abuse counseling in two trials. In both trials, CRA plus vouchers had very striking effects on retention. Researchers were also able to document a high level of cocaine abstinence while patients were in treatment and after treatment as well.

Researchers then wondered whether it was the CRA or the vouchers that were responsible for the outcomes. Two trials showed that the vouchers were an active contributor to the positive outcomes seen with the treatment. They were increasing retention and cocaine abstinence. Indeed, their effects on retention were so robust that they were obscuring interpretation of the mechanism responsible for their effects on abstinence. What was unclear was whether vouchers were increasing abstinence by increasing retention and thus the amount of counseling services received or through a direct reinforcement effect of vouchers on abstinence.

To determine how vouchers produced initial and longer-term cocaine abstinence, 70 cocaine-dependent adults were randomized to two treatments. All patients received both CRA and vouchers. However, in one group, the vouchers were contingent on the urinalysis results indicating that they had abstained recently from cocaine use. The other group received the same amount of vouchers, independent of urinalysis results. Retention rates were the same across the two treatment groups, and yet cocaine abstinence levels differed during treatment and for 15 months after completion of the voucher intervention. Those results provided strong evidence that vouchers increased cocaine abstinence through a direct reinforcement effect and not by increasing the amount of CRA counseling services received.

The single best predictor of long-term abstinence in that study was achieving a period of sustained abstinence during treatment. As the amount of abstinence achieved during treatment increased, the probability of being abstinent during follow-up increased. This same relationship also holds for the control condition: during-treatment abstinence predicts longer-term abstinence, and it does so with about the same likelihood as in the contingent-vouchers condition. The important difference between the contingent-vouchers and control conditions that results in better longer-term outcomes in the former is that more patients achieve an initial period of sustained abstinence with contingent vouchers than control treatments.

The key to voucher effects on longer-term abstinence appears to be the promotion of a period of sustained abstinence early in the quit attempt.

Contingent vouchers are also efficacious in promoting smoking cessation during and following treatment in pregnant and recently postpartum women.

The heart of the Therapeutic Workplace is a Web site for Hopkins Data Services, which is a company that provides data entry services to researchers at Johns Hopkins and the surrounding community. The Web site looks like any other business, but the business exists not to make money but to employ adults who have long histories of poverty, chronic unemployment and drug addiction. While providing employment is clearly a good thing, there is little reason to think that employment in and of itself would affect drug use. So, to affect drug use, Hopkins Data Services uses a fairly simple contingency: every Monday, Wednesday and Friday, when data operators report to work, they have to provide a urine sample under observation. If the urine sample is negative for drugs, they can work that day to earn a salary. If it’s positive, they have to leave the workplace and go home for the day. A critical part of the contingency is that the employees are never fired for drug use but are always encouraged to return the next day or any day thereafter to try again. This intervention has been evaluated, and there is some evidence that it can be effective in promoting and maintaining abstinence and in increasing employment.

This work was conducted in Baltimore, which has been disproportionately affected by high rates of heroin and cocaine addiction. Methadone is probably one of the most effective treatments for heroin addiction, and it is used widely in Baltimore. However, many methadone patients do continue with the use of other drugs. One of the most effective interventions to promote abstinence from these drugs has been abstinence reinforcement contingencies.

Earlier studies have shown that interventions can be most effective when high magnitudes were employed. Two studies showed that abstinence can be promoted in highly treatment-resistant patients who failed to respond to an initial intervention in which they could earn about $1,000 in vouchers over a 12-week period. The study found that tripling the magnitude of reinforcement to about $3,400 over a 9-week period promoted sustained abstinence in about half of the patients who failed the first intervention. Additionally, in methadone patients, many patients relapse to drug use when the voucher intervention is discontinued.

Dr. Silverman conducted a study showing that abstinence could be maintained over extended periods of time in this population by sustaining the abstinence reinforcement contingencies over the period of 1 year. However, financing these interventions can be a problem, so the Therapeutic Workplace intervention was developed to address this problem.

People are hired into this model program to come to work 5 days a week. Every day when they report to work, they are required to provide a urine sample under observation. If it is negative for drugs, they can then work and earn a salary, but only as long as they remain abstinent. Treatment and research has been focused on poor, chronically unemployed adults who have few, if any, job skills. For those people, a two-phased treatment has been created. In the initial phase, their job is to participate in an intensive job skills training program designed to teach them the skills that they will need to perform the job. They earn vouchers instead of cash to reduce the chance that they will use their earnings to purchase drugs early in treatment. As soon as they acquire the necessary skills and become abstinent, they move on to the second phase in which they are hired into an income-producing therapeutic workplace business. They perform real work, and they receive regular paychecks every 2 weeks. This is potentially an unlimited duration phase of the intervention. The hope of this intervention is that the therapeutic workplace business could be financially successful. If it can be financially successful, then this phase of the intervention could be sustained for extended periods of time at relatively little cost. This whole intervention could be built around different kinds of jobs, such as construction or other service jobs.

One study evaluated this treatment in pregnant drug-dependent individuals who were on methadone treatment. The study included patients who were in methadone treatment, but who continued to use heroin and cocaine despite the exposure to that treatment. It is an intensive treatment of 7 days of inpatient treatment and full-day outpatient treatment for extended periods of time.

These patients were randomly assigned to a therapeutic workplace group or to a usual care control group, and there were 20 per group. They were enrolled initially for 6 months and then repeatedly enrolled every 6 months for more than 4 years. In phase 1 of the treatment, they were invited to come to work 3 hours per day 5 days per week. Urine samples were collected every day in that phase. Their job was to participate in training to get the skills that they need. They could earn a base pay for completing work shifts and bonus pay for learning, productivity and professional demeanor. Over a 6-month period, they could earn about $4,100 in vouchers. When they went on to the business, they worked 6 hours per day 5 days per week. Urine samples were collected on Monday, Wednesday and Friday and were tested for opiates and cocaine. Their job was to enter printed data into custom data entry software. They were paid by standard paychecks. They could earn the wage of $5.25 per hour. They also could earn a productivity bonus of $5.00 per batch of data entered minus $0.08 per error. A batch represents about 2,000 characters that they could do in about an hour. If they ever had a drug-positive urine sample, they weren’t allowed to work that day or any other day until they provided drug-free urine. They also got a reduction in the batch production bonus from $5.00 to $1.00 per batch.

At intake to this study, none of the women were employed, and only one was married. Almost all of them engaged in behaviors that placed them at risk for HIV infection. All were opioid dependent, and most were cocaine dependent. During the first 6 months, the therapeutic workplace significantly increased abstinence from opiates and cocaine relative to the control group, essentially doubling the rates of abstinence during that time. These effects could be maintained for 3 years.

To measure the long-term outcomes, urine samples were collected once per month from every subject in both groups. The therapeutic workplace participants had about double the rate of abstinence from cocaine compared to control patients. That was maintained for 4 years. There were similar results for patients who were abstinent from opiates for this time. During the fourth year after intake, overall, the therapeutic workplace essentially doubled the rate of abstinence from cocaine compared to controls.

The patients became very skilled data entry operators. In the first year of the business, they entered about 15 million characters and maintained an accuracy of 99.6% correct. Although the business has not yet been financially successful, there have been indications that it can function as a reasonably successful business.

The control patients show no signs of either increasing their employment or abstinence under normal circumstances.