by Carl Hart
I thought back on what I’d learned about behavior and how it is affected by punishment and reward, going back to B. F. Skinner. Were drugs really that different from other reinforcers or pleasures? I looked at the existing data on that question. In the animal research, the graphs representing how hard an animal is willing to work for a food or drug reward were almost identical: make access easy and provide few alternatives and animals will definitely eat a lot of sweet or fatty food or take a lot of cocaine or heroin.
However, the harder they have to work for any reward—whether it’s a natural pleasure like food or sex or a more artificial one like drugs—the less of it they will tend to seek. This is true whether the animal being studied is a mouse, a rat, a monkey, or a human being. And both in humans and in other animals, these responses will vary depending on the presence of competing reinforcers.
For example, studies have found that when rhesus monkeys have to repeatedly press levers to get either a cocaine injection or a highly desirable food (banana pellets), their responses vary with both effort and dose. Quite sensibly, the monkeys will work harder to get a higher dose of cocaine and put in less effort for a lower dose or placebo. They will also choose larger quantities of banana pellets over smaller doses of cocaine. Even at the highest dose of cocaine offered, these animals will never choose cocaine over banana pellets more than 50 percent of the time.2 Addictive behavior follows rules and is shaped by situations just like other types of behavior. It’s not as weird or special as we make it out to be.
You may say, “Yes, that’s fine with a drug like cocaine that doesn’t produce obvious withdrawal symptoms. But what about a drug like heroin?” Indeed, physical withdrawal symptoms can be seen in chronic opiate (for example, heroin, morphine) users if they abruptly stop drug use. The symptoms usually begin about twelve to sixteen hours after the last heroin dose and look something like a case of the twenty-four-hour, or intestinal, flu. Most of us have experienced these symptoms at some point in our lives: nausea, vomiting, diarrhea, aches, pains, and a general sense of misery. While this condition is most unpleasant, rarely is it life-threatening or accurately depicted in films that suggest the sufferer is on the verge of death.
Throughout the 1960s, drug addiction was defined solely on the basis of the presence of physical dependence (a withdrawal syndrome). About that same time, a group of researchers began publishing findings that questioned this dominant view. They reported that: (1) monkeys would begin and maintain lever-pressing for opiates without first being made physically dependent; and (2) monkeys who had given themselves small amounts of a drug and who had never experienced withdrawal symptoms could be trained to work very hard for their opiate injections.3 More recently, researchers have demonstrated that monkeys’ lever-pressings for heroin injections do not correspond with the timing or severity of their withdrawal symptoms.4 These findings, along with others, underscore the notion that physical dependence isn’t the primary reason for continued drug use.
I started to put these ideas together as I was trying to make my way in academia and dealing with a very unpredictable experience of reinforcers and punishers of my own. Although research careers are rarely presented this way when we are trying to attract youth to science, the reality is that the field is intensely competitive and many highly qualified people do not wind up with tenure-track jobs or even jobs in industry that take advantage of their skills. At UCSF and then even more so at Yale, I came face-to-face with the fierceness of this competition. It was extremely demoralizing at times.
This fight for status was worse than what I’d seen on the street or on the basketball court, where it was at least clear when people were competing and what territory was in dispute. In academia, no one said anything to your face: it was all sneaky stuff, all easily denied or explained away as a “misunderstanding” or “miscommunication.” Men didn’t fight like men; they stabbed you in the back instead. The rules were actually clearer and easier to follow in the hood. But one of the true advantages of my background was that it made me sensitive to social signals, no matter where I encountered them. I was able to learn those used in academia and use them to win, even on such a convoluted playing field.
Nonetheless, there were definitely times when I came close to giving up, when the low salary and grueling work hours with no guarantee of a definite payoff wore me down. The work at UCSF had been disillusioning: as James Baldwin had put it, when you learn a craft well, you get to see its ugly side, and that’s what happened to me, starting there. I felt that the research we were doing on craving was poorly conducted and not productive, that the link between what we were measuring and what happened in real-world drug-using settings was not strong enough to matter. Dr. McCance-Katz was at UCSF on sabbatical at the time and I mentioned these concerns to her, which is how I got invited to do my second postdoc, at Yale. Even there, however, I still had no clear path to that elusive goal of a real job, a permanent tenure-track position. I wasn’t sure I’d ever be able to support my family doing the work I loved. And now, I sometimes hated it. A job at Walmart started to look good by contrast.
To make matters worse, after only months, I learned that Dr. McCance-Katz was soon going to be leaving Yale, which meant my job there would end as well. The viciousness and underhandedness of the competition I experienced during this postdoc was beyond anything I’d ever been faced with before. For example, when I learned that Dr. McCance-Katz was leaving Yale to accept a job elsewhere, I met with a senior member in the department who promised me a faculty position within the department. Later, when I attempted to follow up on the position, this person claimed to have no recollection of our previous conversation, saying that I must have misremembered.
Fortunately, it was at this point that I met Herb Kleber, who was then the director of the division on substance abuse in the department of psychiatry at Columbia. I had a friend who worked with him and said that his program at Columbia was going to be expanding. She introduced us at a scientific meeting and he tried to recruit me with the promise of a faculty position. I was especially excited about the idea of working at Columbia because his wife, Marian Fischman, studied crack cocaine administration in humans. She’d published a paper in the prestigious Journal of the American Medical Association showing that crack and powder cocaine were pharmacologically indistinguishable.5 I eagerly prepared to visit New York for my interview.
However, when I met with Marian, virtually the first thing she said was “I don’t know what Herb told you, but we don’t have a faculty position. We can only offer you another postdoc.” Given the amnesia I was starting to see at Yale, I ultimately agreed to do a third postdoc at Columbia. I didn’t know when this job limbo would end or for how long I could stand it. I certainly wasn’t receiving the rewards of a scientific career that had been expected.
Marian, however, promised that she would do everything she could to help me get a permanent position. She was true to her word. It was at Columbia that I would ultimately get a tenure-track job and reach tenure itself. And in my research there I began finding, as I’d suspected, that humans do respond to cocaine quite similarly to how they respond to other reinforcing experiences. Like the rest of us, people who are addicted to crack cocaine are sensitive not only to one type of pleasure but also to many. While severe addiction may narrow people’s focus and reduce their ability to take pleasure in nondrug experiences, it does not turn them into people who cannot react to a variety of incentives. I began the work that illustrated this as a Columbia postdoc, a job I held from September 1998 through June 1999.
In the study I briefly described in the preface to this book, cocaine users were given a choice between various doses of cocaine and various amounts of vouchers for cash or merchandise.6 On average, on the street, our participants spent $280 a week on cocaine. These were not casual or irregular users.
Marian Fischman’s research group when I arrived at Columbia in 1998. From left, Marian is the fifth person standing. Herb Kleber is seated next to me.
/>
Our procedure worked like this. First, we recruited frequent crack users through ads in the Village Voice and from referrals by other users provided by those who replied to the ads. Then we screened the volunteers for health problems that would ethically preclude their participation in cocaine research (for example, heart disease). We also screened their urine to ensure that it was positive for cocaine, though we did not reveal that we were confirming their use in this fashion.
Those who were cleared to participate were paid to stay for two to three weeks in a ward at Columbia-Presbyterian Hospital in Harlem (now New York–Presbyterian). Before we did any of this, of course, we’d applied for and received special licenses to work with illegal drugs on human subjects and been cleared by an ethics committee called an institutional review board (IRB). Then we obtained the cocaine from a pharmaceutical company, keeping it locked in the pharmacy with other controlled substances, using careful procedures to account for all of it.
On days participants were scheduled to smoke cocaine, each one would sit in a small room with a computer at a desk, where we could observe them through a one-way mirror. A nurse was in a nearby room, monitoring her or his vital signs and lighting the crack pipe when cocaine was chosen. When they smoked crack, participants were blindfolded so that they couldn’t see the size of the rock they were getting. We didn’t want them to have visual cues that might amplify or diminish their expectations about the hit.
At the very start of each day, before having to make any choices, participants had a “sample” trial. That meant that they were allowed to try the dose of cocaine we were making available that day and to see and hold the cash or merchandise vouchers on offer. Both the researchers and the participants were blinded as to whether actual cocaine or placebo was placed in the crack pipe. After the user had sampled the day’s dose, he or she would participate in five “choice trials,” spaced fifteen minutes apart. When a choice was available, an image of two squares would appear on the computer and the participant had to either click the left (crack) or right (voucher) side of the mouse to indicate their choice.
In order to actually get the drug or voucher, they then had to press the space bar on the keyboard two hundred times. During their first four choice sessions, the choice was between a voucher for five dollars in cash or the day’s cocaine dose; during the last four, they had the choice of the dose or the five-dollar merchandise voucher.
Again, the results were similar to those seen comparing different rewards in the animal literature and in earlier human trials. When larger cocaine doses were available, users almost always chose cocaine over the cash or merchandise voucher. So far, this was congruent with the idea that addiction makes people place drugs first. But the rest of the data demolished that theory, showing that lower doses were often resisted. Despite the popular conception that addicted people will choose any dose of drug over any other experience—especially once they’ve already had a taste of it to kindle their craving—this is not what we find in the lab. Even around drugs, addicted people are not simply slaves to craving. They can make rational choices.
This was the case even though the alternative in each choice had only a maximum value of five dollars. In total, our participants could earn up to fifty dollars each day by participating in two complete sessions, which was a significant sum given their usual low income. But if the “first hit produces irresistible craving” theory were true, any dose should have had infinite value during the moment of choice. The cocaine users shouldn’t have been able to think beyond the five dollars to the fifty—or about the particular dose, if the idea that people with addiction are totally out of control once they start using drugs is true.
Nonetheless, on average, users in our studies smoked two fewer doses of cocaine when the alternative was cash as opposed to merchandise.7 This meant that cash money was 10 percent more effective than vouchers in suppressing cocaine use. The conventional wisdom about addictive behavior being completely irrational couldn’t at all account for this result. If people addicted to cocaine always went for drugs no matter what, there should have been no difference.
Because our findings were so different from what most people have been taught about drugs, critics sometimes argued that they only really showed that these crack users were saving their money to buy more cocaine on the street later. That itself, however, doesn’t even support the conventional view of addiction. Weren’t addicted people supposed to be unable to resist drugs that were in front of them and be incapable of saving up for drugs or anything else later? And why would someone turn down pure pharmaceutical cocaine in a legal setting in favor of possibly being beat on the street and getting stepped-on (adulterated) drugs illegally in the future? That would truly be irrational under the logic of the idea of addiction as something that “hijacked” the brain and took control of the will in favor of immediate drug-seeking.
Alternatively, some folks predictably claimed that the users we recruited “weren’t really addicted.” People who were genuinely addicted would never have turned down free crack cocaine, they said. If we’d studied participants with genuine drug problems, they argued, we would have had very different results. Our participants, however, clearly had arranged their lives around crack. They weren’t rich folks who had an extra few hundred bucks a month to spend on cocaine: they typically had unstable living arrangements and few or nonexistent family ties. Many had been convicted of crack-cocaine-related crimes and all had tested positive for cocaine on multiple occasions during the screening process. Most could tell you where to get the best and most inexpensive cocaine in the city. If this wasn’t “real” addiction, what was?
The more I studied actual drug use in human beings, the more I became convinced that it was a behavior that was amenable to change like any other. So why did it seem so intractable in neighborhoods like the one where I’d grown up—and why did people there rarely even question their beliefs about drugs? A key problem is that poor people actually have few “competing reinforcers.” Crack isn’t really all that overwhelmingly good or superpowerfully reinforcing: it gained the popularity that it achieved in the hood (again, far less than advertised) because there weren’t that many other affordable sources of pleasure and purpose and because many of the people at the highest risk had other preexisting mental illnesses that affected their choices.
And that was why, despite years of media-hyped predictions that crack’s expansion across classes was imminent, it never “ravaged” the suburbs or took down significant percentages of middle- or upper-class youth. Though the real proportion of people who became addicted to crack in the inner city was low, it was definitely higher than it was among the middle classes, just as is true for other addictions, including alcohol. Money has a way of insulating people from consequences. In addition, it carries with it more reasons for abstaining—there are things a high-socioeconomic-status person has to do that are incompatible with being intoxicated. Becoming an addict is tantamount to disavowing one’s social niche.
High socioeconomic status provides more access to employment, and alternative sources of meaning, purpose, power, and pleasure, as well as better access to mental health care. The differences in the prevalence of crack problems are mainly related to economic opportunity, not special properties of cocaine. While drug use rates are pretty similar across classes (and often, actually lower among the poor), addiction—like most other illnesses—is not an equal-opportunity disorder. Like cancer and heart disease, it is concentrated in the poor, who have far less access to healthy diets and consistent medical care.
Moreover, research on alternative reinforcers has now shown repeatedly that they can be effective in changing addictive behavior. This kind of treatment is called contingency management (CM). The idea comes from basic behaviorism: our actions are governed to a large extent by what we are rewarded for in our environment. These cause-and-effect relationships where a reward is dependent (contingent) upon the person either doing or (in the case of drugs) not doing a parti
cular behavior can be used to help change all types of habits.
In fact, part of the reason we wanted to compare the responses of crack users to vouchers for cash in our study, as opposed to vouchers for merchandise, was ultimately to understand what types of reinforcement would work best to aid recovery. There is now a whole body of literature showing that providing alternative reinforcers improves addiction treatment outcomes. It is far more effective than using punitive measures like incarceration, which often is less useful in the long run. Although while incarcerated many people stop or at least reduce their drug use, jail and prison themselves don’t provide positive alternatives to replace drug habits. When heavy drug users return to their communities, they are not better equipped to find work and support themselves and their families; instead, having a criminal record and a gap in their résumé makes finding work even harder.
Reward-based CM treatments are sometimes controversial because they can be portrayed in the media as “paying addicts to stop using.” Many people think it’s unfair to those who “do the right thing” by not taking drugs to see drug users getting paid to behave the way they should behave anyway. Cash rewards are especially touchy, since the users could presumably simply buy drugs with the money.
But I see it differently, and here’s why. Indeed, we’ve all probably observed how people respond to rewards in multiple areas of life. It’s often seen most clearly in parenting: for example, if my sons want a new computer, I expect them to maintain a certain GPA. In most workplaces, if the boss offers a raise for achieving certain goals, employees will do their best to hit those targets. Because drug use is governed by the same principles that govern other behaviors, contingency management treatment uses these ideas to change addictive behavior.