Giving a gift or a cash incentive to someone to give up an addiction sounds like a prize for behaving badly, but the practice works. The real challenge is deciding who should pay for it.
Anna smokes almost two packs of cigarettes per day. Her brand of choice is Marlboro, full flavour, no menthol. Anna is 14 weeks pregnant, so that is what her fetus smokes, too.
Aged 33 with short, curly blondish-red hair, and a hot-pink iPhone dangling from the pocket of a hooded sweatshirt, Anna (not her real name) is at the University of Vermont's Center on Behavior and Health in Burlington to learn about a smoking-cessation study for pregnant women. She knows that smoking is bad for her baby. She also knows that she cannot afford the habit, which costs her more than US$7,000 each year — approximately half of what she earns as an in-home childcare provider.
At the laboratory, Anna learns that she will be randomly assigned to one of two treatment groups. Treatment A will involve three brief educational sessions and nine phone calls from a trained smoking-cessation coach. At each visit, including the one today, she will earn $50; completing all 9 calls will earn her an additional $65. She will also receive money for having ultrasound scans taken of her baby at 30 and 34 weeks and for participating in 8 additional assessments in the year following the birth of her baby. Participants receiving treatment A can be compensated by up to $655 for their time.
In treatment B, the 'incentives for abstinence' condition, the protocol for participants is the same as for those in treatment A, but they are also screened for drug use through frequent breath and urine tests at the clinic, and are rewarded when their results are negative. The women earn gift cards that can be used to pay for things such as baby clothes, films, fuel for their car and groceries. In addition to the $655, participants in treatment B can earn $1,200–$2,400 in incentives. “You guys definitely got the right idea, throw money at people,” says Anna, who is hoping that she will be assigned to treatment B.
Giving tangible rewards to reinforce positive behaviours, such as abstaining from drugs, underlies a field known as contingency management. Studies1 have shown that rewards that are contingent on good behaviour help people to refrain from addictive habits, including taking drugs or eating unhealthy food. The rationale is that financial incentives activate the same reward systems in the brain as addictive behaviours.
Moreover, even though quitting an addiction improves a person's life in the long run, research has demonstrated that people are biased towards immediate gratification1. In other words, receiving an instantaneous cash reward is more desirable than saving the money that would have been spent on drugs and medical care over a long period.
You're providing tangible rewards to people who were engaging in illegal or unhealthy behaviours.
But despite proven results, treatment schemes like this are rarely offered. “I think in some ways that relates to the controversial nature of it,” says Nancy Petry, an addiction specialist at the University of Connecticut in Farmington. “You're providing tangible rewards to people who were engaging in illegal or unhealthy behaviours.” To help to change that perception, an array of scientists are working on ways to improve contingency management by tweaking how it is offered, in the hope that it will finally cross the boundary to an acceptable and commonplace treatment.
Offshoot of the cocaine epidemic
Contingency management arose out of the work of psychologist B. F. Skinner, who challenged the idea that all behaviours stem from free will. Instead, through experiments on rodents that he began in the 1930s, Skinner showed that behaviours can be reinforced by systematic punishments or rewards.
Addiction seemed to be a suitable disease on which to test Skinner's ideas. Addiction is “a concrete behaviour, one that's very difficult to manage, but can be very easily measured”, says Maxine Stitzer, an addiction-treatment researcher at Johns Hopkins University in Baltimore, Maryland.
Case studies dating to the 1960s and 1970s provided tantalizing evidence1 that financial and other incentives might help people to quit drugs or lose weight. But the field really took off in the 1980s with the rise of the cocaine epidemic. Many pharmacological and psychosocial interventions failed, but financial incentives dramatically increased cocaine abstinence.
In the 1970s, Stitzer, one of the first researchers to study contingency management with drug users, and her colleagues developed a simple protocol for people who were addicted to a substance: rewards for negative drug tests and nothing for positive tests. She experimented by providing people who are addicted to opiates with special privileges, such as letting them take methadone (a medication used to control cravings) at home rather than at a clinic. In one study2, one group of patients could automatically take methadone home, but another group had to earn the privilege through abstinence from other drugs. Over a four-week period, Stitzer found that people who had to earn their privileges were four times more likely to abstain from drugs (32% abstinence versus 8%) than those who received automatic privileges.
Vouchers allow the individual to express what's rewarding for them.
But researchers soon began to favour cash incentives. “Everybody likes money,” Stitzer says. “And more importantly, you can vary the amount.” Also, unlike Stitzer's methadone approach, which was specific to people who were dependent on opiates, cash incentives can be applied to any addiction. In the 1990s, Stephen Higgins, director of the centre that Anna visited in Vermont and a former postdoc in Stitzer's lab, developed a financial incentive programme for cocaine addiction. Outpatients with negative test results received vouchers that could be redeemed for retail items. By giving vouchers instead of cash, says Higgins, patients can buy something meaningful without using the money for drugs. “What's a reward for you might differ a lot from what's a reward for me,” Higgins says. Like cash, “vouchers allow the individual to express what's rewarding for them”.
By the 2000s, Stitzer had combined the methadone take-home privilege with vouchers. In one experiment3, she divided cocaine users on a methadone treatment programme into three groups: one group received take-home methadone privileges for negative test results; a second received the take-home privileges plus vouchers totalling up to $5,800 over 52 weeks; and a third followed the normal protocol of taking methadone at the clinic. Both sets of people with the take-home privileges submitted three times more drug-negative urine samples than the normal maintenance group over a one-year period. Furthermore, those who received vouchers showed longer periods of abstinence.
A number of meta-analyses4 have shown that incentives work considerably better than more conventional addiction-management programmes, particularly when vouchers are added to the protocol. What is more, says Higgins, vouchers are the only treatment so far that works for cocaine users.
Contingency management is popular in helping people to stop smoking, an addiction that is seen as more amenable to low-intervention treatment than addictions to other drugs, such as heroin or cocaine. To give an example, in Scotland's deprived Greater Glasgow region, 20% of women report smoking during pregnancy. David Tappin, who specialises in clinical trials for children at the University of Glasgow, recruited more than 600 women, all of whom smoked. At their first prenatal appointment, he randomly assigned half to the routine smoking-cessation programme run by the UK National Health Service (NHS), and the others to the service plus the possibility of receiving up to £400 (US$610) in shopping vouchers. By the end of their pregnancies, Tappin found that just shy of 9% of women in the control group had stopped smoking, compared to almost 23% in the incentive group5.
But moving contingency management from the lab to the real world has proved to be challenging. The problems are multifold. For one, asking participants to travel frequently to a drug-testing centre is often untenable. Even though Anna lives only 8 kilometres from the University of Vermont, getting to the lab took her one hour by bus. Then there is the issue of acceptability. To many taxpayers, giving money to people addicted to a drug seems baffling — or morally wrong. But the real hitch is: who should pick up the tab?
Financial incentives make sense in closed health-care systems in which a single entity — such as the NHS — covers all of an individual's health costs, says Petry. Although the initial pricetag may seem high, contingency management is cheaper than treating addiction with more conventional programmes down the line. It can also thwart the sorts of chronic health problems that can arise from prolonged use of a given substance.
Tappin estimates that it would cost the NHS £400,000 to roll out a contingency-management programme for all pregnant women who smoke in and around Glasgow. By comparison, he says, the NHS already spends £4.6 million per year on cholesterol-lowering statin drugs in the same region. Compared to other treatments or procedures already in place, “it's quite a cheap intervention”, he says — and cheaper still if it improves the long-term health of mothers and their babies.
In countries such as the United States, which lack a single-payer health-care system, contingency management can be more difficult to implement. But it can work for specific groups. For instance, the Department of Veterans Affairs, which covers all the health needs of military veterans, adopted a contingency-management programme across its outpatient substance-abuse clinics in 2011 (Petry is an independent consultant for the department). Individuals with negative test results are rewarded with gift cards to the Veteran's Canteen Service, which sells numerous goods at a discount.
Researchers have also started to experiment with alternative methods to overcome the cost and delivery barriers associated with giving every participant a high-value reward when he or she has a negative drug test.
One way to reduce costs is to encourage people to buy into an incentive programme with their own money, a practice known as a deposit or commitment contract. That money then has to be 'earned' back through negative urine tests.
In a study published this year6, researchers put the concept to the test with more than 2,500 employees of CVS Caremark, a pharmaceutical chain based in Woonsocket, Rhode Island, plus their relatives and friends who wanted to stop smoking.
They divided those individuals into several groups, including one whose members could receive $800 if they abstained from smoking for six months and another group whose members had to pay $150 to participate. They could then earn back that $150 plus a $650 reward if they did not smoke for 6 months.
The researchers found that both the reward and deposit groups quit at higher rates than those receiving more conventional care, such as nicotine replacement therapies and counselling. Moreover, those who paid the $150 deposit were twice as likely to quit than those receiving a straightforward reward. But there was a catch. Because participation was optional, 90% of those assigned to the rewards group agreed to give abstinence a try, compared with just 14% in the 'penalty' group. “The data are overwhelming in showing the limitation of deposit contracts. The vast majority of folks will not join them,” Higgins says. “The straight incentive programme is the way to go.”
Other solutions have proved to be more successful. For instance, the Department of Veterans Affairs uses a 'fishbowl' model developed by Petry in which participants with negative test results are eligible to draw a slip of paper from a bowl in much the same way as a raffle. Rather than giving everyone the same lump sum, prizes on the slips vary — half of the slips carry no reward, others are for vouchers ranging in value from $1 to the cost of a new TV. The fishbowl model substantially reduces costs, Petry says, with results rivalling those seen with conventional voucher programmes7.
Researchers at Johns Hopkins have developed the 'therapeutic workplace', in which jobseekers who abuse substances and present negative drug-test results are rewarded with the opportunity to work and earn wages.
Others are looking at ways to ease the delivery of contingency management. Jesse Dallery, a behaviour analyst at the University of Florida in Gainesville, has developed a model in which people who smoke are sent a carbon monoxide testing kit in the post. Using a web camera or smartphone, participants video themselves blowing into the machine and holding up the value. Negative results are rewarded with vouchers, or by returning a cash deposit made by the participant (see 'Smoke free').
Dallery's former graduate student Bethany Raiff, a behaviour analyst at Rowan University in Glassboro, New Jersey, is now looking to couple that online test with video games — one for smartphones and one for Facebook — in which people who smoke earn rewards in the game instead of money or vouchers. She has partnered with video-game designers who specialize in making health-focused products. “You have to make the games fun and addictive,” Raiff says — and then corrects her choice of word — “but not in a bad way”.
Each of Raiff's games lasts about five minutes — the time it takes to smoke a cigarette. To stop people smoking while playing, Raiff and her colleagues are incorporating preventative strategies, such as two-handed playing. In the long-term, Raiff would like to work with insurance companies who have a vested interest in keeping clients healthy. Some, she notes, already cover preventative care, such as gym memberships.
Back in Vermont
Although contingency management is more effective than conventional treatments, similar issues arise in both approaches — patients struggle to stay abstinent. Bolstered by his short-term success in getting people who have become dependent on cocaine to quit, Higgins began searching for a population in which short-term changes could produce substantial benefits.
Twelve years ago, Higgins began his research into helping pregnant women who smoke to end their habit. Even short-term abstinence, he reasoned, could dramatically improve their babies' health. So far, his research has shown that women in the incentive group carry larger fetuses, as measured by ultrasound scans at 30 and 34 weeks, and give birth to larger babies. Now, he is offering incentives to women to continue to abstain from smoking for three months after they give birth, and is considering extending that time frame. His research indicates8 that providing a reward after childbirth helps women to breastfeed for longer and decreases the risk of postnatal depression.
Back at the centre, Anna finds out that she is in treatment A and will receive compensation for her time but no extra incentives. “There are benefits to both,” Allison Kurti, the postdoc running the study, tells her. “If you've got a real busy schedule, it's kind of nice if we don't have to see you as often. Some people find that more convenient.” Convenient, perhaps, but as effective? Anna's not so sure. Just as Skinner observed in rodents so many decades ago, rewards and reinforcement really are key. “I know if someone's on my ass,” Anna says, “I'm more apt to do it.”
Higgins, S. T., Silverman, K., Sigmon, S. C. & Naito, N. A. Prev. Med. 55, S2–S6 (2012).
Stitzer, M. L., Iguchi, M. Y. & Felch, L. J. J. Consult. Clin. Psychol. 60, 927–34 (1992).
Silverman, K., Robles, E., Mudric, T., Bigelow, G. E. & Stitzer, M. L. J. Consult. Clin. Psychol. 75, 839–854 (2004).
Benishek, L. A. et al. Addiction 109, 1426–1436 (2014).
Tappin, D. et al. Br. Med. J. 350, h134 (2015).
Halpern, S. D. et al. N. Engl. J. Med. http://dx.doi.org/10.1056/NEJMoa1414293 (2015).
Petry, N. M. & Bohn, M. J. Sci. Pract. Perspect. 2, 55–61 (2003).
Higgins, S. T. et al. Prev. Med. 55, S33–S40 (2012).
Related links in Nature Research
Related external links
About this article
Cite this article
Gupta, S. Contingency management: Why it pays to quit. Nature 522, S57–S59 (2015). https://doi.org/10.1038/522S57a