Science Watch

It was a chilly February day in Concord, N.H., and the audience of a hundred judges, prosecutors and probation officers facing psychologist Douglas Marlowe, JD, PhD, was perhaps even chillier. They'd invited him to speak because newly enacted legislation made it tough to incarcerate nonviolent addicted and mentally ill offenders. Instead, the criminal justice system would provide alternative programs, including drug courts that reward good behavior and punish missteps — a system that many of Marlowe's audience members viewed skeptically.

"I know what a lot of you are thinking right now: ‘You want me to take these high-risk antisocial, addicted individuals and give them," — Marlowe switches to a mocking voice — "positive reinforcement." You're thinking, ‘Why should I give them rewards for doing what the rest of us are legally, ethically and morally required to do anyway? Who rewards me for not breaking the law? Who rewards me for not doing drugs?' But you have to remember there was a time in all your lives when people did reward you for doing simple, basic things. People applauded when you went poopie in the potty. ... Your families did an extraordinarily good job of socializing you, and that is why you are all judges and lawyers and corrections officers today."

It's a talk Marlowe has given many times as chief of science, policy and law for the National Association of Drug Court Professionals. And it works, he says. If you remind people of their own learning histories, they become more receptive to the idea that even hardened criminals can change. Then, he tells them how to use psychology's fundamental laws of reinforcement to help drug-addicted offenders quit — the technique psychologists call contingency management.

For example, some judges send offenders to jail the first time they fail a urine test. But research shows that it's more effective to have gradually escalating penalties. Positive reinforcement — in the form of token gifts and pep-talks from judges — also improves abstinence and recidivism rates. But perhaps the most important quality of a good drug court is regular drug testing and swift consequences, Marlowe says.

"The courtroom is the perfect place to deliver a contingency management intervention," says Marlowe. "I think Skinner himself would have loved a courtroom where punishment and reward were delivered systematically."

The courtroom, however, is just one of many venues where psychologists are applying contingency management on a larger scale than ever before. After spending the better part of a century germinating in psychology labs, psychologist-designed programs are finally taking root in the wider world, especially in drug treatment programs and company wellness initiatives. The results, so far, are nothing short of staggering: Homeless people with HIV are remembering to take their medications, cocaine addicts are showing up to work on time and drug-free, and already healthy workers are becoming even healthier, by increasing their gym attendance and refilling prescriptions on time. As for drug courts, those that faithfully apply principles pioneered by B.F. Skinner are reducing recidivism by upward of 35 percent, according to a research review by Marlowe (in the Chapman Journal of Criminal Justice). That success has spurred a huge uptick in drug court participation nationwide, to the point that every state now offers drug courts, says Marlowe.

In essence, Skinner is scaling up.

"We're on the cusp of a new generation of contingency management," says Joshua Klapow, PhD, a public health professor and contingency management researcher at the University of Alabama at Birmingham. "We're taking mom-and-pop projects, designed for a few hundred people at most, and learning how to apply them to whole populations."

Pay to play

Contingency management programs are often deceptively simple. Most reinforce good behavior by giving people cash or vouchers. For example, a program developed by James Sorensen, PhD, a University of California, San Francisco, psychology professor, paid HIV-positive methadone patients for taking their antiretroviral medication. By giving 66 participants vouchers worth an average of $5 per day, Sorensen and his colleagues increased pill taking from 56 percent to 78 percent, according to a study published in Drug and Alcohol Dependence (Vol. 88, No. 1).

Getting people to take lifesaving medications may not seem like a major accomplishment, but many of the study's participants were homeless, which makes keeping track of medications — not to mention refrigerating them — difficult, Sorensen says. Previous attempts to increase medication compliance by getting people into housing and providing counseling were popular with patients, but didn't improve their medication taking.

"HIV is not the highest thing on your priority list when you are hungry and living without shelter," Sorensen says.

It may seem obvious that paying people will encourage them to do what you want — that is, after all, how all jobs work — but the details are critically important, Sorensen says. For instance, you can't just ask people if they took their medications; you have to objectively verify it and provide reinforcement as quickly as possible. To do that, Sorensen uses MEMS caps, medication bottles that record the time and date of every opening. Also, you don't just want to pay a flat rate — escalating payments for consecutive instances of good behavior are far more effective, according to more than a decade of research by Stephen Higgins, PhD, a psychology professor at the University of Vermont.

"You want people to be more invested as time goes on," Higgins says. That technique worked well in a 2010 study published in Addiction. In the study, 166 pregnant smokers provided a urine sample twice a week. If that sample showed no evidence of smoking, the woman earned a shopping voucher worth $6.25, a sum that grew by $1.25 for each consecutive clean sample. If a participant slipped, that reset the payments to $6.25, though participants could get back to their highest-previous payout for returning two more consecutive clean samples.

This payment system may seem complicated, but it helped 34 percent of the women quit smoking, compared with 7.4 percent in the control condition. It also resulted in healthier babies, with women in the incentive program having low-birth-weight infants only 6 percent of the time, compared with 19 percent in the control condition.

So, if contingency management is so effective, even with people who have otherwise intractable problems, what's the catch? They can be more expensive than treatment-as-usual, and many people worry about the ethics of paying some people to do what everyone else does for free. However, such qualms are lessening in the face of the growing health-care crisis, Higgins says.

"There is some resistance or discomfort with the concept of economic incentives, but people are beginning to see that we all pay for unhealthy behaviors anyway," he says. "If somebody is engaging in cigarette smoking, or not managing any chronic illness, or engaging in a sedentary lifestyle, we all end up paying for it through insurance pools, Medicaid or Medicare."

Secrets of sustainability

Another criticism of contingency management programs is that, like most interventions, effects tend to fade after the program ends. For example, HIV-positive participants in Sorenson's study returned to taking their medications only half the time as soon as the study ended. One solution is to continue the program indefinitely — a tack which, while expensive, would save the health-care system money by reducing the spread of AIDS, Sorenson says.

Another solution is to have contingency management programs pay for themselves. That's the tack Kenneth Silverman, PhD, a psychology professor at Johns Hopkins School of Medicine, is taking. In his lab, drug-free urine samples are the price of admission to a data-entry workplace that pays an average of $10 an hour. The company, Johns Hopkins Data Services, provides its employees — all of whom have drug problems — with a higher base pay rate for workers for each consecutive clean urine sample and gives bonuses for good job performance.

Corporate and university clients defray the cost of the program, but it still requires grant money to run. As a result, people can only work at Johns Hopkins Data Services for the duration of a study — a few years, at most. However, three Maryland employers have expressed a willingness to hire program graduates, who would continue to take random drug tests as a condition of employment, says Anthony DeFulio, PhD, associate director of the program.

"We are harnessing the power of wages to maintain drug abstinence," DeFulio says.

Another way to fund contingency management programs is to have people pay for them themselves. That's the business model of the website www.stickK.com, where you set a goal for yourself and put a price on that goal — say, to lose a pound a week for $5 a week. If you meet your goal, you keep the money. If you don't, that $5 goes to what stickK.com founder Jordan Goldberg calls an "anti-charity," a group that you have identified that you strongly disagree with.

"If you're for gun control, you'll be extra-motivated to keep us from sending your money to the National Rifle Association," says Goldberg, adding that you can choose other consequences, such as betting against friends or sending money to causes you support. (The company makes money through advertising and corporate partnerships.)

These kinds of automated systems are driving the movement of contingency management from small-scale studies to major population-level applications, says Joe Schumacher, PhD, a professor of medicine and contingency management researcher at the University of Alabama at Birmingham.

"Incentive systems are going high-tech," he says.

Schumacher is best known for his work using contingency management to get homeless people with crack addictions to quit, but he's now extending that work to larger populations through a Birmingham-based company called ChipRewards.

ChipRewards recently collaborated with Chattanooga, Tenn., to create a program for 3,200 of the city's workers. They adapted software originally created for business loyalty programs to monitor how often employees, for example, go to the gym, refill prescriptions or attend preventative health screenings. The program automatically issues employees good-behavior points, which can be exchanged online for a variety of products.

That means the same computer program that encourages you to eat nine burritos to get one free may soon help you lose the weight you gained eating all those burritos.

ChipRewards has been hired by several large companies and is being used by more than 100,000 employees, says Klapow, ChipRewards' chief behavioral scientist. In the future, programs like his could create large-scale medication adherence programs for pharmacetical companies, since they stand to earn money if people remember to refill their prescriptions. Adherence also helps people better manage chronic illness and reduces the nation's health-care costs, Klapow says.

Of course, not everyone is enthusiastic about the idea of having companies electronically monitor and reward healthy behaviors. It's one thing to use contingency management to help people with drug addictions, but it's quite another to apply these programs to the wider population, says George Loewenstein, PhD, a behavioral economics professor at Carnegie Mellon University.

A case in point, he says, is weight loss. We can pay people to eat more healthfully — in fact, he did just that in a randomized controlled trial in the June Journal of General Internal Medicine. In the study, participants in the contingency management condition lost an average of eight more pounds than people in the control condition. But to address widespread obesity, it's better to lower the cost of fresh fruits and vegetables and raise the cost of processed food, he says.

"Contingency management tackles the problem at the individual level, but we risk losing sight of the real underlying causes of the problem and possibly even blaming the victim," he says.

Then, there's the larger issue of free will. As contingency management systems spread, will we begin to see ourselves as nothing more than rats in Skinner boxes?

"There's the potential of going overboard," Loewenstein says. "Ending up in a ‘Walden Two'-type society doesn't seem like such a great outcome to me."


Tour the drug-free workplace designed by Johns Hopkins psychologists at our digital edition.