It’s one of the paradoxes of being human: even when we know what’s good for us, we often make choices that are less than optimal. For example, patients with prescriptions routinely fail to take their medication, leading to expensive complications down the road. Could offering financial incentives as a reward for making good choices help change people’s behavior in the long run?
Associate Professor of Business Economics and Public Policy Iwan Barankay researches how incentives influence behavior. With the help of high-tech gadgets that make it easier and more accurate to track whether patients are taking their medications, Barankay investigates the effectiveness of using financial incentives to motivate choices. He finds that simply offering a reward is not enough; for the behavioral change to stick, there must be a deeper cognitive connection as well.
Professor Barankay sat down with the Mack Institute to discuss his findings. A transcript of the interview follows.
Economists have studied for a long time financial incentives and understand that they can have predictable ways to shape people’s behavior and decisions. But what we now want to go to is an understanding of how we can shape behavior with time-limited interventions. Put another way, if I pay you to do something, you will do so because it is in your financial benefit for a while. But what happens if I take this reward away? Will you continue to do it or will you relapse to your prior behavior? We now know that there are ways to design these incentives that keep you on that path of behavioral change even though financial incentives have been removed.
A key application of these time-limited financial incentives is now in the health domain. The Affordable Care Act now gives a legal background that has shaped many decisions, both inside insurance companies and independently inside organizations where they use financial incentives to help their employees to make healthier decisions. Here, the key challenge is figuring out what can we do to shape these decisions with time-limited interventions. A key example is diabetics who are at high risk of cardiac arrest or cardiovascular diseases: they have prescriptions, but they often fail to take them even though if they have a heart attack they regret the decision.
The question is, what can we do to help these patients to take their medication? It’s not only in their own interest, but it is very expensive if they have a heart attack and end up in the emergency room, and this will be a very costly intervention for the whole organization.
We have studied financial incentives for patients that have statin prescriptions. We pay them to take their medication. This is possible now with advances in technology where we use remote pill boxes that send through a transponder and the cellular system information about whether people have taken the medication or not. In this way, we can condition feedback and payment to them.
“We now also know that cognitive load, when we force people to think, can have beneficial long-term consequences.”
We now know that if we use these incentives, not only to target the pill taking, but also to design the incentives in a way that generate a deeper understanding of the health consequences or that forces people to think more about their health, they’re more beneficial for long-term behavioral change. An example of this is a study where we try on the one hand paying people to take their medication. On the other hand, in another group, we pay them more if they take their medication before they receive a reminder from us. This gives them an added benefit to take their medication, but it teaches them to be more autonomous and to learn to behave in a way that keeps up with the prescription and builds a routine.
All these studies are still a nascent research agenda. They build on understanding that incentives do work, but patients typically relapse to prior behavior once they have been discontinued. But we now also know that cognitive load, when we force people to think, can have beneficial long-term consequences. The combination of these two approaches, the financial reward but also the financial incentive to think, could yield to these long-term behavioral changes that we hope to see.
All my research uses these financial rewards, but they use them in a way where we do not really interact directly with the individual members. Instead what we rely on are new gadgets, technologies that help us to remotely monitor people and give them conditional and tailored feedback to their current behavior and their current needs. The frontier here is twofold: one, that these gadgets become even less intrusive than they are today, that they become part of the body as much as they might be just part of a smartphone. On the other hand, we will build more and more ideas for machine loading inside so we can tailor the intervention much more closely to people’s current data and their current needs. This is still something that is new, and we don’t quite know how to build that into the system now. Nevertheless, this is well understood to be within reach in the next five years.