Wharton's Mitesh Patel discusses his new research on using lotteries to encourage employees to lose weight.

From gym memberships to free Fitbits, companies are offering an array of incentives to encourage their employees to get healthy and lose weight. It’s an admirable goal with a financial bottom line: Workforce wellness translates into fewer absences and reduced health insurance costs for the employer. Mitesh Patel, a professor at the Perelman School of Medicine and a professor of health care management at Wharton, thinks there’s a better way to motivate employees than with freebies. His most recent research reveals that cash lotteries are powerful incentives for people to lose weight. The paper he coauthored on the topic is titled, “A Randomized, Controlled Trial of Lottery-based Financial Incentives to Increase Physical Activity Among Overweight and Obese Adults.”

Patel, who is director of the Penn Medicine Nudge Unit, the world’s first behavioral design team embedded within a health system, recently spoke about his team’s findings on the Knowledge at Wharton radio show on SiriusXM. (Listen to the podcast at the top of this page.)

An edited transcript of the conversation follows.

Knowledge at Wharton: Why did you pursue this particular research?

Mitesh Patel: Rates of obesity are rising. Nearly 70% of adults in the United States are either overweight or obese. We know that better diet and more exercise can help people lose weight, but the challenge is getting people to actually adhere to those strict regimens. The idea is based on something that’s already prevalent in the industry, which is more than 85% of large employers use financial incentives for health promotion. Most are targeting ways to increase physical activity. Many are starting to offer wearable devices or other types of platforms to help employees track their own physical activity and step counts. We wanted to test what’s the best way to design an incentive to motivate employees.

Knowledge at Wharton: Some companies offer incentives such as gym membership reimbursements for meeting health goals, but you wanted to look at a lottery method. Why?

Patel: Many programs are using lotteries or drawings to motivate people. The challenge of the mechanism that you describe is that it requires you to do a bunch of stuff up front, then it gives you a rebate that’s kind of buried along with other things. We wanted to test ways to pull that incentive out, make it more salient, more visible, and leverage the fact that people are so motivated by lotteries. In the United States, we spend $75 billion a year on lottery tickets, so people are already familiar with it, and it really drives people to change their behaviors. This is meant to [test whether] people will place a bet on their own well-being. Will they bet on themselves to be more physically active and change behaviors?

Knowledge at Wharton: How does it work?

Patel: The standard way is either a simple lottery, which is you have a 1-in-4 chance every day of getting $5, or a jackpot lottery, which is a very large amount of money available. In our study, we used $500, but there’s a 1-in-400 chance. People rarely win that.

Those two mechanisms are hard because people learn quickly that they don’t win in the jackpot, so they get discouraged quickly, and the other one is kind of the same every day. So, we added a third one as a randomized trial. People are randomly assigned to one of the three lotteries or a control arm. The third one is what we call a combined lottery or a two-tiered lottery. Essentially, you have a 1-in-5 chance of winning $5, or a 1-in-100 chance of winning $50. We tested which of these three lottery incentives would get people to be more physically active, compared to control.

“This is meant to [test whether] people will place a bet on their own well-being. Will they bet on themselves to be more physically active and change behaviors?”

Knowledge at Wharton: Was there a significant difference among them?

Patel: The combined lottery worked the best. In the control arm, people reached their 7,000-step per day goal 26% of the time. But in the combined lottery, it was 38% of the time — a 12 percentage point increase, almost a 50% relative increase.

In the other two arms I described — the simple lottery or the jackpot lottery – there was essentially no difference from control. We found in the jackpot lottery that people started really high but got demotivated very fast. By the end of the study, they were actually doing worse than the arm that got no incentive to start with.

Knowledge at Wharton: What do you think accounts for the difference with the combined lottery?

Patel: I think the combined lottery in many ways works like the standard lottery that we think of. When you buy a lottery tickets, there’s a $500 million jackpot lottery. But that’s not the only thing, right? If you match two or three numbers, you might win $20 or $100. The method there is to keep you engaged. You know that you’re going to win something. The combined lottery really leverages that. You win something small, which kind of keeps support, but then you have that large jackpot in front of you. It’s a balancing of the two different mechanisms.

Knowledge at Wharton: How can companies use this information to structure their programs for better health outcomes for their employees?

Patel: Employers are doing this for a variety of reasons. They want to reduce unnecessary costs, but they also want to make their workplace a healthy one. They want to build a culture where people are excited about being healthy, so I think promoting these types of efforts works along all of those lines.

One of the challenges is that most employer-based financial incentives for health promotion are focused on standard economics — this idea that you can use a simple lottery. Either you win a small amount every day, or there’s a large jackpot. In truth, we see mostly that people have the large jackpots. What we found, which was interesting in this study, is those can actually be discouraging because people very quickly realize that they don’t have very good chances of winning.

I think what employers could do is take these findings, along with the other findings we found in other studies, and really take a step back and say, “The way we design these incentives really matters, so let’s put some thought into this. Let’s look at the evidence out there, including this study and others, and think about better ways to align our incentives with the goals of our employees.”

Knowledge at Wharton: When the lottery jackpot gets very high, more people jump in and buy tickets because they’re willing to take that gamble. Was that an element of the study?

Patel: In the $500 lottery, people got really excited. The challenge I think that’s key is the difference between a one-time behavior, like getting vaccinated for the flu, versus a continuous behavior, like reaching your step goal every day.

If it’s a one-time behavior, then a larger lottery could work because people get really motivated at the beginning. They might get their flu shot, and if they don’t win, you have a whole year before you have to run the lottery again. But in physical activity or weight loss or diet changes or taking your medicines, these are daily behaviors that require you to be engaged not only every day but for a long-term period. Those types of things require different types of incentives.

Knowledge at Wharton: A part to this process that doesn’t get talked about enough is making sure you have that involvement, that daily connection with employees to help them improve their health.

Patel: Yes, I think there’s a lot of inertia up front in getting people engaged in these programs, so that’s where a lottery or a large jackpot might really help. But then once they’re engaged, you’ve got to keep them engaged. We see a lot of people who set New Year’s resolutions. They’ll start by joining a gym or start a new diet or Weight Watchers, and then they very quickly fall off the wagon. We need to think about how we can design these [programs] so we keep people engaged.

“You’re more likely to eat better if you work, live and interact with people who eat better.”

Ultimately, we’d like to help individuals to build healthy habits, and I think that requires a couple of steps when it comes to incentives. The first is, understanding which design can help people in the short term. That’s the research that we’ve done. The next is asking, “How can we help people in the long term?” Running the incentives longer — that’s research that we’re currently working on. Once we’ve developed that, we can start saying, “How can we deploy a fixed-term lottery that, when we stop, people continue to be healthy afterwards?” I think we’re just starting to move in the direction where we can test that.

We have had some success by combining financial incentives with social incentives, because we know that social relationships and connections to our families and our friends are long-standing. So, there may be a way to marry those two types of approaches together to help sustainability.

Knowledge at Wharton: What are the social incentives?

Patel: It’s things like if you were to go to the gym by yourself, you’d probably have less success than if you went with a buddy. You hold yourself accountable to that person on days when you’re not motivated. On days when the other person is feeling down, you can be [the motivating] person. We can think about ways that we can leverage that with financial incentives. We could make the incentives contingent on competition between those two people, or we could have them work together or support each other. Then once the incentive stops, those two people are still going to the gym together. If you’ve been able to build that dynamic, you have that support system in place.

Knowledge at Wharton: I would think that would improve the overall culture in the office, especially when it comes to well-being.

Patel: Yes, that’s what we would like to see. There is evidence that you’re more likely to be physically active if the people around you are physically active. You’re more likely to eat better if you work, live and interact with people who eat better. If we can find the right people and get a certain proportion of people to change their behaviors, we’d hope in the long term that we’d see spillover effects. I think it’s part of building that healthy culture. Social influences are very powerful. When you see other people doing certain things, it encourages you to do the same.

Knowledge at Wharton: Will you try to expand this research?

Patel:  The studies we’ve done so far are from the range of three to six months. We have a couple of studies in progress that are a year, a year-and-a-half long, testing different types of financial and social incentives. We want to look not just at physical activity and weight loss, but also at clinical outcomes, like changes in your blood sugar control if you’re diabetic, or your blood pressure or cholesterol if you’re at risk for cardiovascular disease. Those are areas where we’re really focusing on leveraging the work we’ve done to expand it for a more clinical scenario.

Knowledge at Wharton: You conducted this study with people working at specific companies, correct?

Patel: That’s correct. One of the benefits of doing that is you have a large population that wants to do this. The employer is aligned with that. In this study, we used overweight or obese individuals. Now that we’ve found things that work in those populations, we’re trying to focus more on people with high-risk clinical things like uncontrolled diabetes, heart attacks, high cholesterol and so on. We also want to look at people who may have lower socioeconomic status or don’t have access to the same technology that someone who has a large employer might have.

Knowledge at Wharton: What has been the reaction of the employers that you worked with?

Patel: I think the reaction has been really positive. These are typically employers who have been working for years to make their programs better. By showing them that if you just do a simple test three different ways — the same monetary amount, but just designed differently — we can have a dramatic change in outcomes. So, [the response has] really been enthusiastic. We now have a number of partners that we work with in the employer and non-employer setting who are interested in scaling these more broadly to their populations.

Knowledge at Wharton: Did you talk to some of the study participants afterwards about their reaction to this research?

Patel: Yes, we surveyed all 209 people who participated in the trial and asked them, “What were things that worked well? What were things that didn’t?” We had a really positive experience in that we had employees who told us they had been trying to become more physically active for years and just hadn’t found the motivation to do so, and our study really helped them to do that.

“I think a lot of our work has shown that simple changes in design can have a dramatic effect.”

In this study, we used smartphones to track step counts. There’s a lot of excitement about wearables, which we also study. But people really use their smartphones a lot because they carry them with them all the time. We had more than 95% of people who started the study finish the entire six-month trial. You don’t typically see that. Usually, you find about half of them stop using it within a few weeks. We were really encouraged by that.

Knowledge at Wharton: Where do you take this next?

Patel: We have a couple of studies currently. We have a clinical one with uncontrolled diabetics. We have one with a large employer — Deloitte Consulting. We have 602 of their employees from 40 states in this trial. Deloitte was interested in partnering with us to test how we could leverage social incentives. We’re working with them to figure out the best way to motivate their employees in a way that we could potentially leverage, to expand not only to their population but also to other employers. We’re doing this on a nationwide level.

If we can get employees to be more physically active, lose weight, take their medicines — it’s beneficial for both the employee and the employer. I think a lot of our work has shown that simple changes in design can have a dramatic effect.

The average large employer in the United States spends almost $700 per employee per year for health promotion, so they’re already spending a lot. The number we used in this study was about $1.40 per day, which over the course of the year is $550. That’s what they were spending back in 2014, when we started planning for the trial. Now it’s over $700. There’s already lots of money out there. The real question is, how do we use that money in the best way? They have thousands of dollars to gain just by getting someone to be more physically active, to lose weight, to get their diabetes under control, to quit smoking. The potential benefits here are tremendous.

Knowledge at Wharton: You mentioned demographics, and therein lies an interesting question. For people in low-income situations, it can be harder to fight what may have been familial patterns in overeating, drinking, whatever it might be. Is that a bigger challenge?

Patel: Each population has different challenges. People with lower socioeconomic status may not have a smartphone or other types of technologies that are engaged in [these types of programs]. Or they may have competing challenges, like you mentioned, in terms of things that are going on in their lives or other struggles and challenges.

For the employed population, we still see challenges. They’re just different. They’re working really busy schedules from 7 a.m. to 7 p.m. in an office, at a desk. When are they going to get the time to go to the gym?

The design of these programs is typically one-size-fits-all. You pick a social incentive or a monetary incentive, and you deploy it to thousands or millions of people. But each person is probably going to vary in the way they respond. [So,] another [aspect] we want to work on is how we can personalize [these programs] or target them to people in a better way. Some people might respond better to financial incentives. Other people might be turned off and respond better to social incentives. That’s part of the work we’re doing. It’s understanding who responds and who doesn’t, so that we can make progress on that front as well.

Knowledge at Wharton: That’s a challenge, because companies do think one-size-fits-all. They want to put out a program, and have the people adapt to that program. You’re saying we need to think of it from the other perspective — having the program adapt to them.

Patel: Exactly. I think the program itself could be adaptable, especially with the rise in technology. Are people responding? If not, we can move them to another intervention.

But there are also simple things [we can adjust]. For example, most programs give people a step goal — let’s say 10,000 steps. We know that’s really hard. It would be better if we asked people to get a baseline level, so we know who’s walking 3,000 steps from the beginning and who’s already walking 9,000. Then we say, “Choose a goal. We can tailor it to you.” By choosing a goal, you feel more motivated because you picked the goal, as opposed to us assigning it to you.