Nudging toward better health Highlights from Nudgeapalooza

24 August 2018

Nudges are being used today in increasingly varied ways to drive consumer behavior.  Similar strategies in health care can lead to better medical decisions and improved patient outcomes, says Mitesh Patel of the Penn Medicine Nudge Unit.

“We're not tricking doctors here, right? We're not pushing, we're not magically making them not know about what’s going on.”

Learn more

Explore the Behavioral Economics collection

Subscribe on iTunes

Listen to more podcasts

Subscribe to receive related content from Deloitte Insights


TANYA OTT: What is going on in hospitals, doctors’ offices, even our own homes—that affects our health without us even knowing it? I’m Tanya Ott and that’s what we’re talking about today on the Press Room.

Do I have your attention? Okay … health care. Back in the day, doctors and nurses used to do a lot of things with pen and paper and face-to-face conversations. But these days we’ve got a whole lot of high-tech in medicine, and its shifting the way medical decisions are made. A decade ago, just one in ten physicians was using electronic health records. Today, it’s closer to nine in ten. And most of us patients have a smart phone in our pocket or purse. More than half of people over 65 have a smartphone. How does affect our health care? Often in very subtle ways that we don’t even notice. They’re called nudges—and some of them are good for us and others aren’t.  

Georgetown University and Deloitte’s Centers for Government Insights and Integrated Research held a one-day conference to explore how behavioral science and data science can help governments, universities, and businesses address some of their most challenging problems. One of the keynote speakers was a guy named Mitesh Patel.

JIM GUSZCZA: Mitesh Patel is a triple threat: MD, MBA, Masters of Science. He's a professor of Medicine and Health Care Management at the Perelman School of Medicine at the University of Pennsylvania.

TANYA OTT: And he's got an appointment at Penn’s Wharton Business School.

JIM GUSZCZA: He’s also practicing primary care physician …

TANYA OTT: … who’s published work in major medical journals and he’s the director of the first “Nudge Unit” with a major medical center …

JIM GUSZCZA: He's a rock star.

TANYA OTT: The Penn Medicine Nudge Unit—which Patel leads—is a behavioral design team that focuses on systematically testing ways that companies—in this case, a health care system—can effectively use nudges to improve decisions and change people’s behavior. The first nudge units started in the UK and did groundbreaking work around getting people to pay their taxes, contribute to charitable organizations, and become organ donors.

MITESH PATEL: I'd like to start off with thinking about the idea of human behavior being the final common pathway for the application of nearly every advance in medicine. No matter how effective a medication is at preventing a heart attack or stroke, and no matter how protective a vaccine is from preventing you from getting a disease, or no matter how targeted we can make a molecule, in order for any of these things to have benefits for patients, really two things at least have to happen. The first is that a clinician needs to recommend or prescribe these to a patient and [second] the patient must both accept them and adhere to them. And because of this, it increases a lot of challenges in health care that we may not see in other spaces. We know that about a third of health care spending is unnecessary and wasteful.1 And we know that there are areas that we do too much of, but there are also areas that we don't do a good enough job in—things that we know that work but we either are not prescribing them to patients or patients have been prescribed but they're not taking them for one reason or another.

We all here are familiar with nudges. These are subtle changes in the design of these environments that can have an outsized impact on our behavior. We usually think of them as meant to either remind, guide, or motivate our decisions towards things that we would want to do and benefit us in the longer term. Some examples might be, and you've heard of some of these already, setting the default to the preferred option of the path of least resistance. Prompting an active choice, so you have to make a decision now as opposed to delaying it into the future. Or thinking about ways that we can frame information either through increased transparency around costs and other issues or understanding how our behaviors relate to others around us, social comparisons that folks have talked about today.

TANYA OTT: That’s where the Penn Nudge unit comes in. It develops platforms and engagement strategies that use nudges to improve daily life. But one-size nudge doesn’t fit all, so they’re really interested in figuring out how to customize nudges. Unlike for-profit corporations that are working in the nudge realm, everything the Penn Nudge unit implements is designed to be submitted to a peer-reviewed medical journals.

MITESH PATEL: I always get this question: How do we identify opportunities? There are a couple of different ways. We first go to the people at the front lines of care—the clinicians, the nurses, the medical assistants that are taking care of patients—because they see the problems that are arising and they have first-hand visibility into what we might be able to do. What are the design flaws that we could try to fix?

When we launched the Nudge Unit in the summer of 2016, we held an innovation tournament, [which was] a way to crowdsource ideas around where we could implement nudges to improve health care. In three weeks, we got 225 ideas, and these ranged from attending physicians to people who are escorting patients around the hospital. And we had a tournament, and got to the top three, and eventually implemented a bunch of them.

TANYA OTT: They were particularly looking for ideas that leveraged technology—specifically, electronic health records and smart phones. But first, they needed buy-in. Physicians want to have autonomy over how they practice medicine, so the nudge team didn’t want to force interventions on doctors or patients. Doctors had to be a part of the process and have input on the design of the intervention. The design also needed to be scalable to a large population, not just a handful of patients. To do that, the Penn Nudge team needed top health system leadership at the table, as well the IT team.

Once they had that, they got to work. The first thing they considered was something called “defaults.” Remember, some of that early nudge team research in the UK centered around organ donation. In Western European countries where residents had to opt out of being an organ donor, the rate of organ donation was much higher than in the US where residents have to opt in.2

MITESH PATEL: And so how do we think about this using this in health care? Well, one challenge we found in medicine was that the generic prescribing rates at Penn were actually in the worst in the county compared to the peers, year after year.3 There's a lot of health systems in Philadelphia and nobody wants to be the worst in any particular area, so you can imagine this wasn't the best thing for us to hear. So Penn was proactive about this. They spent several years educating physicians. How can we get that doctor to change? What we found was after two or three years of doing this, still the last place in the region. And so internal medicine was a little bit more proactive. They decided to change their default in the electronic health record, and I'll show you an example of that design. Around the same time, family medicine was [being prescribed] at a similar rate, but [they] didn't decide to change their default. And so we had a nice comparison group that we could compare to. So, here's the electronic health record for both internal medicine and family medicine before the intervention. This physician is trying to prescribe a brand name medication [for] which [there’s] a generic medicine available. Now, if anyone's been to the doctor's office or anyone here practices, you know you're not just doing one of these. You're done 5 or 6 or 10 or 15, so you're going through them pretty quick. You don't have time to think. You're just rapidly making decisions renewing medications. And the path of least resistance takes you [to] just pick the first one on the list. You've gotta use your mouse and go from one area to the other, and this is one reason why a lot of brand name prescriptions are being prescribed.

So internal medicine changed this. Here, the same physicians [are] prescribing. This is after the intervention and you can see that it only shows you the generic options. Remember, the premise of a good nudge is that we don't restrict people's choice. Particularly with default, you can opt-out. And so you can press this in the top right, the database lookup button, and this will expand the list and show you the full list. Now, this is a button that physicians are generally used to expanding. They do a search. They can't find something. They expand their list. So it still preserves your ability to pick the brand if you want it, but gently guides you to prescribing the generic, just making it easier.

TANYA OTT: Mitesh Patel says before they changed the default, internal medicine prescribed the generic 75 to 80 percent of the time. Family medicine was a little bit better, between 80 and 85 percent. Once they changed the default choice to the generic, internal medicine jumped immediately to 80 to 85 percent.

MITESH PATEL: So we get about five to ten percentage [point] increase, depending on which generic class you look at. We looked at beta blockers, statins, and PPIs (proton pump inhibitors). This was great, really simple. Actually, [it] took no money to do this. An administrator went in and just checked a box.

TANYA OTT: And for family medicine—where they didn’t change the default for the box setting—the level of brand name prescriptions didn’t change at all.

Once the team showed results by tweaking the default choice, they turned an eye to active choice. The idea that a physician can’t avoid making a choice on something—they either have to say yes or no. They can’t just leave it blank. The retail industry does this all the time. And one major pizza chain is particularly good at it.

MITESH PATEL: They know you're hungry. You're ordering pizza. You're hungry. You go through, you order your pizza, you might add a soda, and then you go to hit pay. And they're able to look and see that you haven't added a dessert. And so, they ask you, would you like to add the chocolate lava crunch cake for just 4.99? We know you're hungry, you haven't had dessert. And in order to make it more enticing, they show the chocolate. I don't know if you can see this; it's kind of seeping out of the chocolate cake. It's gooey, it's warm, and there's a choice here. Now what people to feel—that behavioral economic concept that people feel—is an anticipated regret. This idea that they're gonna finish their pizza, and they're anticipating that they’re going to regret the fact that they didn't order that dessert. And so more people are gonna say yes. And you know what they [have] done now, actually, is instead of having the buttons saying no and yes, there's just a big green button that says yes and the x in the top right-hand corner doesn't appear for about five seconds … so it makes it even more aggressive in terms of pushing you to order other stuff. So you can think about how we might do that in health care.

MITESH PATEL: Another area where you've probably seen this is trip insurance. I'm sure everybody here has bought a flight [ticket] at some time or another. You cannot buy a flight without making a decision on trip insurance. And they do several things. They say, protect your trip to New York for $51; it's highly recommended, what ever that means. I don't know who's recommending it.

They put several reasons why you should do this. It saves you time, you can get your money back, all this stuff. And then they put this quote, “Buy trip insurance and it can protect you against cancellation costs.” And that sentence is just very generic and average, but they put the words USA Today, May 2016,4 I think, and that gives it some credibility. The idea being: all of these things [are used] to try to get you to pay for trip insurance, which we know is not worth it, in most cases.

So how do we apply this to health care? We went back to the clinics and we looked at what are some of the problems that they're having? Where can we apply this? And we found shocking statistics in preventative care, within our primary care practices. Patients who come in to see their doctor and are due for annual mammograms, breast cancer screening, nearly 60% of them haven't had a mammogram ordered by the time they leave the office visit, with their annual visit, with their primary care doctor.5 Patients who come in and are due for colon cancer screening come in to see their primary care doctor. Seventy percent of them walk out the door never having colon cancer screening ordered before they walk out the door.6 Patients come in now during flu season. Flu vaccines are recommended to essentially everybody now who's not allergic to them. Eighty percent of patients are not even ordered for a flu vaccine by the time they leave the clinic.7

TANYA OTT: Maybe the conversations are happening, but a patient says, “I just don’t want to have a colonoscopy” or “I’m not getting that vaccine today”…

MITESH PATEL: But, 60, 70, 80% is quite high. There’s something going on in terms of the fact that physicians only have seven to ten minutes with patients.  Patients come in with a list of five or six things they want to talk about and none of these things are on that list. Right? It’s back pain, chest pain, it’s something else that they want to talk about. So how can we use active choice to address this?  Essentially, what we did was we created a smart alert which would prompt the physician to say your patient is due for the flu vaccine. It would check to make sure they haven't already gotten it, they're not allergic to it, they haven't declined it already. Would you like to order it, yes or no? And they just had to click yes; it'd open an order set, they'd just click a button to sign it, and it's all set; or they click no and they can continue doing what they're doing.

TANYA OTT: Some doctors complained about the number of screens they had to click through. But Patel and his team studied it over time and found the approach worked. And not just for flu vaccination. It also increased the percentage of patients who got a colonoscopy or mammography.

MITESH PATEL: We worked with the health system and they rolled this out to all primary care practices in the City of Philadelphia. But before they did that, we made some design tweaks. This was being delivered to both doctors and medical assistants. The medical assistants could template orders. The doctors could either decline them or accept them or the doctors could enter them.

When this rolled out, we had it just go to medical assistants. And so the physicians would never get a prompt with an alert. The medical assistants [are] templating the order, having the discussion, so for a physician, if we have a patient who wants to get the flu vaccine, if the order's there, you just have to sign it. You actually have to do more work to remove it. It's like a default. And the physicians are now, instead of spending time dealing with alerts, talking to the patients who don't want the flu vaccine or don't want the colonoscopy to figure out why, to see if they can address any concerns or to talk about something else.

TANYA OTT: Eventually, Mitesh Patel and his team turned their attention to what happens outside of the doctor’s office or hospital, when patients are at home.   Many insurers offer financial incentives to get people to do biometric screenings, to be physically active, and to make other healthy choices in their daily lives. But the design of those incentives weren’t really tested. They wanted to test different ways to frame financial incentives.

MITESH PATEL: We took 280 overweight and obese employees at our university. We asked them to download a smartphone app and strive for at least 7,000 steps a day. We chose this because we didn't wanna get people who were already going for runs or the power walkers to just walk more. We wanted to get everybody above the minimum level. Seven thousand steps is an area that the American College of Sports Medicine says is kind of the minimum amount of activity you need to start getting your health benefits from physical activity.8 And so we followed them for three months and had a three-month follow-up period, but we randomly assigned them to a control group where you just got daily text messaging. Congrats, you met your 7,000-step goal or sorry, you didn't.

A standard gain priming. This is the way most wellness and insurance programs are launched still today, which is if you do something, we'll then pay you for it after you do it. So every day you meet your goal of 7,000 steps, you get a $1.40.

The next was a lottery where people had a 1 in 5 chance of winning $5 or a 1 in 100 chance of winning $50, and the lottery runs every day regardless of what you do. Mathematically, this comes out to $1.40 a day over the course of a month.

Or a loss-framed arm where we tell you that $42 has been put in a virtual account with your name on it but each day, you don't meet your goal we take $1.40 away. This idea that people are motivated by losses rather than gains.  

Now what's interesting is mathematically, these are all the same. No matter what intervention arm you're in, if you meet your goal ten times at the end of the month, you're gonna get about $14 per day. In fact, we don't actually pay you any different. Everybody gets paid at the end of every month for their accumulated earnings, so the payment’s not different. But even though mathematically, they're the same from a psychological perspective, they're really different, because as much as people like to get $1.40 per day, they really hate to lose $1.40 a day.

TANYA OTT: Here’s what they found. The control group—the ones who just got a text saying, “congrats, you did it!” or “sorry, you didn’t”—met the 7,000 step a day goal 30 percent of the time. The gain incentive group—the ones that got $1.40 for each day—they met the goal 35 percent of the time. Statistically that was no different than not paying people at all. The lottery group did a little bit better, but the loss incentive group—those who were told they had $42 in a virtual account and lost $1.40 every day they didn’t make the step goal—they did the best!9

MITESH PATEL: Again, no difference in the way we paid people. Just the difference in the way we framed it. Nobody could be worse off than when they started, because you could only make money. You actually had no money to lose.  There was no virtual account. We just told people that, but it makes a big difference in terms of how people behave.

TANYA OTT: Here’s one to ponder. One company uses this loss aversion form of motivation by selling employees an Apple watch for $25, but there’s a catch.10 If the employees don’t meet their monthly fitness goals, they have to pay back the entire cost of the Apple watch. Talk about leverage loss aversion!

MITESH PATEL: So in summary, medical decision-making is often suboptimal.  I’ve shown you many examples of that. Subtle changes, just changing the default, asking physicians to make a choice now versus later, nudging people with prices, can have an outsized impact on our environment. And unfortunately, most of the environments have not really been evaluated for their design. And whether we like it or not, the current design is nudging us;it just may not be in the way that we want it to be. And so Nudge Units are behavioral design teams that can really think about this process, use some of the expertise, and implement this in a systematic way to figure out how we can improve patient outcomes and ultimately save patients’ lives … 

TANYA OTT: Dr. Mitesh Patel, professor of Medicine and Health Care Management at the Perelman School of Medicine at the University of Pennsylvania. He’s also director of the Penn Medicine Nudge Unit and was a keynote speaker at Deloitte’s Nudgeapalooza event last year at Georgetown University.

If you hit up our archives you’ll hear more about why we make the choices we do about everything from what we eat, to how we save for retirement and who we love. Maybe, we don’t have nearly as much choice in the matter as we think.

TEASE BITE:There are controversies. There are questions marks. There are issues. People talk about ethics. People talk about the sustainability. People talk about the applicability in different contexts. But that’s what makes this even richer and more wonderful to examine.”

TANYA OTT:  You can find the archive at Deloitte Insights. Tweet us at @deloitteinsight (no S) or tweet me at @tanyaott1. I’m Tanya Ott for the Press Room. Thanks for listening!

This podcast is provided by Deloitte and is intended to provide general information only. This podcast is not intended to constitute advice or services of any kind. For additional information about Deloitte, go to