Mitesh Patel is a practicing physician, as well as a faculty member at both the Penn Medicine Center for Health Care Innovation and Penn’s Center for Health Incentives and Behavioral Economics. Dr. Patel is also an Assistant Professor of Medicine and Health Care Management at the Perelman School of Medicine and The Wharton School at the University of Pennsylvania. He is well known for his research on behavioral economics where he and his colleagues are discovering ways to improve and elicit healthy behavior. Dr. Patel’s thought leadership has been featured on CNN, NPR and in The New York Times, and his scientific findings have been published in several prestigious journals including the Annals of Internal Medicine, the New England Journal of Medicine, and the Journal of the American Medical Association.
1) As a physician interested in health, what do you make of the recent UCLA study that suggests BMI is a poor performance indicator? Although the extremely high recidivism rates we hear in lay media are generally inflated, programs that focus solely on weight loss programs seem to be falling out of favor. Is there a better approach to gauging and influencing toward behavior that contributes to wellness?
The challenge with using BMI is akin to the challenge of using any kind of score or metric for a population of people. There is always going to be a gray area. For instance, someone with a BMI of 29.9 is overweight, but someone with a BMI of 30 is obese. Even though there is a very, very small amount of difference between the two, when you categorize someone through this lens it can be classified as a significant difference. So this challenge I just described with BMI will be comparable with a lot of other standard measures.
So what many companies, employers and insurers are trying to do is find more holistic ways of looking at people’s health. That is where it gets complicated, someone might have a low BMI, but have diabetes, and the right intervention is weight loss. This is an example of why using any metric in isolation is challenging. I do believe outside the context of the BMI measure, losing weight for overweight individuals is generally known to be beneficial. There is generally never harm in getting your BMI down to a lower range if you are above 25. However, that said, you certainly can find people with a BMI of say 32 that live to be over a hundred, but on average people in our current population are healthier if they lose weight.
A common problem with some wellness programs is they are often one-size-fits-all. For instance, lose 10 pounds and get a reward, but really we need to do a better job at personalizing to the individual. This highlights the importance of paying attention to how these programs are designed. We are facing complex problems, and oftentimes we are meeting these problems with solutions that are frankly too simple.
2) Outside of monetary incentives, what do you believe is most important for a company/organization to get right to best set themselves up for positively supporting employee well-being?
This brings us back to the importance of the overall design of the program. Is the program designed in a way that it will produce the results the company is expecting to get? Let’s say the goal is to increase everyone’s activity level, so the company gives everyone a free Fitbit, sets up a leaderboard to see how much everybody is doing and then creates a competition because competition can drive people to change behavior. The problem with this hypothetical solution is the program will motivate the people who are the top of the leaderboard — the people that tend to be already motivated — and demotivate the 95 percent of people that are not at the top of the leaderboard. I don’t think this is the right approach because it excludes the people you want to reach the most. We have done a couple studies where instead of setting a high bar, we set a threshold instead. For instance, in one study we set the threshold at 7,000 steps. The average American gets 5,000 steps, so the goal (in this particular study) is about a 40 percent increase in steps for most users. What this does is create a program that will reach more sedentary people than simply people who are already highly motivated to begin with.
3) What excites you the most about how technology is being used today to influence healthy behavior? And, where is it failing?
I think technology possess great potential to help us change behavior. One of the main reasons is that we could not measure these behaviors up until [roughly] 5 or 10 years ago. We didn’t know how many steps people took, we didn’t know if they took their medication (we can now with connected pill bottles), weight measurements were self-reported and often inaccurate. Technology has given us the opportunity to passively monitor, and we can now do that at a large scale. We can measure thousands of people with very low manpower because it can all be automated through technology. The greatest promise of technology is being able to, on a large scale, automate this idea of passively hovering and get a rich data set so that we can see what is working (and what is not). Furthermore, we can do this while the only expectation for the participant is to continue doing what they are doing, which if you think about it is a big deal.
Where technology is failing is we have not taken the step beyond measuring. How do we actually get people to change their behavior using technology? I call this the “technology delusion.” People sometimes think that you can take someone who is overweight — who is inactive — give them a wearable device and all of a sudden they are going to be a new person. This might work for me or you who are engaged with this stuff, or Quantified Selfers, but it will not be true for people that have an inherent lack of motivation. These devices have not been shown to increase motivation in at-risk populations. That is why the studies I am a part of couple a behavioral change strategy with a technology. The technology is good for recording, maybe helping with feedback loops, but the behavior change component is what is often missing from organizational workplace wellness strategies.
4) There is research to suggest that extrinsic rewards are episodic, and in some cases extrinsic rewards can alter motivation in ways that are counterproductive. Most of this research is based on carrots (incentives) opposed to sticks (penalties), does using the fear of loss mitigate any of the risks generally associated with extrinsic motivation? Besides proving to be more effective, are there other attributes to penalties that position it as a better choice than rewards?
Intrinsic motivation is of course desired, if we can get people to increase that kind of motivation it is where we would start. The problem is it is fairly hard to influence intrinsic motivation, and then sustain that increase. The person really needs a good reason, many times that reason relates to a family event, or a life-changing event; whatever it is, the intrinsic motivation has to come from within the individual.
Extrinsic motivation, giving somebody some type of reward, is generally meant to jump start new habits and then hopefully we can remove the extrinsic motivators. There are some that believe you have to leave the reward in place to see sustainable behavior change. We have found evidence that people who get extrinsic motivation that’s well-designed get better results than our control groups. Furthermore, in some instances we have removed extrinsic motivation and we don’t really see that those people do worse than the control group either. We performed one study where we positioned the reward as a loss, allocated the money up front, and then took it away if the participant did not meet their goal. What is important here is that the lever was not a penalty — no one lost money out of their pockets. So this was not a stick per se but more like a “frozen carrot.” We told all three groups in that study at the end of the month they will get a check in the mail, and they could earn about $42 (a month). The reward was the same among the two non-control groups, but for one group the incentive was framed you get something for your behavior, the other group it was framed you start with a reward but it can be taken away. What was nice about that was it was a reward kind of masked as a penalty, and it made people feel like the money was theirs, a concept called the endowment effect. We find time and time again when people have skin in the game they are more likely to change their behavior.
5) Addressing the potential negative aspects of penalties, how do you coalesce your findings of successfully using the fear of loss to elicit behavior change, with the ethical notion that people should not be (or at least feel) penalized for personal choice?
Certainly there are ethical things to think about when one group is going to get something and another group is not. Those concerns should be discussed and addressed. One way to determine if the reward is causing harm is asking the question, “Do people disengage?” People are generally concerned about framing a reward as a loss, the belief being a group (subjected to the loss) is not going to like it or consider it punitive. We found in our study that even with a frozen carrot, 96 percent of people finished the study and stayed actively involved even 3 months after we turned off the incentive. This engagement is much higher than you would see in many wellness programs currently in use. If the incentive was perceived in a way so punitive that it made participants drop out that might give us pause. However, because of the success of the study it makes us believe that this method is scalable. I am not saying it will be for everybody. We still need a way to make these incentives more personalized. Some people will respond better to losses, some to gains. What we learned at the population level is it appears more respond more favorably to losses, but at the individual level a patient-centered approach will help us further by identifying the right incentive for a particular person, which in turn will increase efficacy.