Why Wellness Programs Don’t Work So Well
Many companies offer wellness programs in a bid to improve the health of their employees and control the costs of providing medical insurance. Employees check in regularly with a nurse or health coach, participate in a yearly wellness fair to track their weight, blood pressure and cholesterol levels, or are given incentives to walk 10,000 steps a day. But do these programs really work in changing behaviors?
New research from the Perelman School of Medicine at the University of Pennsylvania reveals a fundamental flaw in the design of such programs — they’re generalized plans that don’t take into account some people need more help than others. Also, they assume people always act rationally on the information they get. It’s not that simple.
Joining the Knowledge@Wharton show, which airs on SiriusXM channel 111, to talk about their research findings are two doctors: David Asch, a professor of medicine who is also a Wharton health care management professor, and Shreya Kangovi, professor of medicine at Penn who is also a senior fellow at the Leonard Davis Institute of Health Economics. Their findings are also detailed in the recent paper, “Behavioral Phenotyping in Health Promotion Embracing or Avoiding Failure,” which appeared in the Journal of the American Medical Association.
An edited transcript of the conversation follows.
Knowledge@Wharton: What motivated this research?
Shreya Kangovi: This was less of an empirical research study and more of a thought exercise and observation. The underlying motivation was drawn from other studies that both David and I have done. We focus our research on trying to promote healthy behaviors because that really matters for health outcomes. There’s a lot you can do that is effective. But even when you have a really proven type of health behavior intervention, there are folks who don’t respond. We asked ourselves, who are these individuals and what might be happening here?
A lot of these health behavior promotion programs are based on this premise that information is power. We give patients signals of how they’re doing on health behavior change, whether it’s asking them to weigh themselves all the time or check their sugars. We just assume that people are going to see their weight is going up or their blood pressure is high, and they’re going to do something about it. But we challenged that assumption when we realized from our observations this wasn’t always the case.
Knowledge@Wharton: This is an important area of research because so many companies now have these wellness programs?
David Asch: That’s right. They are new, they are pervasive and they’re extremely well-meaning. Employers fundamentally want their employees to be healthier — frankly, people want to be healthier — so it’s perfectly natural for these kinds of programs to exist. Typically, most of us spend a lot more time with our employer than in other settings, so these are often the best settings in which to think about encouraging healthy behaviors.
The challenge is that a lot of these programs are designed with the idea that we’re perfectly rational people. That a little bit of feedback — you could lose a few pounds or you should eat a little bit less — is naturally going to fall upon a rational human being who’s going to say, “You know, you’re right.” The trouble is, I don’t necessarily want to get on the scale in the morning and get that feedback. Sometimes that feedback isn’t so helpful. Sometimes it’s a little aversive.
Knowledge@Wharton: These programs are designed to help employees feel better, but they also help companies lower healthcare costs.
Asch: It’s not obvious that these programs will do a lot for the bottom line of companies. They might. But it’s less likely that they’ll do that than promote a healthy workplace, which may provide benefits to employers in other ways. Of course, employers want to be the employers of choice and want to attract healthy employees. Even if they don’t save money, and they may not, they could be good for society to the extent that these programs work.
But I think the challenge is that their effectiveness might be limited by the way they’re structured. As Dr. Kangovi pointed out, it’s possible that they could be structured a little bit better so that they don’t rely so much on the feedback that may work well for some people but not well for all people.
Knowledge@Wharton: Based on your research, how should companies tweak these programs?
Kangovi: What we need to look at in these trials is a simple question, who benefits and who doesn’t, rather than just assuming that everybody benefits. We need to figure out who are the people that really are motivated by this feedback and have that rational response of trying to regulate their behavior, and who are the folks and under what circumstances do people become discouraged by that type of feedback? Does depression, for example, predispose you to becoming avoidant and not wanting to step on a scale when you know that you’re going to see a number that you don’t like? Are there other factors involved? We need to figure out who benefits from feedback so that we can start to tailor these programs for the individuals who are most likely to benefit. I think that’s the key thing.
The other piece is that we all can become avoidant under the right circumstances. If we’ve been on an eating binge, we just don’t want to step on that scale because we know that the number is not going to be pretty. You have to more broadly help employees cope with failure because that’s such an inevitable part of any type of health behavior change. You’re going to step on a scale and see a number you don’t like. You’re going to check your sugar and see that it’s high. How do you not beat yourself up over that?
There might be a couple of strategies borrowed from psychology that can help people cope and be resilient against their own failures. One is called positive affect indication. It’s really just simply random acts of kindness, small compliments or self-affirmation that can help boost your mood after you step on that scale and see a crummy number. It’s just telling you that you’re not a bad person just because your weight is a little higher than you thought it would be. Think about some other areas in which you’re proud of yourself or have had success.
The other piece is something called attribution retraining. That is teaching people to view their failures as controllable rather than uncontrollable. In education, for example, students will fail tests and say, “I’m just stupid.” That’s a really hard thing to overcome. Attribution retraining coaches students to think, “It’s not that you’re stupid, it’s that you didn’t study because you went out on a date that night.” Breaking failure down into these concrete things that are controllable is a way of bolstering people against the inevitable challenges they’re going to face and helping them get back on the horse.
Asch: That’s why this is so exciting, because we could look at the distribution of success or failure with these programs and throw our hands up, or we can recognize that we’re probably reflecting different kinds of traits and states that people are in and then use that to tailor the appropriate intervention.
We’re in the age of precision medicine. When you think about cancer treatment, people don’t have just lung cancer, they have different kinds of lung cancer. Now that we know that, we tailor different therapies to different kinds of lung cancer, so that information is incredibly valuable. I think the same thing is true here with human behavior. It was probably silly of us to think that a one-size-fits-all program could ever work. There’s something optimistic and progressive about this.
Knowledge@Wharton: Why have we fallen into this one-size-fits-all mentality?
Kangovi: Everybody prefers to think simplistically. It’s easier to hope that there’s a one-size-fits-all and walk away from it. The other thing that’s even maybe more interesting is sometimes we do a little bit of blaming the victim. The people that it doesn’t work for, we might say they’re just lazy and they don’t care. We’re learning from our research that the people who fail maybe care too much. Maybe they’re just so hard on themselves that they don’t want to see those bad numbers, and maybe we’re making it worse by forcing this feedback and information down their throats. What if we were a little bit kinder and gentler to those folks and didn’t blame them? I think we could get better results across the board.
Asch: This discussion, in some respects, is two steps away from where most businesses currently are. Most health promotion programs from any source are fundamentally based on the idea that people are perfectly rational actors. “I just have to educate you that there are 600 calories in that bran muffin and then you are less likely to eat that bran muffin.” Or “I just need to tell you that smoking is dangerous. Of course, you’ll quit smoking.” No one is against the idea of educating people about risks or health-promoting activities, it’s just that in most cases we already pretty much know. Yet these programs that we see are typically designed around the idea of educating people or encouraging them in very rational ways.
The field of behavioral economics, which is incredibly strong here at Wharton and at Penn, recognizes that people are not so rational most of the time. They’re often irrational. The advantage of behavioral economics is that we are irrational in highly predictable ways. It’s that predictability that helps us. The work that Shreya is describing takes it one step further, which is some of us respond to different interventions for our irrationality than others. So, it’s two steps removed from the typical program, which is we are all the same, we are all rational, and therefore an education program works. That’s way too simplistic.
Knowledge@Wharton: Monitoring takes two different forms: You may have to check in with a medical professional who may be linked to the company, or you can self-monitor. How does monitoring work in these programs in general?
Kangovi: There are two questions. First, at a high level, what is monitoring doing here versus just other types of health promotion? And what’s the difference between self and other monitoring?
Any time you have a health promotion campaign, there’s going to be fall off. If the campaign is to walk 1,000 steps a day, you’re going to see 80% of people doing it the first week, and then 50%. What we’re pointing out is that the monitoring itself might be chasing people away. It’s not just the usual loss of interest, it’s that there’s a group of people that might have kept walking if they didn’t have a signal in their face all the time or if they weren’t worried about a signal telling them that they were failing.
Because for a certain group of people, that signal is so demoralizing. That’s really the striking part here. Here we are thinking that, again, information is power, let’s tell people stuff, it’s definitely got to help. Maybe there’s a subgroup for whom it’s actually harming them. That’s one piece of why the monitoring part is really important.
We don’t know that much about the differences in this response to failure with self-monitoring versus other monitoring. You do have to assume that any kind of avoidance that you feel when you have a bad result is going to be amplified if you have to share that with your nurse, your doctor or your community health worker or in any kind of public forum. There’s some evidence to support that.
Knowledge@Wharton: How can we get companies to use this research to re-evaluate their health programs?
Asch: There’s nothing so compelling as a really good idea. These programs are built on a good idea that people should know their numbers and move forward. That’s the sort of headwind or ballast that keeps employers from taking a critical look at the effectiveness of these programs or the fact that, when they are effective, those positive results may be distributed highly unevenly across their workforce.
The first step is to get people to be critical and examine what’s actually happening as opposed to feeling good about the appealing notion of a workplace wellness program. I’m enthusiastic about workplace wellness programs if they’re designed carefully and with some thought to the way real people make real decisions. But I’m not so positive about them as they’re currently designed. I would never buy a scale that would sit in my bedroom and tell me my weight vocally. “You weigh 200 pounds.” I don’t think that would be something I’d want my wife to overhear, particularly since I’m 5 feet 5 inches. If we always think that everyone should know your numbers and somehow that will be motivating, that assumes that we are a kind of animal that isn’t very human, to be honest.
Kangovi: First, just acknowledge that we’re poking a hole in a pretty pervasive dogma here. We’re in the age of knowing your numbers, whether it’s employee wellness or Fitbit or all of the digital devices that can tell us every single thing that’s going on. What we’re seeing is that those efforts really work for some people, but they tend to be the ones who might be healthier in the first place. What I’d like to see happen is that we can continue to use those, but when people fall off or fail those efforts, we don’t just shun them or put them to the side. We engage them.
If I’m running a company and I have everybody with a pedometer for a month, and 50% are crushing it and the other 50% stop walking or stop using the pedometer, let’s get a focus group together. Let’s figure out what happened and maybe put in some other strategies like coaching or the random acts of kindness so that if you don’t check your pedometer one day, you get a text that says, “It’s OK that you didn’t make your steps today. Tomorrow is another day.”
Knowledge@Wharton: Can technology assist in this process if the approach is right?
Kangovi: Technology is what we make of it and what we make it to be. We need to develop a little bit more science and understanding of failure and health behavior change. Then maybe we can get technology to assist us by doing some of the tailoring work that we just talked about.
Asch: This is a great point, and I’m very optimistic about technology. But I want to see it as an enabler. It’s not a leader. Fitbits don’t make you walk more, but they can enable a lot of other strategies. When the early self-monitoring technologies came out, they got picked up by the quantified selfers. These are the people who log every single step they take and photograph every meal they’ve eaten. Those people have a totally different disease; they have some kind of obsession.
It was natural but flawed for people to think we just need to give everyone Fitbits and turn them into quantified selfers. But you’re not going to turn couch potatoes into quantified selfers. They’re fundamentally different creatures. How could we use technology to enable the people who wouldn’t have been intrinsically motivated to do that in the first place? That is the challenge.
Knowledge@Wharton: But what about the people who can’t afford technology?
Asch: You’re totally right. It’s another example of the digital divide. While it is true that technology is getting cheaper and much more ubiquitous, it will always be the case that the more resourced people are going to be able to be ahead of the curve. To the extent that we think that these are fundamentally important principles, then you would think we need to find ways to distribute this technology.
We’re still ahead of the game because I have yet to see really well-designed programs that use technology to move things forward in effective ways. If we got there and found those things and found that we simply couldn’t distribute them to broad populations, then we’d think this is an agenda for some kind of social recovery.
Kangovi: I totally agree. There’s still a role for human touch in all of this. Most of my research focuses on the use of community health workers. These are trusted laypeople who typically come from within low-income communities and who are making initial contact with patients or employees and helping them think broadly about health behavior change. I think particularly for marginalized patients or those who don’t have the resources, that human touch can still be really important.