Most people know they should exercise more, but discovering the right way to motivate them is difficult. An interdisciplinary team of Stanford researchers has taken on that challenge and in the course of their work has conducted the largest ever assessment of physical activity.
“Our goal is to motivate physical activity on a global scale. And the team is producing tools and knowledge that can be deployed at a low cost on a planetary scale to improve both mental health and physical health of millions of individuals,” says Scott Delp, a professor of mechanical engineering and of bioengineering at Stanford, who spoke about the research at the university’s recent Reunion Homecoming Weekend festivities.
Working with a mobile app company, the team gathered minute-by-minute step counts for 2 million people in 100 countries. An important finding of the survey is that there is inequality in activity across populations. When activity inequality is high, women tend to be less active, a condition that shortens their lives, Delp says.
Delp notes that there is no shortage of devices one can use to measure physical activity, but most of the “nudging” by those devices simply doesn’t work. “In most cases, people use physical activity monitors over the short term and then throw them in a drawer,” he says.
One of the team’s key findings came from a study led by psychology professor Alia Crum, who revealed a placebo effect in exercise physiology. A group of subjects took a test that monitors how much oxygen an individual takes in and how much carbon dioxide they put out during intensive exercise. Blood was drawn after the exercise ended and the subjects were asked to return a week later.
When they returned, some of the subjects were told the blood test showed they had a gene that is excellent for fitness; the rest were told they didn’t have it. The subjects who were told they didn’t have the gene performed significantly worse when they redid the exercise test. Those who were told they did have the gene performed at the same level as earlier. “What this demonstrates is that the way you deliver information can have a powerful effect, not just on people’s behavior but on how their bodies respond to physical activity,” Crum says.
Other experiments have shown that personal interventions, even when conducted by a virtual advisor, resulted in subjects adding more than four hours of exercise a week. Also motivating: working out with friends, even online friends.
The researchers plan to share their findings with developers who can then incorporate them into commercial devices and applications. “Methods that are validated scientifically can provide a new paradigm for low-cost medicine that can be disseminated worldwide to improve physical and mental health,” Delp says.
The research project is funded by the Stanford Catalyst for Collaborative Solutions, an initiative launched in 2016 to inspire campus-wide collaborations to tackle some of the world’s most urgent challenges.
Mark Horowitz: This session will be given by Scott Delp, and Scott’s been leading a program about motivating mobility, and what he means by this, he’ll explain more, but we all know exercise is good for us, but we all don’t do it as much as we should, and so his project is trying to deal with that conundrum. Scott.
Scott Delp: So our goal is to motivate physical activity on a global scale, and the team is producing tools and knowledge that can be deployed at a low cost on a planetary scale to improve both mental health and physical health of millions of individuals.
Like Mark said, we know that exercise is good for us, but in the U.S. alone approximately 260,000,000 people do not get sufficient physical activity to maintain their bodies and their minds. That’s 78% of the population according to the Center for Disease Control.
Now the cost of this inactivity is staggering. In a recent article, Min Lee and her colleagues, reported that 5.3 million deaths per year can be attributed to physical inactivity. 5.3 million deaths per year. So, you can make a calculation like that because 18 major diseases and disorders are linked to physical activity. There’s strong correlation between physical activity and things we might consider, cardiovascular disease, diabetes, yes that’s true, but also six types of cancer are related to physical activity as is anxiety, depression, and cognitive decline. That’s the bad news.
The good news is that physical activity is potent medicine. One of our team members, Abby King, chaired the national committee on physical activity that just produced this fantastic report that summarizes all of the evidence that’s been gathered over the last 40 years, and it makes very clear the potent effect of physical activity, and what I wanted to do is just read you a single quote from this report. So here’s the quote, “A single bout of moderate “to vigorous, physical activity will reduce blood pressure, “improve sleep, reduce anxiety, and improve cognition.” You can reduce your anxiety and improve your thought clarity with a single bout of physical activity. Going for a bike ride, playing a game of tennis, that’s awesome! That’s incredibly motivating.
The other thing to know is that small doses have a big impact. This doesn’t mean that you need to be running six miles a day, six days a week. If you get up from your desk and you move about, there’s very clear evidence that you begin to accrue benefit. So, how are we doing?
We recently conducted the largest worldwide assessment of physical activity that has been performed. We worked with a local mobile app company, and what we gathered was minute-by-minute step counts for two million people in a hundred countries. This was the biggest survey of physical activity by a factor of a thousand. We were able to characterize planetary scale trends in physical activity.
One of the things that we found was that there’s inequality in activity across a population. Just like there’s income inequality, there’s rich people and poor people, there’s physical inactivity inequality, and when their inactivity inequality is high, women are differentially disadvantaged. There’s more activity poor women, and if you relate that to lifespan when there are activity poor women, it costs women in terms of life years. So, again, it’s a big problem.
The other thing we discovered is that if you want to intervene, the place to have the biggest impact, is where there’s activity poor individuals. Now there’s no shortage of devices to measure your physical activity, but unfortunately, most if not all, activity interventions are failing and they fail in part because they provide lots of information, but simply providing the information doesn’t help you move more. If you survey people about, “Well, why can’t you get more activity?” and people, no one ever responds, “Well, it’s because I don’t know exactly precisely how many steps I take each day.” And in fact, informing people of these quantitative metrics, for some people, can produce a mindset of inadequacy. There’ll frequently be nudges that might come during a meeting or an inopportune time, and so soon you begin to ignore them. In most cases, people use physical activity monitors over the short term and then throw ‘em in the drawer, and that’s in part because it’s a one size fits all. We’re motivated by different things. We respond to different things, and there hasn’t been good personalization.
Our team is developing scientific evidence and is based on evidence to replace information overloads with an understanding of mindsets that is how you think about the world influencing your behavior and your physiology. To replace constant nudging with well-validated behavior change theory.
We know about behavior change. There’s decades of research and almost none of that has been incorporated into current activity monitoring. To achieve sustained engagement by storytelling. People love stories. Engaging in a story as you move through the story by being active is one of the approaches we’re taking, and also data-driven personalization. So, we’re bringing these approaches together in our catalyst project, and what I wanted to do is just give you a little bit of insight into the evidence that we’ve developed so far in the first year of the project.
First about mindsets. What are mindsets? They’re lenses through which we see the world, and they influence your attention, your motivation, and amazingly, your physiology. The placebo is a great example. You can take a sugar pill and if you think it’s gonna get you better, in many instances, there’s profound healing effect from taking essentially nothing other than a placebo. This is shown across almost all fields of medicine, the profound placebo effect. So no particular medicine has as strong a physiological response. We wanted to investigate how getting genetic information influences your cardiovascular fitness.
Brad Turnwald, who was a student in Ali’s lab worked with folks in my lab where we brought them in, we did a VO2 max test. We have them run on a treadmill to exhaustion, and while they do that we measure their ventilatory gases. How much oxygen they take, how much CO2 they put out. They came back a week later, and we done a blood draw, and we randomized them into telling them whether they had excellent genes for exercise or not.
We then ran the same test, and if we told them they were excellent, they didn’t change. If we told them they had the gene that was not excellent, their capacity was statistically significant and significantly reduced. So, they were working just as hard, but things changed that you cannot control it. Your changed voluntary oxygen to CO2 and things that you have no voluntary control over. What this demonstrates is that the way you deliver information can have a powerful effect not just on people’s behavior, but on how their body’s respond to particular physical activity, and we’re wanna use the information to design messaging that’s personalized that affects mindset in a positive way.
We also need to get lasting behavior change. Getting behavior change for a week or a month, or even three months doesn’t work, and Abby King has had good success with long-term behavior change with implementing theory. This is not a fancy app. This is kind of an old school, sit at a computer and have Carmen talk to you and she did an intervention, Abby did an intervention in a community center that served low-income, Latino, older adults, and the average session lasted six minutes. They came back for four months, and this was the result they got. This is the change in minutes of physical activity per week. So, there was a control group that didn’t change from zero, and Carmen, this virtual advisor, was able to get 250 minutes of additional activity. Four hours of additional activity per week. That is a ton of additional, this is the biggest change I’ve ever seen! If you can get four more hours of physical activity per week, you know, half an hour a day, eight days a week, that’s doing really well. So, we really think that implementing this behavior change theory can have a big impact.
James Landay is on the team and he’s working with Paula Moya from the English department on developing narratives to promote long-term engagement. What James has been able to show is that simply state-of-the-art detection and feedback is not enough, that simple stories are compelling, and that the multi-chapter narratives where you engage people over the long-term and the story advances by you engaging in physical activity can potentially change long-term behavior.
Jure Leskovec is an expert in artificial intelligence and personalization. He developed the Friend Finder algorithm on Facebook. This is Jure now. He must’ve developed that algorithm when he was about 12. He also does personalization for Pinterest to get you to figure out how to click on various ads and engage over the long-term there. Here he’s applying his expertise to examine how to personalize feedback to users of health applications to engage them over long-term dynamics, and one of the things that we care about is whether you have a social network and you know your friends can help you exercise, but can virtual friends help you exercise?
The answer is yes. What’s plotted here is the average daily steps and you see when the experiment starts that they go up, and they go up in both the blue and the red. The blue is a group of individuals who were motivated, asked a friend to join them, the friend didn’t join them, but they still improved. The red is the additional boost that you get from a friend joining you. This is a single friend and you see that you get a boost in activity for a period of three to four months. That’s pretty good. When you add a second friend, they don’t do quite as much and it dies off a little earlier and of course it varies from person to person. Interestingly, if you do competitions and you add friend groups, the groups that do best are mixed gender, 50/50 men and women, and also if the groups are matched. If you get people who are all performing at a high level, they compete with each other, or groups that are really trying to build up, they compete, but if you have a mismatched group, it doesn’t work very well.
These are powerful approaches that we’re applying to a range of populations, older adults, people with pre-diabetes, individuals with arthritis, and also with cardiovascular disease. Because we can bring in people from English, Computer Science, Psychology, Bioengineering, like myself, together with people in the medical school, we can intervene on these populations.
Our impact is to provide fundamental insights that we can disseminate and also the app developers out there in the world would make our information freely available and can be incorporated into commercial applications. Methods that are validated scientifically and really provide a new paradigm for low-cost medicine that can be disseminated worldwide to improve physical and mental health.
Thank you very much.