Sunday, January 8, 2017

My data, myself: My fitness tracker helps motivate me to exercise, but clinical research suggests your mileage may vary

It is the beginning of a new year, which for me means calculating my yearly mileage and thinking about next year's goals. I have been obsessed with tracking my performance in fitness since I started running in 2000. It started in a very low-tech way, writing down my mileage each day in a calendar and then adding it up each month and then at the end of the year. With the advent of the app MapMyRun, I became even more enthusiastic about tracking my running, which helped me to run more consistently.

Last Christmas, I got a new Garmin VivoActive smart watch, which has really changed the way I look at my activity levels. Like most smart watches, this one can track my exercise, steps, and sleep cycles. It also buzzes me if I have been inactive for too long. This is particularly useful when you have an office job. In a completely unsurprising Pavlovian response, I have now started to anticipate when the watch will buzz, so I get up and take a few laps around the office until the "move bar cleared" buzz comes in. The watch also gives me a little fireworks display when I reach my daily step goal (or multiples thereof, which are particularly edifying). When I sync my watch, the Garmin website will analyze the data, so I can see how I am doing for the week, month, or year. To some people, this may sound exhausting, but as a runner, bike commuter, and science nerd, I love all the charts and graphs.
2016 distance totals
I always thought that my obsession with my fitness data was some vestigial interest in analyzing data due to my background in science, but it seems that many people find fitness tracking to be motivating and fun. My data analysis has been fairly simple; I have only looked at trends over time (e.g., my running increased during the stress of grad school and decreased around the time of my pregnancy). It's amazing to see how other people have quantified their lives so thoroughly. Quantified Self has some great examples of people data mining their own lives to understand and improve themselves, one woman has analyzed more than two decades worth of data about her fitness, diet, and weight.
Based on my experience, I would think that fitness tracking apps and wearables can help people to stay motivated. However, the results of some of the published clinical trials suggest that it isn't clear if fitness trackers can help everyone achieve better health (as measured by activity levels, weight, and metabolic variables). There are currently 21 clinical trials using fitness trackers registered on Clinical Trials.gov, so more information about their effectiveness should be available soon. In September 2016, a study published in The Journal of the American Medical Association received a lot of buzz for reporting the results of a trial that compared the weight loss with the standard regimen (i.e., calorie restriction and physical activity) versus the standard regimen plus fitness tracking. They found that the fitness tracking group lost less weight than the control group, suggesting that wearables may not help people lose more weight, especially in the long term (excellent coverage of this study in STAT). There are some variables that can make activity trackers improve fitness; one study showed that cash incentives led to some increase in activity levels. However, this increased activity did not continue once the cash rewards were gone. An earlier study, sponsored by Fitbit and using an opt in approach to recruit people for the wearable group, found that people who use Fitbits for more than a year have larger decreases in insurance costs than non-users. Another randomized control study in postmenopausal women confirmed that the use of the Fitbit increased activity levels, but they did not collect data on weight loss or fitness levels. A major problem with fitness trackers is abandonment: about a third of people stop wearing their fitness tracker after 6 months.
Distance totals 1999-2016

These results suggest that, for most people, wearable devices will not encourage them to exercise or lose weight. Indeed, I would expect that it would be a very particular personality type that would be motivated by a watch's fireworks display or an online competition. It remains to be seen if people who already have a strong exercise pattern are more motivated by fitness apps and trackers. The challenge for fitness tracker designers and doctors alike is how to motivate people to make changes for their health.

Perhaps more interesting is the potential for the usage of wearables in the future. According to Kat Arney in Herding Hemingway's Cats and here in shorter form for BBC science, the data from our smart watches could ultimately be combined with the data from our personal genome sequences: "Combining the power of modern genetic analysis with bio-monitoring" could improve care, save lives, and revolutionize genetics. Larry Smarr, profiled in The Guardian, has been tracking his life for 15 years following more than 150 variables. Such complex bio-monitoring generates a lot of data and turning that data into usable information is the challenge. This is one reason why tech companies are getting into the health data market. A recent editorial from Nature warns that there could be unforeseen problems with sharing our health data, particularly in terms of personal privacy. In addition, they speculate that it could widen existing inequities and biases. In addition, it is not yet clear what benefit (if any) these data-driven approaches will have.

I don't expect to abandon my GPS watch anytime soon. Instead, I imagine the future me using a multi-function tracker that combines data from my metabolic profile with my genomic data to make recommendations on exercise, fitness goals, and vitamins. I may be living in a science fiction fantasy, but 15 years ago when I started tracking my miles with gmaps pedometer and a calendar, I would never think I would have a watch that could do all that for me.

No comments:

Post a Comment