Study says: Don’t buy a fitness tracker, just use your phone

Pretty much every fitness tracker on the market does the same thing: Using an accelerometer, it tracks how many steps you’ve taken, and from that accelerometer data, usually can extrapolate distance traveled as well as calories burned. Of course, your smartphone has an accelerometer, so why do you need a Fitbit, or a Jawbone Up 24, or a Misfit Shine? According to a new research letter published in the Journal of the American Medical Association, you don’t — smartphones can track steps just as well as a dedicated fitness tracker.

The study looked at 10 different trackers — four smartphone apps and six fitness trackers, including the Moves app (now owned by Facebook) for iOS and Android, multiple Fitbit models and the Nike Fuelband. The researchers — some of whom are still medical students — strapped several trackers to the subjects, who then walked on a treadmill for 500 steps, and then 1,500 steps, twice. Ultimately, the 14 participants in the study ran 56 trials (four treadmill runs each) meaning that there were 560 gadget step-reading data points.

The study found that phones, either running the Moves app, Fitbit app, or the Health Mate app, were as accurate as the dedicated step tracking hardware, and most of the trackers were within 10 percent of each other — except for the now almost-discontinued Nike Fuelband, which recorded steps that were over 20 percent lower than the observed steps and other devices.

One interesting tidbit from the study: In eight of the 560 device trials, the gadget wasn’t properly configured to record steps, which lines up with my personal experience that your step tracker will not be working around one percent of the time. There’s also a chance that the study’s findings could be affected by configuration settings — for instance, Fitbit’s option to tell it that you’re wearing it on your dominant wrist.

This study’s not going to be the be-all and end-all for step tracking accuracy. In fact, this study observing 10 women and four men recruited at a college isn’t all that different from certain anecdotal evidence, like this informal experiment conducted last year by science journalist Rachel Feltman. In my experience, most wearable tracker manufacturers know there’s a roughly 10 percent difference between various step readings, which is more than accurate enough for early adopters and techies. But this study underscores the fact that as wearable devices and step trackers infiltrate the healthcare system, more academic research will be required.