7 Reasons that Fitbit Data in Court Should Terrify You (Part 1)

fitbit dataAs originally reported in The Atlantic, the first case using Fitbit data as evidence is currently underway in Canada. The plaintiff in a personal injury lawsuit, a personal trainer, wants to use Fitbit data to support her case. The plaintiff believes that the Fitbit data shows that her physical activity level is lower than baseline for her age and profession.

However, instead of using the raw data, plaintiff has retained a data analytics company called Vivametrica to provide expert analysis. Using data collected from “industry and public research,” Vivametrica claims to be able to compare one person’s Fitbit data to the general population. It’s potentially game-changing evidence. It should be great. It’s not.

Fitbit is a neat little device for tracking your activity. Importantly, Fitbit (and similar devices) creates, and stores, data. A lot of data. About you. And that data is a tempting target in litigation. But be warned, the implications of using Fitbit data in court should terrify you!

1) There is No Uniform Standard.

Fitbit is merely one of many available wearable health stat devices. Jawbone has the UP, Nike has the Fuelband. None of them track exactly the same data. While the Garmin Vivosmart measures your sleep, the Jawbone UP24 claims to recognize between light sleep and deep sleep.

fitbit data

Which way do you think we’re going? Image by Stuart Miles

Some track calories burned, but do so without tracking your heart rate. Some track stairs, some just track elevation change, and some don’t track vertical change at all. While that may not matter much in Chicago or New York, imagine walking around San Francisco and not taking hills into consideration. Do you know which one you use?

It’s coming into evidence anyway. Terrified?

You’ll need to know the difference. For your cross examination. To explain to the jury. It’s even an ethical obligation! But knowing everything may not help, because…

fitbit data2) Fitbit Data Is Easily Manipulated.

People can be competitive. And bastards. People can be competitive bastards. When winning is at stake, the most competitive look to gain any advantage. (See ALL OF BASEBALL… and every other damn sport).

Even when there’s nothing at stake other than bragging rights, we’ll look to cheat. We even do it when we’re only competing with ourselves! (And since Fitbit is used to potentially lower employee contributions for health insurance, there are tangible benefits to cheating!)

And for people looking to cheat, there are tools like this one! It’s a downloadable package with instructions for manipulating your Fitbit data. It’s even ranked on the first page of results for a Google search for “manipulating a fitbit.”

fitbit data

For a low-tech cheat, according to this blog, just wear spandex.

fitbit data

The power of Spandex

Beyond manipulating data, let’s not forget that these are wearable devices. They’re not chips implanted under the skin. So how can you guarantee that what is recorded represents that person’s activity? Did they take off the wearable at any point? Did someone else wear them? Did they flap their arms to mimic exercise? Or do they have the wearable that interprets the act of writing as strenuous physical activity?

It’s coming into evidence anyway. Terrified?

The data you are relying on or hoping to challenge may have been altered. Dramatically. And you’ll never prove it. Unless it doesn’t make sense, which would be great if not for…

fitbit data3) People are Snowflakes, and so are Vital Signs.

So the plaintiff in the Canadian lawsuit wants to use her Fitbit data to demonstrate that her physical activity has been below normal since the injury. And a data analytics company has told her it will. Even transactional attorneys should be able to spot some pretty significant holes in this one!

Vital signs are unique to the person, and are not easily compared to others. For example, the average normal heart rate of a 15-20 year old ranges from 60-100 beats per minute. Adults range from 50-80 beats per minute. Who’s to say whether your number is truly more or less than “average”?

It’s coming into evidence anyway. Terrified?

Even if it’s your evidence, you’ll never really know what the data means. But certainty was never required under Daubert! Speaking of uncertainty…

4) Human Memory vs. Glitch-filled Algorithm… which sucks worse?

“Ultimately, the Fitbit case may be just one step in a much bigger shift toward a data-driven regime of “truth.” Prioritizing data—irregular, unreliable data—over human reporting, means putting power in the hands of an algorithm.”

As The Atlantic article discusses, the decision to use a Fitbit creates a digital record of a person’s day. However, there’s no guarantee that the data stream created will be the same as the experience of the person wearing it. The divergence between the two sets of information could be the result of many factors – including the weakness of human memory and the problems with imperfect data tracking.

When a jury is left to evaluate the divergence, what analysis is proper? We all know that juries place far too much faith in human memory, which is generally a less-reliable form of evidence. But should the jury instead believe the GPS system that tracked this person’s supposed stroll right to the middle of the River Thames?

Who is correct? Here’s where you’re really going to get into trouble, because this is entirely a credibility battle!

It’s coming into evidence anyway. Terrified?

Which one is right? How do you know? Figuring it out will probably cost you even more money! Maybe if it’s unreliable, you can keep it out! Except…

See part 2 here.