Slashdot Mirror
Class Action Lawsuit Filed Against Fitbit For 'Highly Inaccurate' Heart Rate Trackers (nbcnews.com)
An anonymous reader quotes a report from NBC News: A class action lawsuit against Fitbit may have grown teeth following the release of a new study that claims the company's popular heart rate trackers are "highly inaccurate." Researchers at the California State Polytechnic University, Pomona tested the heart rates of 43 healthy adults with Fitbit's PurePulse heart rate monitors, using the company's Surge watches and Charge HR bands on each wrist. Subjects were then hooked up to a BioHarness device that produced an electrocardiogram (ECG), to record the heart's rhythm against the data being produced by Fitbit's devices. Comparative results from rest and exercise -- including jump rope, treadmills, outdoor jogging and stair climbing -- showed that the Fitbit devices miscalculated heart rates by up to 20 beats per minute on average during more intensive workouts. The study was commissioned by the Lieff Cabraser, the law firm behind the class action suit that is taking aim at three Fitbit models that use the PurePulse heart monitor, including the Fitbit Blaze, Fitbit Charge HR and Fitbit Surge. "What the plaintiffs' attorneys call a 'study' is biased, baseless, and nothing more than an attempt to extract a payout from Fitbit. It lacks scientific rigor and is the product of flawed methodology," Fitbit said in a statement posted by Gizmodo.
Wow...only a day or so after that study was released and the lawyer vultures are already circling.... smh...
You're messin' with my Zen Thing, man.....
Overpriced fad gadgets turn out to be crap - film at 11.
I was not sad to see fools parted with their money here except when employers started buying into them and tying health care costs to inaccurate electronic tethers. After a couple of those experiences, I know the technical groups I worked in learned how inaccurate these things were, or gamed the telemetry systems behind them (hint: their web services are usually crap/hackable) so they could avoid health care price increases without actually taking the 3000 steps, checking their pulse, or playing whatever other nanny games the system wanted that month.
I'm skeptical to think this will work to well. The fitbit isn't classified as a medical device but as a health monitoring device. Without any sort of certifications it tends to mean they legally have to be fairly scant on their claims for fear of the FDA cracking down on them, so any lawsuits based on their accuracy are going to have a hell of a time. It'll all come down to how sympathetic of a jury they get is.
These devices were designed by Nathan Myhrvold. They are accurate but you have the wrong pulse.
Some drink at the fountain of knowledge. Others just gargle.
Well, then. I was thinking of getting one to help me get and stay close to max-rate for interval training. 14% inaccuracy would be wildly dangerous in that case.
The lawyers may be scum, but I'm glad to have read these articles.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
But not for the burns and rashes? Weird.
Expect the prices of these things to soar as their liability insurance costs far outweigh their costs of production and advertising.
The shepherds did so well protecting the flock that the sheep no longer believed that wolves existed.
Sure if it's a device you're hooked up to in hospital you ideally want an accurate bpm measurement but for something designed for the every day joe to wear while exercising, is high accuracy really that necessary? Just pay attention to your body and not a piece of equipment. If your chest feels like it's about to explode how's about you slow down a bit?
I'm pretty sure people in the past hadn't been blowing their hearts out exercising before heart rate monitors become mainstream.
Its a workout tracker, does it need to be terribly accurate? It should be able to track if your heartbeat is low, average, somewhat exerted, highly exerted, or "CALL AN AMBULANCE". It needs to portray a number for most people to grasp what is going on, but that number is probably only a rough idea and unless FitBit is advertising it as medical grade accurate I don't see an issue as long as it can get you in the basic area of your heartbeat. I see this all the time in my line of work (mapping), people see a line on a map and assume that its certified accurate to a millimeter when in reality it is more like a dash GPS, it'll get you close a majority of the time but don't trust it completely unless you want to eventually end up in a lake.
The fallout from this will be regardless of whether or not the disclaimer was there, but rather NO MEDICAL ANALYSIS can be performed on a non-certified device!! This is just paving the way for future legislation to solidify that notion. What will be interesting - and bring a big fucking bag of popcorn - is how both Google and Apple will team up against Federal regulatory action. You want to talk about political nuclear war, baby, this would be it; right up there with encryption!
Life is not for the lazy.
I have noticed myself while working out and running on a treadmill that the treadmill would often say my heart rate was 15-20 bpm faster on the treadmill then on my Fitbit charge HR. I purposefully would end up checking both at times when my to check my heart rate and was surprised to see it was nearly always off while running. I purchased my Fitbit charge HR to be able to keep myself in my cardio zone while working out. This finding completely supports what I was figuring out that this device should never have been purchased for this reason, although it is advertised for this use!
How does the FitBits' accuracy compare with the many other consumer-grade heart rate monitors on the market? e.g. The ones with a strap you wear around your chest.
If the FitBits do a bad job of measuring heart rate - to the point of being worthless noise - then I agree they ought to be sued for selling a product that doesn't do what it advertises. But I'd like to see a less biased party - such as a fitness magazine, or Consumer Reports - do the testing rather than a lawyer chasing a paycheck.
I was wondering why Garmin (FitBit competitor) was proeminently displaying the disclaimer below on their Web site.
Now I know why..
Activity Tracking Accuracy
Garmin activity trackers are intended to be tools to provide you with information to encourage an active and healthy lifestyle. Garmin activity trackers rely on sensors that track your movement and other metrics. The data and information provided by these devices is intended to be a close estimation of your activity and metrics tracked, but may not be completely accurate, including step, sleep, distance, heart rate and calorie data. Garmin activity trackers are not medical devices, and the data provided by them is not intended to be utilized for medical purposes and is not intended to diagnose, treat, cure, or prevent any disease. Garmin recommends you consult your doctor before engaging in any exercise routine.
Accuracy of Wrist-based Heart Rate (Elevate)
The optical wrist heart rate (HR) monitor for Garmin wearables is a valuable tool that can provide an accurate estimation of the user’s heart rate at any given point in time. The optical HR monitor is designed to attempt to monitor a user’s heart rate 24 hours a day, 7 days a week. The frequency at which heart rate is measured varies, and depends on the level of activity of the user. When you start an activity with your Garmin optical HR device, the optical HR monitor provides feedback more frequently as the optical sensor is on all of the time and is trying to measure heart rate on a continuous basis during a given activity period. The intent is to provide the user with a more frequent and accurate heart rate reading during a given activity.
While our wrist HR monitor technology is state of the art, there are inherent limitations with the technology that may cause some of the heart rate readings to be inaccurate under certain circumstances. These circumstances include the user’s physical characteristics, the fit of the device and the type and intensity of the activity as outlined above. The HR monitor data is not intended to be used for medical purposes, nor is it intended to diagnose, treat, cure or prevent any disease or condition.
http://www.consumerreports.org...
The new testing confirmed our earlier results: Both the Charge HR and Surge were very accurate when compared to the reference Polar H7 ECG monitor. During nearly every trial, the variance between the chest strap and the Fitbit devices amounted to no more than three heartbeats per minute.
However, there was one exception: When our female tester wore the Fitbit Charge HR on her wrist and got up to higher intensity levels, the margin of error crept upwards. During one run, when the chest strap read 150 bpm, the Fitbit Charge HR read 144 bpm. During the second run, the device read only 139 bpm. That problem went away when she wore the Charge HR on her forearm. (And the Fitbit Surge was accurate no matter how it was worn.)
I'm sure there's plenty of caveat emptor to go around, but the very first thing I did when I got my Charge HR was to test it by taking my pulse myself and comparing it to the fitbit number.
The resting heart rate is dead on. (not their calculated resting heart rate--that's a dumb, arbitrary number they come up with, but the real time resting heart rate is accurate.)
I've had a few troubles with the workout heart rate, but not because it can't count, but because the wristband moves around after sweat dislodges it. It can slip down my wrist and the sensor will lose contact with my skin. But when that doesn't happen, the number is pretty close.
For a while at first, I wore my under armor heart rate monitor on my chest along with the fitbit, and they were pretty much in sync, too.
This lawsuit is why we can't have nice things.
It's not about max heart rate. some people set a range and try to run within that. don't know why.
"Well, then. I was thinking of getting one to help me get and stay close to max-rate for interval training. 14% inaccuracy would be wildly dangerous in that case."
I do hope you're being sarcastic. If not then you might want to consider listening to your body instead of reliying on some stupid device. You know, what athletes people have done for millenia. If you're so disconnected from what your body is telling you perhaps you shouldn't be training at all.
The lawyers sponsored the study, and it looks like it's actually pretty decent methodology--comparing the pulse from the EKG (which I certainly hope is reliable) to the FitBits on the wrists at the same time means you've got one of the gold standards of study design, since the pulse ought to be the same in the same person no matter what point you measure it from (and how you measure it) so you make it a lot easier to analyze the data. You've controlled both for the issues if you used different subjects and for if you measured at different times, so the only real question when doing the analysis is "How well do the parts of each set align?" (You do multiple subjects to make sure you didn't have something rather weird happen like somebody somehow pulling off 'different pulse depending on where measured,' and the number looks actually rather good--a bit more subjects than I'd expect a study like this without money being thrown at it to have, actually.)
The only thing that seems particularly weird here is that I'd have expected the company to have sponsored this study or something very much like it during its own R&D cycle, because it would be both good for marketing and good practices. Doing this sort of basic study is pretty...basic, if you're trying to make something that monitors {foo}. About the only reason I could see for not doing that would be "FDA might get panties in a wad," except that is something that I'm amazed hasn't happened already...and I would be wanting to develop a medical-grade version anyway. It should sell, especially if I can add in remote monitoring; that would probably net me sales to both health care facilities and in professional sports...
Seems easy to do, doesn't it? You take the reciprocal of the interval between heartbeats scale that to beats per minute and there you go.
Except if you've ever designed software and had to look at a problem like this, you'll understand that it's not nearly so simple. The heart isn't nearly precise as the quartz oscillator you're using for your timing reference. Sometimes it throws in an early beat or late one, because it's an electro-chemo-mechanical oscillator that works by the flow of ions across a membranes of countless cells. So the rate is going to jump around a little bit second to second, which will only confuse the user. What's more any sensor that isn't stuck on with adhesive is going to miss a few signals now and then, or maybe get a spurious one.
So what you do is take a moving average to smooth out all the kinds of noise you have in your signal. Having a fitbit myself, I find it surprisingly accurate when compared against a manual pulse when your heart rate is steady. Watching the figures as I exercise hard, it's clear there's a lag in response as your heart rate accelerates in particular (since the heart slows this is less of a problem as you slow down), which indicates that the device is giving me some kind of moving average.
And it's OK. The point is to give users useful feedback about how their heart is doing during periods of strenuous exercise, not to give them a beat by beat accounting of what the sensors are picking up. It doesn't have to be laboratory-precise or instantaneous because there's nothing useful users can do with that precision. I've lost count of many times have I had to explain these two things to clients: you don't need feedback that's an order of magnitude faster than you're capable of responding too, and you shouldn't take action based on statistical noise; data doesn't have to be perfectly precise, they just have to tell you whether you're in the right ballpark or headed in the right direction.
The fitbit works very well for the purpose it's intended and the way users use it. You take a series of readings (which are sometimes off, but recognizably so) to establish what your heart is doing, and then shift to a new level of effort and take several readings as your heart settles into a new equilibrium.
Anyhow, note the weasel phrase "Up to 20 beats per minute." This means in all their testing the very worst discrepancy they ever found between the readings taken from glued on electrodes from the loosely cinched bracelet monitor was 20 beats per minute. That actually sounds pretty good to me for a worst case result.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
For max rate training you'd be far better equipped with a Polar heart monitor band and a watch that syncs to it.
Yes, the chest band isn't as comfortable, but they're a lot more accurate.
Zephyr's BioHarness is found to return pulse rates that are off by as much as 20 BPM.
Gamingmuseum.com: Give your 3D accelerator a rest.
"miscalculated heart rates by up to 20 beats per minute"
"Up to" -- like "as much as" -- is a marketing phrase, not a description of results. I haven't read the study -- I didn't see any useful information in the linked NBC News article as to how to find it.
and nothing more than an attempt to extract a payout from Fitbit.
This is an accurate statement, even if the watch has problems.
(-1: Post disagrees with my already-settled worldview) is not a valid mod option.
Once again I come here to defend my Fitbit Charge HR: I'm trying to lose weight, and it works great for keeping a gross measure of calorie burn vs intake. Its "imprecise" HRM is a lot better than a step counter at counting calories spent, even on average - even if the instantaneous heart rate is 10-20 BPM off from the actual one, you're likely still in the zone it uses to calculate calories. And it isn't off that long from my own experience. I have simultaneously used twice or "thrice" a Charge HR, Mi Band Pulse and chest HRM on workout sessions, and despite both imaging-based devices rarely being on the same as the chest one, they correct themselves to "close enough" status after seconds, and you don't change zones in seconds. The Fitbit is much better than the Mi Band as it clearly scans with a much higher frequency (I can see just from the leds lighting more often on exercise mode on both). At the end of the day, apps that use the chest monitor will basically provide the same calorie output as both the Charge and the Mi Band. I believe all this ruckus from over-expectationers is because they couldn't distinguish the type of product they were buying - people in the cities, more likely to purchase such devices are also more prone to use gyms and practice amateur sports, thus require better accuracy found in a chest HRM. But in part, this was also Fitbit's mistake: selling the product as such and targeting a market too focused on unimportant markers such as second-accurate BPMs.
Exactly how has this harmed anyone?
Everyone knows pedometers aren't accurate. This is just a continuation of what we all knew in the 1980s and 1990s.
I don't see how this lawsuit has any merit whatsoever. Whom has it harmed, and how?
What about all the other pedometers, like the ones health insurance companies issue to employees with which points are earned to obtain discounts?
Kriston
Good to see some actual studies done on this, but I think that many people who work out already knew about this. Here's a review of the Fitbit Charge HR which shows that during periods of intense exercise, that the fitbit doesn't track very well. It seems to do a pretty good job when you aren't exercising, but compared even to a heart rate strap worn on the chest, the accuracy is lacking.
Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
All measurements come with a measurement error. At what point do we say that something is "accurate" or "inaccurate"?
If I have a heart rate of 150 bpm while exercising and the fitbit reads 130 (off by 20 beats per minute), it is reporting the correct value to within 87% That seems pretty good to me for a cheap device. How much would it have to improve to be "accurate"?
For whatever reason it does not seem to pick up efforts on the bike properly, often I'm 170-180bpm (puffing hard and putting in a good effort) and the Fitbit is reading 110-130, and if the Fitbit forum is anything to go by a lot of other people have issues with it not really working properly on the bike.
There is a lot of ground between the reality and the claim here. The claim is that it is not a medical device, but that leaves a lot open to interpretation. Common sense would indicate if it shows you a number then it must be a useful number. I've used a polar heart rate monitor and it flashes if it feels it is not getting an accurate reading. Sadly, the quest for profits knows no common sense.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
The electrical signals that are driving your heart are very small to start with, and what's worse, they have to be sensed through your skin, which has a very high impedance to start with. The best way to acquire that signal is differentially, preferably with sensing electrodes as widely separated as possible. A FitBit sits on your wrist, and the electrodes are very close together, and more to the point they're on the same side of your body. Even a traditional heartrate monitor chest strap (i.e. Polar, etc) at least has electrodes separated by several inches, and is close to your heart. Next, the signal is small, and requires much amplification to get a usable signal out of it; the further away from your heart, the more gain is required -- which also amplifies the noise, by the way. Then there's the ever-present problem of filtering noise (60Hz noise is the most prevalent, but all noise is a problem). Finally, based on my own experience with heartrate monitor straps and physical exercise, just having the electrodes bouncing around against your skin can cause false readings, typically higher than normal. Oh and by the way, all the above problems are worse when your skin is dry; the electrodes need to be moistened to really work at all. I can't imagine it being anything other than way worse if it's on your wrist. Therefore: no way in hell is a FitBit going to be terribly accurate, unless you're sitting still and keeping the electrodes moist.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
http://www.consumerreports.org...
They want to maximize fat burning.
Don’t Wear It Too Tight
Make sure your band isn’t too tight. Wear the band loosely enough that it can move back and forth on your wrist.
If you use Fitbit Charge HR, Fitbit Blaze or Fitbit Surge, for better heart rate readings during exercise, we suggest wearing the band so it’s secure, but not too tight, and wearing the band higher on your wrist (about 2-3 finger widths above your wrist bone). Lower the band on your wrist and loosen it after exercise
https://www.fitbit.com/product...
I have a similar (non-fitbit) watch I use for running. I like it. I understand it's not 100% accurate. It's accurate enough for its cost, and I'm happy with that. If lawsuits like this continue to fly, such products will get more expensive, or harder to find.
My Apple Watch doesn't do a great job either. I've looked at a number of these wrist worn optical devices and none of them are great.
I think people have false expectations of what these things can do. Fitbit, Apple and all these companies are well advised to set expectations properly.
That said, I'm glad to have it even if its not completely accurate. I use it as a guide and motivational tool. I use my phone as a flashlight sometimes. Its not a great flashlight, but its handy when I need one. Same thing with the heart rate monitor on the wrist; its not perfect but good enough with the convenience makes it worth having for me.
Anybody who needs something accurate should look at something like the Polar Chest strap technology.
Greed is the root of all evil.
I'm not here to defend health gadgets but to condemn the impression that traditional measuring devices are accurate. According to TFS "Subjects were then hooked up ... "--in other words there was a sequence where the gadget was used and later a supposedly accurate device was compared. A simultaneous comparison might have been valid but sequential is questionable.
I test myself about five times every morning over a roughly ten minute period. Each reading is different, often significantly. Not just heart rate, but blood pressure too. In fact after years of frequent tests of all my vital signs, I find that there is so much variation that only a monthly average gives me a useful number.
Don't condemn the gadget without a proper simultaneous test with a proven accurate device.
...omphaloskepsis often...
It's a fair cop, guv'nor. You got bang me to rights.
I think everything about the fitbit charge hr is innaccurate including the miles, steps, calories and hr! Piece of junk!
The early fitbits were basically a computer connected pedometer for goodness sakes (!) yet they charged $100 for it. /accurately/ measures heart rate, altitude, gps coordinates, ambient temperature and steps (if even possible??) could probably give semi-reasonable estimations on health benefits or calories burned but even a highly accurate device wouldn't be perfect. However the consumer stuff fitbit shovel out is weak at beast.
I am sure a device which
I've had no big heart rate monitoring issues with my Charge HR. Or rather, I should say the three different Charge HRs I have had wherein two had batteries that just completely failed (replaced under warranty) to retain a charge after a few months use.
I am 67 years old and questioned the accuracy of the Fitbit HR as I was constantly showing a heart rate of 64 to 68 at rest. I haven't had this heart rate since I was in my teens and in prime rate as I played multiple sports. My doctor monitors my heart rate at rest and it usually measures 68-72 but never 64. In fairness I haven't tried it when exercising but am schectible about the accuracy.
This study was not peer reviewed, and it appears to me that it has problems that would make it unpublishable. The methods section of the report describes data collection in a way that makes it clear that instructions for taking accurate heart rate readings in the Fitbit manual were purposefully ignored. These instructions describe stopping, staying still for a few seconds, and then taking a reading. This paragraph is just above the paragraph containing dominant and non-dominant wrist instructions which the report methods do pay particular attention to, so it could not have been missed. The description of data collection makes it clear that stopping and staying still was never part of the protocol since data was collected continuously during active exercise. The methodology used by Fitbit and all other wrist devices using light will never be accurate under these circumstances. The device is measuring the very small motions of blood vessels under the skin. If the device itself is moving, these measurements are disrupted. The scientists performing the study must have known this, so it is difficult to understand why there was at least no discussion of the Fitbit manual instructions and why they were dismissed in the text of the report. A reviewer, like me, would reject a paper with an omission like this. I performed my own series of scientific tests of the Blaze accuracy in March. Wearing both my chest band and my Fitbit for a week of 4 mile runs, and following the instructions, l found a difference of 2 beats per minute (bpm) with a standard deviation of 3.2 bpm at 95% of my maximum heart rate. I can understand consumers not reading the manual, initially, but eventually they should. A scientist performing a study, on the other hand... https://www.researchgate.net/p...
I have a Fitbit Surge and a Garmin 405cx with a heart rate strap that I've used for years. I always wear both when I work out. [Sitting/Walking] - Fitbit Surge matches my HR if I check my pulse with my hand [Running], even up to a 7 min/mi pace for 4-6 miles - the Fitbit Surge and Garmin are within 1-2 beats of each other [Weight lifting] - Fitbit is HORRIBLY inaccurate. Typically 30 beats off, but oftentimes up to 50-70 beats off right after a tough set. I'll regularly walk around between sets and see the Fitbit at 95-105 and the Garmin at 130-150. When I check my pulse, it is always within 5 beats of the Garmin's HR. I once read that it was a design choice to make a smaller HR monitor that uses less battery in exchange for some accuracy. Don't know how true that is, but the differences are there.
They make NO False claims. They disclose the innaccuracy in the devices Manuals themselves.
Fitbit Charge HR Manual.
"With high-intensity interval training or other activities where your wrist is moving vigorously and non-rhythmically, the movement may prevent the sensor from finding an accurate heart rate. Similarly, with exercises such as weight lifting or rowing, your wrist muscles may flex in such a way that the band tightens and loosens during exercise. Try relaxing your wrist and staying still briefly (about 10 seconds), after which you should see an improved heart rate reading."
https://staticcs.fitbit.com/content/assets/help/manuals/manual_charge_hr_en_US.pdf
Can we say frivolous lawsuit?
... Most of the time when I'm exercising I can't get a number out of mine.