Does anyone know if the Charge 5 ECG is at least somewhat accurate?

[Over65]

I'm on the fence about returning this device.  Heart rate during exercise is off by an hilarious amount.  For example, the Charge 5 typically shows something like 108 BPM while pulse at the wrist and carotid agree on 150 BPM.  Now I'm wondering if the ECG is also way off as well. 

Has anyone out there who has access to a real ECG device made any comparisons?  What did you find?

[GoGoGlade]

It's not a medical device. It skips on the wrist if worn so it doesn't cause irritation. It is not a toy for children but it is an adult toy. I doubt there is a device that is accurate that can be carried with you while exercising.

Gogo glade

[Over65]

Just replying to myself: I did a search on the Apple Watch and discovered it also suffers from heart rate monitoring problems -- looks like the wrist devices just aren't there yet.  Knowing this, I'm more forgiving of the Fitbit.  Also, wrt ECG -- Fitbit seems to compare reasonably well against the Apple watch and data showed ECG to be surprisingly accurate.  Should have looked at this more carefully before spinning up the Community.

[Bike-man]

Hi Over65’

I responded a while back with another Charge 5 users concern for ECG accuracy.

Devices such as the Charge 5 and Apple Watch are not medical grade devices.  Consequently, you can expect mixed results.  Much depends on the placement on the arm, temperature, and knowing what your normal ECG is .  The ECG associated with using a wrist device may be able to demonstrate changes in heart rate (HR) rate and rhythm but may not always be 100% reliable.  
I use mine to track my resting heart rate.  When cycling, I believe changes in rate may be fairly accurate as I have little motion artifact due to my arms and hands being generally in a fixed position.   When doing other exercise such as running, there is a possibility that excess motion may interfere with skin contact and thus have an effect on overall HR rate accuracy.  
My opinion is based on my experience conducting CardioPulmonary exercise studies in a hospital setting. 
Hope this helps.  

[DPinOz]

Hi,

The ECG app is accurate, however the ECG app does not determine your resting or spot pulse rate.

The ECG app uses electrical sensors on the sides and back of the watch to create a circuit that passively monitors the electrical activity across your heart. This monitors your heart activity accurately for 30seconds, and most importantly provides a trace of that activity. This trace can be downloaded, and can be used to spot heart irregularities, and is FDA approved for detection of Atrial Fibrillation. The report shows electrical wave forms PQRST. But this trace is a Single Lead trace, and not as diagnostic as a 12 Lead trace performed on a typical ECG machine. The report will also show you a pulse rate that is highly accurate for the 30 seconds it records.

Your regular pulse rate, resting rate and averages are determined by light sensors on the back of the watch, but I'm guessing it's an average over some time frame, which is why you may see different results at different times, medically you really are interested in various pulse ranges, and over 180 beats per minute for a prolonged period is worrisome. This light sensor reading is most helpful for determining if you are in an exercise zone (Fat burning, Cardio, Peak) during and post expercise.

Watches and devices without electrical sensing and ECG only give you a hint that you may have an irregular heart beat over time. I have found my Fitbit's sleep data helpful, as you get no or incomplete data if you don't have a regular pulse. The ECG app is then helpful for sorting this out.

So, yes, your Fitbit if used as intended is accurate. 

FYI, I have intermittent Atrial Fibrillation, and I find this device very helpful as it is always on my wrist. I have compared the ECG with my Kardia 6L EKG (which gives a better 6 lead ECG trace) but I use my Fitbit for a quick and less sensitive determinations more often. I have no affiliation with either Fitbit or Kardia, just a user.

[DPinOz]

Oh...

Forgot to mention, when in doubt take you pulse properly with your fingers from your wrist or carotid. You just need your fingers and a watch/clock with a second hand. (Annoyingly Fitbit Charge 5 watch faces are pretty useless).

You can feel the pulse, and spot an irregular beat because it will mess up your count. It's just harder with higher pulse rates.

You could buy a stethoscope and listen too, but that's possibly obsessive.

[charleskn]

I don't believe the Charge 5 has the ECG feature. The in-accuracy of wrist worn heart rate monitors, using photoplethysmography, is well kown. The Quantified Scientist on YouTube does excellent reviews. In general Fitbit does fairly well, but has typical limitations:

• The wrist placement puts the optical monitor over bone and not much "meat"
• Darker skin tones and tattoos affect readings
• Exercises such as weight lifting can cause inaccurate readings, and HR during HIIT can be hard to track. My Charge 4 does much better at tracking HR during cardio (bike, rowing, etc) due to the relatively constant HR.
• The HR reading tends to lag reality

My solutions:

• I keep my Charge 4 on the inside of my wrist and a tad further up my arm, but that means I have to make sure it stays snug
• I now use a Polar optical arm-band as my main HR monitor for exercise

[Over65 · Accepted solution]

I'm replying to my own post because I did some research instead of being lazy.  Here's the lowdown on both HR accuracy and ECG accuracy:

1.  Heart Rate.  Finally RTFM and low and behold, HR accuracy shot up markedly when I did what the company recommended!  HR recording accuracy went from hugely wrong - off by 30%-40% to within 5%!  Fitbit tells us to make sure there are at least two finger widths between the device and the ulna's styloid process (the big bump bone on the top and outside of your wrist).  Here it is:

When I put the Charge 5 in that spot, voila!  Massive improvement in HR accuracy.  It's still off, but by a massively smaller amount.

2. ECG accuracy.  For detecting aFib, it's fairly accurate and specific.  It has FDA approval to market this so the device is not a toy by any means.  I compared my doctor's ECG Lead 1 recording and it looked almost identical to the Fitbit Charge 5's ECG recording.  Fitbits says its ECG is comparable to a full-up medical ECG's Lead 1.  In particular, I could clearly see my R and T waves, for those who know what I'm talking about (I'm a retired researcher).  The Q wave and S waves were also sometimes visible.  P waves were the hardest to detect versus noise.  It was interesting that I found one research article saying the Fitbit Charge 5 was more accurate than the Apple Smartwatch (Apple is much pricier than the Charge 5), and another article showing the opposite, but this may be due to version changes as both product lines mature.  

Bottom line: considering the price of this thing, it's pretty **ahem** good.  I should have RTFM in the first place and should have looked up Pubmed data before I asked the question about ECG accuracy.  Apologies to the Community, but hoping I made up for it with this bit of investigation.

[charleskn]

Color me silly for missing that the Charge 5 has the ECG function. I am not surprised that the plots are accurate, but noisy, for the ECG. The optical heart rate monitor has so many other variables along with the algorithm that has to interpret the optical signals.

[fazer600]

I have just done a tredmill test (last saturday) and whilst i was cooling off i checked my pulse rate to the machine that measures everything whilst you run/walk (not sure what they call it) anyway i must say my fitbit charge was pretty much spot on it was abt 2 beats pm min out as for the ecg i`d love to know that they are accurate as i have heart issues.

[Over65]

Another update: BPM Heart Rate is very accurate when one's forearm isn't moving such as during cycling.  However, if running or using an elliptical in the gym, BPM is still way way off.  Sometimes as much as 40%.  I'm requesting Fitbit allow us to manually enter BPM and exercise duration.  I for one know exactly how to find my pulse over time and can easily estimate it over thirty to sixty minutes.  This will make Fitbit *much* more useful until the technology evolves.  For example, Fitbit has no idea how many calories I should eat, rendering this measure irrelevant, because it has no idea how many calories I'm actually burning.  Where is Fitbit marketing on this issue?  Don't they want to sell more of these things?