How does Fitbit calculate Resting Heart Rate?

[BigHaole]

I'm finding this very frustrating to be sitting at my desk, and see my HR showing as 74, but my Resting HR being 80.  Clearly Fitbit is not using the conventional definition (from Wikipedia):

"The basal or resting heart rate (HRrest) is defined as the heart rate when a person is awake, in a neutrally temperate environment, and has not undergone any recent exertion or stimulation, such as stress or surprise."

This definition would lead me to expect my reported resting heart rate to be the low value reached in the early morning, or at least the low value I reach, during the day.  Instead, it's above both of these.  I would like to know how it is being calculated, so I can know if my Fitbit is reporting anything useful when this number goes up or down.  Over the recent new years holiday, I got more sleep and more exercise, with less stress, so I was expecting this to go down, but it has gone up and I do not understand why.

[Metsci]

The Microsoft fitness band claimed that it takes your resting heart rate as your heart rate 30 minutes BEFORE you wake up. 

Since it logs the data, it waits until it thinks you woke up and then looks back 30 minutes.

Sounds like a pretty good rule.

Did not find anything, yet, that tells be how FitBit actually does it.  As some other poster mentioned, my FitBit recorded one does not change during the day, but always shows the rate it tells me about first thing in the morning.

Maybe the original poster has very restless sleep some mornings just before he/she wakes up. 

[Kagakusha]

The only problem being that is NOT the normal medical definition of ‘resting heart rate’!
That would be ‘sleeping heart rate’.

Your resting heart rate is defined as when AWAKE but at rest. In fact poorly defined, as it does not accurately define what are rest’ means - for how long for example - but it’s unequivocally AWAKE.

Moderator edit: personal info removed

[Metsci]

Apparently Microsoft doesn't think it's "unequivocally AWAKE".

[Kagakusha]

Microsoft is a computer software company, not an authority on heart rate.

The normal medical definition is awake.
In fact it is poorly defined medically, because there is no detailed definition of what ‘resting’ means, but it is invariably ‘awake’.

One reason is straightforward - until fitness watches appeared, unless hooked up to monitors in hospital, it was inevitably ‘awake’, because you had to be awake to measure it!

Measures using data whilst asleep may have their uses. They might even be more reproducible than the traditional measure. But they are different and cannot be compared to it.

Moderator edit: personal info removed

[Colmcq]

Fitbit RHR 56.

About an hour ago: 41

It's not even a bit wrong; its 40% wrong.

The could just calculate resting heart rate more accurately by saying: "what was lowest 2 mins during 24 hours" and have done with it. The current algorigthm is a joke.

that sounds perfect. But it isn't what's used.

Moderator edit: merged reply

[Metsci]

Neither are you.

[Kagakusha]

And I wonder on what basis you reach that conclusion, in the absence of any data.

In fact as a professional scientist I have full access to the scientific literature, and have researched this as it interested me. I read the original papers, not just Wiki etc. You might like to translate my pen name from Japanese.

The result is perfectly clear; the normal definition of resting heart rate is *awake*.
That’s unsurprising, as until recently that’s when data was available other than for hospital patients - who are presumably ill.
But as a scientist I regard it as poorly defined, as there is no accepted definition of how long you should ‘rest’.

The Fitbit definition can be used to compare to other measurements using the Fitbit algorithm. But it is meaningless to compare it to anything else.

Moderator edit: personal info removed

[SunsetRunner]

So remember fitbit is essentially for soccer moms who want to measure their steps and help remind them of fitness level for the day to become more health conscious. These devices are not edge of tech medical instruments for measuring every facet of heart activity.  You could get a chest strap if you are a little more serious which obviously you take this very seriously. As a "professional scientist" I would expect you to be able to gather the data your fitbit collects and analyze it yourself. Fitbits are essentially 100-300 dollar glorified pulse oximeters. I know that and I'm not a pro scientist. I don't understand why people expect the world handed to them when they spend a little money on something. The rhr algorithm isn't going to be perfect, nothing ever is stop,expecting perfection. I would expect a scientist to at least use a chest strap or for someone as obsessed with the subject matter to send the day at their desk hooked up to a 12 lead ecg.

[SunsetRunner]

BTW from personal experience I believe fitbit devices work great and feel they accurately represent data collected

[cctennant]

Just to add another comment here.  It's absolutely clear, as Kagakusha says, that RHR is supposed to be calculated when you're "awake and at rest"  It's also clear that there's no standardized way of doing that. How could there be?  What it you've just drunk 5 espressos?  Are stressed out about a job interview?  Have rested for 60 seconds, or rested for 30 minutes?

I've been dual-wearing a Fitbit Ionic and an Apple Watch 4, for 9 months now.  The Apple Watch does make an attempt to calculate resting heart rate while you're awake.  From what I can see, the AW doesn't use your absolute lowest heart rate during the day.  So I might have  a lowest heart rate of 54 for the day but it lists my resting heart as 57 or 58.  At any rate, given the fact that there is no standardized way of calculating RHR, the AW algorithm seems a reasonable one to me, and the results have made sense to me as I've tracked them over the past nine months.

Because I wear the Fitbit every night, and the AW every day during the day, I'm able to compare the two RHR numbers.  Fitbit's number is generally close to the Apple Watch, usually within +/- 4-5 BPM.  I don't think that's because either one is a bad measure, but just because there's so much inherent variability in how you sleep, and even more so in what happens during the day.  However, to Kagakusha's point that the FitBit number is measuring *something* even if it isn't precisely RHR, the FitBit number does tend to be more consistent over time because the conditions when I sleep at night tend to be more consistent than what happens during the day.  So while the FitBit number isn't technically a resting heart rate, I find it a very useful measure to track over time, looking in particular at how and when the rate changes.  So for example it's very useful to see when I'm overtraining, because my Fitbit RHR tends to go up (and so does the Apple Watch calculated RHR for that matter).

There's another issue with the Fitbit RHR number, which is that their algorithm does some averaging over multiple days.  This means that big changes in your night-time heart rate don't show up immediately and fully in the RHR number.  You can tell this because you can get access to the raw HR data through the FitBit API, so I'm able to compare Fitbit's calculated numbers to my actual heart rate during the night.  You can see this yourself when you look at the heart rate graph in the FitBit app.  It's best to look at the graph right when you wake up -- if you look at it later in the date, the small changes while you sleep get swamped when the graph re-scales with the large changes in heart rate that typically happen during the day. 

But even with this caveat, I do find the FitBit RHR number useful, so long as I look at trends over multiple days.

[Kagakusha]

Broadly I agree with those comments.
The Fitbit number can be useful, but because it uses an unknown algorithm involving sleeping heart rate, it differs from other measures and should not be compared to them.
In my view probably only trends over quite long periods have much meaning. Day to day fluctuations aren’t considerable.

From my iPhone - please excuse spelling checker induced nonsense!

Moderator edit: personal info removed

[Oregonmama]

It can't be based off of my heart rate when I am still. My rate when I am still is about 10 BPM lower. I bought the device to monitor my RHR. I have Graves and it seemed like an easy way to get regular monitoring, but now I wonder if it is worth having. It's pretty close on the days when I sit quietly enough in the early morning that the Fitbit thinks I went back to sleep, but if I get up and start doing things right away, it's about 10 BPM high. In the evening, I will often sit quietly and get the lower number. Unfortunately, I can't export my heart rate data (another big disappointment) so there isn't a way to track my real resting heart rate without poring over the graph each day and manually recording the number. Again, this is the whole reason I have the device, so when I see Fitbit continue to propagate this bit of misinformation, instead of helping those of us who don't hit the snooze button, I'm wondering why I'm keeping it.

[Kitesurfer]

Secret recipe? Are you joking? This is our health you are playing with! Now, please, go and publish your formula.

[jpsfla]

Well at least the resting heartbeat monitor was able for me to see my arrhythmia spike of 173 beats per minute back in December. Plus the sleep modes would not work regularly because of an irregular heartbeat and PVCs. My doctor even put me on a Holter fora day.

Drinking beet juice and the resting heart rate has came back to 58 to 60 (normal for me) instead of 35 to 40 and the sleep function works now. So regardless of how they do it, it works. No meds either!

Sent from my iPad

[Oregonmama]

As a professional engineer, I can tell you this can't be done, because Fitbit won't let you collect (export) the data. I tried when I found the algorithm wasn't telling me what I needed to know. I'm Peri-menopausal and my sleep is all over the place as my hormones fluctuate. I bought the Fitbit to track my RHR, since that seems to be a good indicator for my hyperthyroidism, but it is only sort of useful, because of the way the algorithm works.

[ZumbaSteph]

Thank you so much. At 54, I am deep in menopause and still having hormonal fluctuations, and hypothyroid since I was 14. So I just decided that a different avenue would be better, and went with Heart Rate Variability technology and research for health indications. I prefer this and feel that I get better data and I can figure out my own RHR. Especially after I fgured out that my old Charge 3, and my newer Versa indicate completely different RHR!Sent from my Verizon, Samsung Galaxy smartphone

[Oregonmama]

ZumbaSteph Sounds interesting. What are you using? I can get a rough idea by looking at the Fitbit graph, but would sure love something that makes it easier.

[Eileeniebeen]

Wearing my Fitbit to bed or not doesn't affect my Fitbit reported resting heart rate and I think it's very inaccurate.  For example, as I sit here typing this and drinking coffee, my heart rate is 65-66 and yet my rhr is supposedly 70. Typically when I sit and read a book my hr is 59-60.

[CarlaDelgado]

I have a question, what period of time of inactivity but lowered heart rate makes Fitbit change your resting heart rate? Ie I have recently migrated from a Garmin vivisport where 10 minutes of a constant lowered heart rate during a period of inactivity reduced or changed the resting heart rate for the day, what period in light of the above does Fitbit need to recalculate resting heart rate?

[Kagakusha]

Since Fitbit do not publish the algorithm, nobody except them knows.
It is meaningless to compare a Fitbit resting heart rate to any other measure of this parameter - which is in any case poorly defined.
There is probably some meaning to watching long term trends in resting heart rate as measured by Fitbit.
But none in comparing it to other measurements.