Your smartwatch is lying to you 

If you thought you’d seen the last of Dr Rohin ‘Medlife Crisis’ Francis and his magnificient moustache, think again. And I’ve actually made more Medlife content since this vid so check out the smartwatches playlist for me of me and Rohin chatting wearables: ru-vid.com/group/PLg0VbZ0kyCHl0yKBAQZ0NaI-Gxax6_0oX

That moustache has it's own resting heart rate.

You can find more clarification on Apple's RHR figure in their 'Health' app. Specifically, it says "Your resting heart rate is the average heart beats per minute measured when you've been inactive or relaxed for serval minutes." and continues with "Resting heart rate does not include your heart rate while you're asleep and is validated for users over the age of 18 years."

I've been wearing a Fitbit for over 3 years, and I believe that the Fitbit resting heart rate (RHR) feature is useful for two reasons. 1). Monitoring your RHR over the past year can indicate whether your health is improving or deteriorating. For instance, I quit consuming caffeine six months ago and noticed my RHR gradually decreasing. When I contracted COVID-19 three months ago, my RHR skyrocketed during the 10-day period, even while I was asleep. As my RHR began to decline, I started recovering from COVID-19. 2). I also check the lowest heart rate during my sleep, as I find it helpful in assessing my overall well-being. I've noticed that my lowest heart rate decreases as my body becomes healthier. In my opinion, the exact RHR number doesn't matter; what's important is observing the trend, whether it's going up or down, as it provides valuable insight into your health.

Amazing breakdown of all the data points, you've displayed them all fantastically. I've gotta say, I really do think the the 'cynical' reason you started with is the real reason. But yes, I accept the alternative that it's just simpler to measure overnight, however unlike the cardiac monitors we tend to use for patients, wearable devices have a major extra feature - accelerometers. Surely it would be really easy to find a 5 minute period with minimal activity and then use that for RHR. And maybe repeat several times. I think sleeping HR is a perfectly good metric to track (as you suggest), but people should be made aware of the difference. We've been focusing on the low end, but I would hate people ignoring 90bpm overnight because they think it's in the normal range of 60-100. If someone's HR is that high at night, I'd recommend getting checked for things like sleep apnoea etc.

My AW telled me 2 weeks after I stopped smoking, that my average resting heartrate was going down from 80bpm to 60bpm, since the day I stopped smoking. That was a really cool surprise. :)

while not backed by any data, my theory as to why the clinical world has settled on 60-100 instead of 50-90 as their RHR range, is because of the lack of time in a clinic setting. they take your heart rate & blood pressure after having gotten up, spoken with the nurses, maybe gotten nervous, walked around, etc. they don't have time to let you sit for 10 minutes in a stress-free environment just to get your heart rate.

Whenever I go to the Doctor's office, my heart rate is like 20-30 BPM over what it is normally. I don't want to tell them that it's because all the nurses that work there are ridiculously attractive and I'm trying very hard not to say anything too stupid

Great clear video man! Conclusion; Don’t trust the number, trust the trend… Going up or down indicates more than the actual number. Same for body fat % on scales if you’d ask me :)

As a cardiologist I can tell you *all peripherals* (including BP monitors) under-count the true *heart rate* (i.e. the number of times the heart contracts). This is different to the *pulse rate* How can this be true? Well even fit people get ectopic beats (VPC, PVCs etc) which cause a cardiac contraction, but not a strong enough pulse wave to register peripherally. I'm sure each device has a different threshold for detecting weaker pulse waves. Furthermore, these devices use temporal smoothing - some up to 10 beats at a time. That is why you don't see an instantaneous jump from 60bpm to 100bpm like you would on an ECG. The number is smoothed over time and can result in varying heart rates at any given time.

As a Garmin user I know that my resting heart rate is roughly 45 BPM and when I have went into the NHS for a check up they always correlate that my resting heart rate is in the 40's and they ask me if I am an athelete, which I do fitness train, so its not any other underlying cause. I was always told that Garmin has a 10% error factor built in.

My feeling is that the conventional definition is lower quality. 5 minutes of rest just in from a long run probably settles at a significantly higher number than after a short burst of exercise or a previously rested state, etc. Overnight conditions might tend to be more consistent. Maybe?

The thing is, a traditional ‘taken by a doctor in an uncomfortable clinic chair’ HR is way less likely to be accurate - no-one is relaxing in that environment. So it might be more true to say “traditional medical science is lying to you about what a healthy HR is”. And it’s easy to verify smartwatch measurements. I’ve done this many times to verify that my ‘slobbed out of the sofa in the evening’ HR of 38 or something is accurate. It is. Just now I’ve just sat on the sofa and confirmed my watch’s reading of 43bpm. Which is actually 2 beats lower that it’s been showing me as my RHR this week and 2 beats higher than my overnight sleeping low. So I’d say the watch is doing a good job, and it’s the doctors who need to catch up!

I wonder how reproducible each platform is. If I'm working toward getting more fit, I could be persuaded that it doesn't really matter if my indicated RHR is 58 or 48, as long as the trend is headed downward.

In the Apple health app, German version, I am told, that resting heart rate is validated for persons over 18 years old and is specifically not monitored during sleeping time. So, which is it?

As a data scientist I appreciate the discussion on validation metrics R square!

My SE 2. gen Apple Watch doesn't seem to include sleeping data. I'm averaging around from 50 - 60 resting but I have below 50 when i'm sleeping.

Resting heart and hear rate is a tough number to measure health in my mind. For someone who's numbers are in the 60s while sitting but if i try to monitor it by myself it rises in the 90s. We things called emotions (anxiety) that can alter numbers. So sleeping heart rate seems like a better metric to me.. Mine is like 48-53 while sleeping.

The video is already a year old, but I know for a fact that in the past, the Apple Watch didn’t (correctly) distinguish between sleeping heart rate and daytime resting heart rate before the sleep mode launched. I used to use a third-party sleep tracking app, and once I started using that, my resting heart rate dropped significantly. When the sleep mode launched, the resting heart rate went back up to a more realistic value again.

Waking resting heart rate depends on what I've been doing. If I exercised in the past 30min, probably 80. If I've been sitting for an hour, probably 70. If I've been sitting for a few hours, probably 60. If I've been relaxing for a few days, probably min-50. My sleeping heart rate is nearly the same for every one of these situations. Low 50s.

How do we know YOU are not lying to us?

in apple health app, when you look at resting heart rate. its written that' its measured when you are at rest for a couple of minutes. and that its not a sleep measurement. so I think apple is doing a good job. my Garmin epix pro measures during the night. but in my case they provide the same resting heart rate.

I'm still wanting to know how Samsung Galaxy Watch compares to these. Would be nice to get a followup video on that one.

I have an active lifestyle; I walk 3-4 km a day and try to eat healthy. My Apple Watch gave me Low HR notification while sleeping on 2 occasions which led me to goto ER and get ECG, Echocardiogram, and Holter Monitor done. They all turned out to be normal but it increased my anxiety so much that now I have to take anxiety medication and I am really wary of my heart. Apple Watch is a curse for me. I have stopped wearing it.

As a determiner of cardiopulmonary health, recovery rate from stress to rest is far more instructive.

Very interesting results. If you use sleep tracking on the Apple Watch it specifically separates between Resting Heart Rate and Sleeping Heart Rate. And for me these are about 10 beats apart and in ranges that make sense and match my expectations (by measureing myself). So for me it does not seems to factor the time asleep into the calculations for RHR. But maybe the caveat is that you actually have to use the sleep tracking feature.

Don’t wear my watch when I sleep. Easy. Not sure they are as inaccurate as you say. My Garmin if anything puts my rhr higher than what I see rested. It’s tricky to put out content on these trackers. They are regularly updated in software and tweaked to be more accurate and they take time to ‘get to know you’. Comments here suggest they are pretty good. Glad to see you finally got a Garmin.

Why do only a few persons refer to taking pulse rate with your fingers on the artery and a watch? Is this not the most accurate method?

I have a Huawei GT4 and they specifically state that their RHR is HR while at rest during the day. However, when comparing their measured value to the graphs it's obvious that I never reach that low during the day.

As long as it is consistent right? I use resting heart rate as an indicator for how my body is doing when it is trending up or down. Therefore consistency is important. I’m still looking for a video covering consistency/accuracy of wearables.

I thought I have some heart disease when my Apple Watch showed me that my RHR is at 48-55 when just sitting on the computer and when it's at 40 while sleeping. Turns out I'm just fit because I work out a lot and because my maximum heart rate is at 196 bpm based on several tests which is nice for my age of 25. I also never have issues in terms of dizziness or something like that.

Nice to know. Oldie days measured by feel. 15 seconds try a few times see if consistent count. Not convenient but can reasonableness check HRM. Chest straps tend to be more accurate.

I have had to change the alarm level on my fit bit. It was set to alert me if my heart rate was below 50 bpm for 5 minutes. But it does this several times a week. So I have chaged if tro 40 BPM. The nurses suggested that it was because my heart is being slowed down by my medication.

When I was in the height of cross-country season, my Garmin watch said my heart rate was in the 60s while sleeping, but low 40s when awake

Wow, this title didn't come up to be clickbait. Respect for you.

My FitBit and Apple Watch returned similar resting heart rate measurements. A fun observation of my own health is that my resting heart rate is higher when I’m sleep deprived.

Interesting. I use a Apple Watch 7 and do not wear it at night. My resting heart rate that shows on my apple watch is pretty accurate. It shows typically in the 42-47 range. When I am in (college) class, sometimes I just have it on the heart rate app and it will be exactly that. I have done some informal lab tests where I lay down or sit and it will also be in the same range. Laying down typically results in the high 30s while awake and talking.

My Garmin Epix pro has Garmin newest heart rate sensor and seems to be pretty accurate. When I have occasional been to the hospital and had my heart rate measured I have been asked what I do as my resting heart rate is on average 50bpm or lower if I”m really relaxed. So I believe watch is about right

I use Samsung's Gear S3. It shows me around 54 BPM when I first wake up then usually in the upper 70s and 80s when I'm resting during the day.

I used to have a Fitbit with a RHR reading 46 bpm for resting. I’m assuming that might be alright since I’m a runner in college but I have wondered the accuracy of it before.

So issue is that the algorithm for measuring isn't unified and it gives different figures depending on the provider... Hmm that sucks but you can still see the trend, which is arguably more important.

I just look for trends. Especially if it's HRV. It seems to correlate well with how I'm feeling in regards to fatigue and training. But I'm not a doctor and probably have no idea what I'm doing.

An interesting discussion. It wouldn’t be the first time that technology forces the well known medical gold standard to be reevaluated. One thing I would note, not mentioned in the discussion was the impact of “variable” HRV (heart rate variability) which could easily corrupt getting an accurate resting HR. Different manufacturers may well have chosen different algorithms to manage readings. While we all need to be aware of the limitations of these devices, I push back on them as being “fake”

Wearables are never going to be as good as a proper GE/Dynascope/etc machine you'll find in a hospital but for making you aware that your HR is high, your ECG is off etc and you need to seek medical help they've proven to be quite useful in that regard

If my RHR was 35-48 even while awake I'd probably think "hey, that's neat" while also making a doctor's appointment

My heart beat and resting heart rate on my Garmin 945 LTE is almost identical to what my doctor measured every year (3 years in a row). Whatever algorithm they’re using seem to be working

I was always told that resting heart rate should be measured first thing in the morning before getting out of bed. On that basis my Apple Watch, albeit a model from several years ago, is accurate.

I had to wear a Holter monitor for a month to monitor my heart. I wore my Oura (gen 2 and gen 3) rings at the same time. When I would sleep at night, I would get a call every night from whoever is monitoring the Holter monitor, because my heart rate would go below 30 BPM, and it that would set off alarms. Meanwhile, Oura claimed my sleeping heart rate was in the 50s. Garbage products.

I have a Garmin Forerunner 55. I've tested it against taking my pulse upon waking for 60 seconds and it's spot on.

My daytime RHR as defined by 5min of inactivity changes quite a bit thoughout the day, largely dependent on how long since my last meal (previous exercize aside, of course). So in the early afternoon it will typically be at least >10 as compared to late morning.

Very very interesting. Explains why I always thought my resting heart rate was 72/73 before I got a fitbit and it said it was 64.

Garmin algorithm is described on their webpage. They average the lowest 30 minutes and add some percentage to that, can’t remember exactly. It gives me the result in mid 40s, but it’s not unusual for me to see readings in the mid 30s when I drink my morning coffee. A cardiologist gives me 44, as measured at a visit some time ago. In my case the number seems reasonably accurate.

Take phone, Place over heart, record audio for 20sec, trim video to 20sec exact, count beats, multiply by 3, bpm calculated, mine was 57

So I’m hearing that theoretically if you don’t wear your watch while sleeping, your resting heart rate is more accurate through the app? I don’t wear my watch while sleeping and my resting heart rate seems accurate to sitting still for a few minutes

My Sleeping heart rate is about 8-10 beats lower than true resting heart rate ....that is well known phenomenon. It makes sense that sleep heart rate probably is a better measure .....

I’ve been using an Apple Watch since 2021, when I was about 75 lbs heavier than I am now. I looked back at the RHR trend over the last 3 years, and the graph shape looks nearly identical to my weight loss graph. In 2021 my average RHR was near 60, then dropped to below 50 after losing all the weight in 2022, when I also greatly increased my activity levels. As of this year it’s in the low 40’s. I do keep my watch on at night so I know it’s categorizing sleep HR separately. I just checked with a pulse oximeter and the number matches my watch’s real time measurement exactly (44 at the moment). So I have no reason to doubt the accuracy of my Apple Watch’s RHR numbers.

So if I never sleep with my Apple watch on, is the resting heart rate measurement going to be fairly accurate compared to the medical standard resting heart rate while awake?

A good video that Apple should review and consider the implications. They should always be trying to be more accurate with their measurements. I wonder how to send it to Apple

Some people have problems with their heart rate (too high, too low, or variable), so it is a huge disservice not to show them the rhythm as close as possible to the truth. I had a high pulse after an operation and my Xioami watch said I had 78, and I had 120. In general, the older generations had a latency in presenting the pulse variations, but they were close. Today there is no excuse for major differences. It's just a flaw.

the only reason i got my watch was to look at my HR bc I have tachycardia and i can tell when my heart is being a little weird/fast so i check my watch and sure enough my watch is like 160bpm instead of like 90 lol

It makes sense that sleeping heart rate would be a better predictor. If you define resting heart rate as sitting quietly for 5 minutes, that can be anything from sitting quietly in my dark room after my pre-bedtime meditation, to the minutes they left me to wait in the doctor's office waiting to see my doctor. That's a 30 BPM difference on a good day. How can you get a good predictive value from that?

I’m not sure if you intended it to be funny or not, but that “while you’re awake….“ echo made me laugh out loud😂

I love you bro! such useful work you are doing my G

You should do a collaboration with the Quantified Scientist on this topic. He rigorously tests the smart watches.

well my samsung tells me my heart rate is 80bpm, it never really shows me lower - it always says im stressed, guess what i suffer from anxiety :) not sure why u tested those watched and left samsung out

What Fitbit did you use? The google watch takes measurements every second for heart rate so I wonder if that can be more accurate?

My garmin has always been pretty good when compared to traditional "lay down in bed for 5 minutes with the monitor on". I would hardly call it misleading. Taking the absolute minimum it hits during the night is silly for reliability reasons, but I find nightly average coincides very closely with the traditional definition. I am using the top of the line garmin.

My heart rate is 49bpm while watching this video and I average /around 42 - 45bpm while sleeping. Interesting that mine doesn’t get much lower while I sleep haha. I must not sleep very deeply.

My resting heart rate right now is 92 bpm, but my FitBit is telling me my RHR is 70bpm, which I’m guessing was while I was asleep. FitBit tells me that I’m super fit for someone my age (73), which I can tell you is blatantly incorrect. Maybe to get a more accurate RHR I have to NOT wear my FitBit while I sleep?

Interesting, so while I was in the hospital, waiting for surgery, they had me hooked to their hr monitor, and my Apple Watch is exactly the same. Hmmm. 🤔

My Apple Watch gives me a whole set of readings of different things to do with heart rate and cardio generally. If you read the descriptions of how they are measured, they must obviously be guesstimates considering it deduces them all from one sensor on my wrist. But I don't think the absolute value matters that much so much as how it changes with time. If my resting heart rate as measured by my watch is dropping over time, as mine is due to my improved exercise regime, that is surely a good thing.

Great information When I first got my garmin it seemed to use the lowest value over 24 hours, almost always sleeping but I noticed if I had a high number from a bad night of sleep I could drop the number meditating in the afternoon. Now it seems strictly a night time measurement. So I suppose there is some risk of the algorithm changing and moving your numbers up or down. If you are on the edges of either extreme you should do some follow up. It’s been much easier to measure a resting hr after sitting 5 minutes and quickly repeatable in a medical setting rather than having someone monitor you sleeping for 8 hours so it will remain more important. Garmin could easily give you both sleeping and resting hr and maybe they will update it. Last note, I got the garmin blood pressure monitor and it also gives you a hr for the test which you do after sitting still for 5 minutes. Looking forward to the video on blood pressure.

So the takeaway is to target lowering it no matter what system you use

My Fenix 7 gives RHR of 47. It’s roughly that when I’m sleeping. But it’s saying that it is that now, I’m chilling out on here. Did the old count BPM for 15 seconds and multiply by 4 trick and had 48. So mine is pretty accurate, but is it strange that my sleeping heart rate isn’t lower than my awake resting heart rate?

I prefer to subtract my lowest heart rate from my maximum heart rate over the day. Divide this range by 6 to get the standard deviation. My estimate of my RHR is then RHR =Lowest HR + 1 Standard deviation. This is based on the assumption that my heart rate will follow a normal distribution. This way of calculating my RHR is more consistent across the various heart rate monitors that I owned.

Is a low RHR always a sign of fitness? Could you have a higher RHR than someone but still be considered fitter?

This is because these mean is notorious to discard information about the true distribution of. In fact it’s because unless you’re only doing one thing a day then the distribution of the random var HR won’t be normally distributed. A better model than y ~ a + N(0, sigma) is needed

About 10-15 years ago I measured my resting HR every day to see if I was getting enough rest between running sessions. I also measured in a few different ways just to see how different factors affected it. If I took my HR lying in bed first thing in the morning I had about 5 bpm lower compared to if I got up from bed and wa sitting in a chair. If I measured standing up it was another 5 bpm higher. If I measured later in the day the value would be a bit higher (don't remeber how much off the top of my head) and if I measured relatively close after eating the HR would also be a bit elevated. So lots of ways to get the numbers mixed up. Doesn't really matter how you do it. But you have to do it the same way every day if you want useful data.

...to your other point though. In my data sets I have markers for which period is being measured by which device because like so many things we aren't really getting apples for apples comparisons across manufacturers of biometric data like this.

I'm glad for this video because I'm not super fit and my Garmin says something about 46 bpm and I was starting to think I have a heart problem. when I sit still it is more reasonable about 55.

I am not superfit but each time I go to the doctor they sound quite surprised that I have a low resting heart rate of around 41.

same values here, its probably better to set your own baseline and correlate changes with that, in Apple watch HRV measurement is even worse

You can check your resting heart rate during the day by stopping for a bit and checking. You can’t check your watch during the night, or during your absolute resting heart rate, so that’s why you should get that reading.

Just watched this video again today, and made me think about some of the different apps used to measure sleep. They very rarely guess when I'm *actually* sleeping. So the "numbers" are not really to be trusted per-se. But what I think is helpful is that the comparison is helpful! For instance, if one night is stays I slept more than the previous one, that is generally true, even if the actual number of hours are not accurate. So my question is... if we take those numbers as "arbitrary units", rather than "bpm" or "hours of sleep", is that more accurate/helpful to help people getting in better shape?

Tracking resting pulse can be useful. I'm not sure averages and algorithms help you much though. I've read that resting heart rate can be genetically low or high but a change toward lower can be an indicator of fitness. Drastic changes should be run by a doctor though as it can go down or up for non-fitness related reasons. Comparisons between individuals may not work even though many want this. It may be an internally consistent metric though. Many want it to be an externally consistent metric. This is a lot like most bodyfat testing, you might actually have a 3% drop but you might not really know your actual bodyfat percentage with the same accuracy as your estimation of drop. This is true of a lot of "home" testing.

I see the RHR value from my watch as a not totally accurate benchmark. I can see from one day to the next what is going on, even if I don't take the exact number as gospel. For instance it suggests early to mid 50s but the two times I got covid that daily average crept into the early 60s. Took a few weeks and ramped up then back down. Both times the same.

My problem is that my watch reports heart rates almost double the measured rate. This leads to double the kcal usage predicted by Google Fit. From my own research, Google Fit is outputting a good estimate based on heart rate. But the heart rate measured by my watch sensors is too high.

fwiw: have both oura v3 and fitbit. Oura for sleep and Fitbit mostly for the day Don't really care about the number but the trend. When the trend changes I look for a reason.. Lowest night time heart rate, hrv, temp and resparation are my goto. usually alcohol, stress, infection or exercise. During the day both mostly agree closely somewhere in the 60s.

interesting, as I wear a basic Garmin watch, and its RHR number is consistently 7-10bpm above my measured RHR of either relaxing for 5 min or upon waking.

I only wear my watch during the day and the resting heart rate is still in the 40s, according to the Garmin watch.

Using my sleeping resting heart rate, max heart rate by age and the Karvonen method helped me calculate my most accurate training zones, especially Zone 2. Far from perfect to use this method or train by heart rate zones, but it worked fairly well.

Thanks for your vids. They're interesting but I find them a bit confusing and unstructured and hard to follow...

Doesn't matter what the number is. It's for tracking trends.

My Garmin Fenix 7 says my 7 day avg RHR is 46 bpm... My current heart rate while watching this video fluctuated between 42-55 bpm so it would seem it is kinda accurate? (Garmin that is) I guess even if not, it would still show trends?

When I exercised a lot, I found my resting heart rate could be lower than my sleeping heart rate, in fact it could be 100-110 for hours while sleeping, only dropping to 60 and below just as I was to wake up. I got down to 41 BPM in the mornings, just sitting at the computer with some coffee. But stop excercising and resting hear rate goes up quickly, after a few months I can't get below 60 anymore. I was doing about 10 hours of cardio per week before, always on the edge between aerobic and anaerobic. My watch just reports the actual measured heart rate, once per hour or when I check manually.

Andrew, where do I find table of recovery heart rate? Which value would show that I’m in the clear? And is it also affected by age or is it affected by the peak value that occurred during exercising?

There is a name for slow heart rate - Bradycardia - it can be a serious medical condition and in some cases may require the fitting of a pacemaker. I’m surprised the doc didn’t add the caveat that a slow heart rate isn’t always indicative of an individuals fitness level.

I use Garmin Fenix 7 and according to it my rhr is 45 - 50 but I know it's while sleeping. I work in a hospital and one of the nurses was scared when she saw my rhr and made me wear an ECG holter monitor for 24 hours. My minimum was 30 BPM (at 3:11 AM), maximum 202 (during a 10k run), 24 h average 62 BPM and average while asleep 49. Please don't start yapping about how low and high my HR is because the best cardiologist in my town did my report summary and told me I'm great... The point is it's hard to standardise something big corporations are so secretive about.

I like the fact you have taken toy watches ie Apple, Fitbit Mi-Band etc to use as your sample data set, good drills. I know from having used all of them that what the present is saying is not strickly true, each device has difference quality of hardware & comparing to a Chect Strap is an unfair & bias test / comparrison.

When do people have heart attacks? Do they have them when their heart rate is low during sleep or when their heart rate is high while shoveling snow?