​Optical HR accuracy: The experts speak

We ask the experts what they think of optical heart rate technology
​Optical heart rate tech: Experts speak

A couple of weeks ago I wrote a piece born of frustration with the current crop of optical heart rate sensors. After we reviewed the Garmin Vivosmart HR, Forerunner 235 and Polar A360, it wasn't a surprise that they weren't as accurate as chest straps – but I felt brands weren't being upfront about the shortcomings of sensors.

Read the original: We need more honesty in heart rate tech

Once the article was published, a host of companies got in touch wanting to add their two cents to the issue. And I was interested to find out from the experts whether wrist monitors could ever be accurate, whether the data from them is really junk, and if people are being misled?

I caught up with three industry experts, including the inventor of the optical heart rate sensor, to get their takes on the matter.

Dave Wright - MyZONE

CEO and creator

"The whole optical blood flow thing is great if all you're doing is walking or running. And if that's all you do, they're not too bad.

"But when you start building a live display into ecosystems and you're guiding your training by it, that's when you can't afford to have any delay. Consumer demand for credible data is so high now, they don't expect anything less.

Essential reading: Fitbit Blaze review

"That's the thing with a lot of these optical blood flow wearables. If you're just walking and running then they have good algorithms to get rid of light and movement artefacts, but if the intensity is a lot higher, above 160 beats, then the blood passing is so fast that when you add movement as well, it becomes really difficult to get the right reading.

"It's physiology not technology. It's a matter of the unpredictability of the blood through the wrist in an area you can't control. Ear tech is going to beat the wrist because it suffers less physiological interference.

"Anything more than a one or two second delay isn't acceptable. The reason being that HIIT sessions are about pushing people to the limit. If you're doing 10-20 second intervals and there's a delay of 10 seconds, it's beyond unacceptable.

"Not naming names, but one company have been desperate for us to use their device. I used it for a standard workout, doing burpees and it was taking 1 min and 25 seconds before it caught up with MyZONE. I deemed the data irrelevant.

"Where Fitbit have got into a little bit of trouble is because they're saying they're not a medical device and they say that it's not capable of doing high intensity activity, but the advertising shows exactly the activities that they know aren't accurate."

Chris Eschbach – Valencell

Director of exercise science & clinical trials

"It's in our interest to do detailed validation and testing of all optical heart rate devices on the market. Accuracy is heavily dependent on where the sensors are located and what the person is doing. The wrist has less tissue, more bones, more tendons and if there's not great blood flow then accuracy goes down.

"We took two chest straps (chest straps are awesome for accuracy because they use electrical signals but we put electro gel on the back for good conduction) and we put the chest straps on a group of 20 people. They performed a test on our treadmill with a mix of walking, running, walking, standing and running at a higher speed for dynamically changing heart rate and step rates.

"When we analysed the chest strap data we're getting about 91% of all the data within +/- 5 BPM of each other. Now if we move to earbuds and arm bands they're nearly there too, around 91% accurate compared to a chest strap between +/- 5 beats. Our wrist tech was around 85% within that +/- 5 beat range when running.

"When you go to other activities, it's a whole different ball game. During cross fit or weight lifting, the head is a great location, good blood flow, stable and a good fit, as is the upper arm. The wrist can completely crap out. If you're doing pull-ups or arm curls, they may not be measuring at all.

"We're getting better. It will be equally as good as the chest strap in certain situations, but there will be weaknesses it has and others don't.

"When you're measuring heart rate I want to be within that five beat window to be really meaningful. If we're using heart rate to measure intensity levels or an input for calorie expenditure, or how much training has affected you, you're either measuring it accurately (within +/- 5 beats) or you're not. If it's not, then let the user know the data is worthless. And that's not the case right now. No-one is telling users if the data is horrible, and they should."

Liz Dickinson – Mio Global

Mio CEO and inventor of optical heart rate tech

"For us the most important thing was getting rid of the chest strap. Over time optical will become good enough if not perfect. It's not quite there yet, but it's a lot better than when I invented it a few years ago when it couldn't even do continuous sensing.

"Its usefulness right now depends on what you want to use data for. I would say that optical is useful for running and cycling and some are useful for daily life – but nothing out there today is at the level of the chest strap. That's the fact, but I don't think anyone is saying that's not the case. What we're saying is that if you don't want to wear a chest strap then there is an alternative.

"The more athletic you are the more likely you are to wear a chest strap. What we're trying to do is to democratise access to HR training and the only way to do that is to get rid of the chest strap. We want people to have a great experience when they're working out and we believe heart rate is the most important thing for people to understand in health and fitness.

"In terms of honesty, I haven't reviewed marketing literature of other brands to say whether they're misleading people or not. But you have to be careful. The tech works differently on different people, because body fat or blood diffusion isn't the same as other people."

Wareable verdict

So what have we learned from talking to the experts of heart rate tech? Well firstly, if you're serious about your sport and data today, then steer clear of the wrist monitor. It's a great way for entry level runners and fitness fans to take control of their plans, but for those using the data to train, it's not up to task.

And while runners may one day be able to use optical tech for training sessions, when lag times and algorithms improve, for gym bunnies and HIIT fans it looks unlikely that the wrist will ever be a serious option.

But there does seem to be a light at the end of the tunnel. Our experts agreed that the ear is going to become a key place for future sensors, and it's a place we're already comfortable with placing tech. Hearables here we come.

13 Comments

  • TheWerewolf says:

    So... to sum up - if you're a 'serious athelete' - which is a *tiny* fraction of even the people who actually exercise regularly - which itself is a *tiny* fraction of everyone... then don't use a consumer fitness (and - note.. *fitness*) tracker.

    Got it.

    Personally, I would have thought that was kind of obvious.

    Mind you - if you're the kind of person who's taking this all so seriously that you're willing to strap into a complex (and one expects, expensive) array of electronics for it.. I think there are bigger issues than HR accuracy to worry about in your life.

    • Jonnyg says:

      You would not buy a watch that inaccurately measures time so why buy a heart rate monitor that inaccuracy tracks heart rate. You can afford to miss a few seconds, surely accuracy around HR is quite the opposite. Go buy a sundial

    • j.stables says:

      The point is that the latest Garmin Forerunner 235 (to make a single example), which has optical, is not a lowest common denominator 'fitness tracker'. It's a fitness focussed device, that falls short of expectations with its live HR readouts. It's a great watch for a lot of mid-level runners and offers a good bpm guide, but it's not up to the chest strap. Yet.

  • VictorVMan says:

    I train only new and non-exercisers where data heart rate monitoring accuracy and reliability are critical. They are at high risk for heart disease, with no base of conditioning. We use Polar heart rate monitors on them every single training session, due to risk, what training intensity they can train safely, what their move to rest ratios are and how they respond to each exercise, machine and area or environment of the club. No guessing. Also to know what energy system we are going to focus on, what method and what guidelines to follow.

    I really like the Mio and Fitbit units, and so do they, but we have the same issues discussed in the article.

  • sallyedwards says:

    For the past 40 years, I have trained with a heart rate monitor and raced professionally for most of those years. Polar stated that their chest EKG transmitter was 98-99% accurate and at the time, I felt that was important. But, I no longer believe that we need that accuracy, classified as "medical grade". Rather, we need people using heart rate "sensors" that they can comfortably enjoy and connect without the discomfort, having to wet the pads on the back of the belt, and not having to "lift up their shirt" - which for many women is embarrassing. I am willing to sacrifice medical grade accuracy for getting the every body to start to understand the I-word which is intensity. I like Liz (Mio) and Mike (MyZone) and Chris (Valencell) want everyone for every work out to use a heart rate sensor. Some error is acceptable large amounts of error is unacceptable - and with 4 decades of experience I'll take plus/minus 5 BPM to wear a wrist or arm band versus a chest strap that often drops the signal, cross talks, and can hurt. The day of the chest transmitter belt is over. Don’t every purchase one again.

    SALLY EDWARDS, CEO, Heart Zones, Inc. and Triathlon Hall of Fame - member.

  • JBerkow says:

    Hi: I come from the medical device industry and have spent the last decade using the photoplethysmograph to derive parameters that capture the various cardiovascular (CV)compensatory mechanisms as they relate to maintaining stability in sick patients.  In general, stability can be defined as the continuous adaptation to varying forms of stress to enable adequate oxygen delivery to tissues and organs. In the medical industry, the goal is to enable recognition of when the CV system is heading towards instability where sick individuals have compromised compensatory mechanisms--e.g. poor reserve.  If BP is used for this purpose, a drop below 90 mmHg systolic is indicative that the compensatory mechanisms have become exhausted.  So vitals really offer little value here.

    Question: if we could clearly show which of the compensatory contributors was not working efficiently, would this increase demand for more precise measures?  For example, we know that after 60 years of age, vascular stiffness increases significantly where the vasculature is actually the primary compensatory mechanism (not the heart).  If the consumer could better understand how to preserve vascular health by monitoring it with this type of device, would there be market interest?

    I personally feel that consumer products will converge with the medical self-care products as monetary incentives by employers and payers become more common.  As long as measures are validated but not used for a diagnosis or treatment, we can make these products much cheaper as they do not need to have FDA clearance.

    Jan Berkow, CEO ChironTec (previously co-founder, CTO, and inventor Intelomed)

  • JackKessler says:

    The people who sell these things are not just selling worthless junk, there is an excellent chance they have given a number of people heart attacks and even killed a few.  If you are a 50 year-old overweight white male, your exercise heart rate should be around 80% of 220 minus your age.  (220-50) x 0.8 = 136.  If your heart rate monitor is 15% inaccurate, and many of them are at least that inaccurate, while the heart rate monitor you bought to protect you against a heart attack is reading a safe 136, your actual hear rate might be as high as 156, 92% of your maximum.  92% of maximum is a heart rate that should only be maintained by toned elite athletes.  There is an excellent chance that if the customer will get himself a heart attack or worse if he believes the marketing literature.  

    An example of just how dishonest these marketing swine are is Liz Dickinson who is smiling an lying in the picture in this article.  "In terms of honesty, I haven't reviewed marketing literature of other brands to say whether they're misleading people or not."  Really, Liz?  You're the CEO of one of these companies and you haven't looked at the literature of your competitors' products?  Really?  You have the gall to smile and say that?  You are a liar, Liz. 

    • ThaVix says:

      This is not correct. You can run at 100% of your Fcmax and the point is that most people don't know what it is and there's not a perfect formula to find it. There are tests. Anyway as I said you can run at 100% of your heart and if you are sane there's no risk because heart has system protections. Heart attack often happens while sleeping or after a run. So the Hr monitor is not meant to avoid heart attack. It's not truth that is dangerous going at 90% of fcmax. The hr monitor should be used for specific train based on threshold and goals you want to achieve. 

  • dman4me says:

    Ok. So why doesn't anybody mention the scosche armband. In my tests with a chest-strap and the machines hand grips, it is accurate within +- 1 or 2 BPM. You were this armband just below the elbow.

    Also, what about the spree cap/headband? I have not bought it yet, but read it is reasonably accurate.

    Let us know.

  • kingamesaros says:

    My Fitbit's HR monitor stops working when I start sweating, that's about 3-4 minutes after I start training. I do high intensity and many times the maximum heart rate shown is 140 when I know it is 170+ (as sometimes works...). I think it's at least 20% off in more than 50% of the cases, so at this point it is pretty useless.

    But when I get really dissapointed is when I went climbing on a vulcano. Totally 12 hours+ of climbing. I was so eager to get my data. For the first 2 hours it showed my heart was at maximum or close to it 80% of the time. But then something happened... my heart was measured at 110-120bpm all the time!!! Wow, I feel like superman! So I think in that case my Fitbit was completely off for several hours...

  • Tomas says:

    I count my wrist pulse with fingers for ten seconds and multiply by 6. Doesn't cost much, unaltered by sweat, environmentally friendly and pretty accurate. 

  • Tone says:

    Back in the late 1970's our lifesaving club had a battery operated pulse meter with an optical sensor that you clipped on your finger. I was 16 at the time, Liz must have been 3 when she invented the technology. 

  • NeillMyers says:

    Chest strap technology doesn't work for everyone.  I have tried them all and NONE work for me.  The chest straps rely on measuring the  electrical impulses of the heart and if your heart has an unusual pattern (as does mine) they won't work correctly.

    I now use the Scosche HR monitor.  I strap it to my upper arm and use it for my bike and run training for triathlon.  It is rock solid in terms of results.

    The comment about a 1-2 second delay being unacceptable for HIIT was overblown, IMHO.  When you do HIIT you push as hard as you can for each interval.  I cannot imagine watching the HR so closely that a 1 second delay would have any impact at all.  If you have the mental capacity to watch that closely you simply aren't pushing hard enough!

What do you think?

Connect with Facebook, Twitter, or just enter your email to sign in and comment.