Wearables: How Accurate Are They? — Part 2

July 15, 2020 - 8 minutes read

IoT app development

Thousands of people around the world have taken a more proactive approach to their health thanks to wearables. Through a combination of the Internet of Things (IoT), artificial intelligence (AI), and health app development, these devices allow users to monitor their steps taken, calories burned, heart rates, and more. But just how accurate are they?

In a recent post, we examined how wearables actually “count” your steps and why this metric can be way off the mark. If you missed this article, you can read it here. For this entry, we’ll explore how wearables track heart and calorie burn rates. We’ll also delve into some of the factors that can make these readings inaccurate.

Why Your Heart Rate Readings May Be Haywire

Ever heard of photoplethysmography (PPG)? Don’t fret if you haven’t — it’s an obscure term outside of medicine (and a mouthful to pronounce). PPG refers to the technique that most wearable devices use to measure heart rate.

By beaming a green LED into your wrist, PPG enables wearables to measure your blood volume. Every time your heart beats, more blood flows into your blood vessels. A proportional amount of the green LED’s light is absorbed by this blood. In between heart beats, your blood recedes away from your vessels, causing less light to be absorbed. Wearable devices use this difference in light absorption to calculate your heart rate.

IoT app development

This is certainly one of the most elegant uses of IoT development and sensors in healthcare. But it’s not without its faults. In fact, green LED sensors can be quite unreliable. To measure blood volume properly, the green light must penetrate the skin. But several studies have found that melanated skin is actually more likely to absorb it.

It’s important to note that this is still a hotly-debated topic. For instance, one study didn’t identify any correlation between skin color and accuracy. But it did report an activity error rate 30% higher than when at rest. And other research found that the Apple Watch performed quite well, regardless of skin tone.

Honing in on Your Heart Beat

To address this potential issue, Fitbit works hard on calibrating its device sensors to work accurately for everybody; the company says that the green light its wearables emit are sufficiently strong enough to penetrate through darker skin, and its sensor is receptive enough to accurately detect heart rate signals. But many users with melanated skin have complained that wearables from the San Francisco developer either gave wrong readings or didn’t work at all.

Mikael Mattsson is a Senior Research at Swedish university Karolinska Institutet. He explains that, in research settings, scientists typically calibrate and use seven different light wavelengths to capture accurate results. But fitting all of these capabilities in a small wearable is impossible.

To circumvent this, the Apple Watch doesn’t only rely on green LEDs to measure your resting heart rate; it also employs infrared sensors. While more reliable and accurate than green light, Mattsson says that movement can still cause this to measure vitals incorrectly. Apple seems to be well-aware of this issue; the Apple Watch Series 4 measures your heart’s electrical current directly via electrodes instead of relying on proxy blood level measurements.

IoT app development

But even with more bells and whistles, the accuracy of wrist-worn wearables can vary depending on the type of actions you’re doing. Stable, repetitive activities such as stationary biking are usually fine. But even other relatively controlled exercises like using an elliptical machine with arm levers can throw your wearables a curveball.

For this reason, Mattsson hasn’t even begun to test wearables outdoors yet. Why? “If they’re not good enough indoors, they won’t be good enough outdoors,” he explains. The lesson here is this: Wearables’ heart rate monitor can adequately function under specific conditions and are completely fine for casual use. Just know that they’re still not perfect.

These Same Inaccuracies Make Calorie Burn Rate Unreliable

If you were holding out hope that wearables track your calorie burn rate accurately, we’ve got some bad news from you. Many of these devices utilize proprietary algorithms to calculate your energy expenditure — and these algorithms factor in movement and heart rate measurements (which we all know by now can be quite erroneous).

To make calorie burn estimations more precise, numerous wearables take physical aspects such as height, weight, sex, and age into account. This enables the devices to calculate your basal metabolic rate, the usual number of calories you’d normally burn each day. But throw in the calories expended from physical exertion, and the whole equation goes awry. That’s because this is the most unreliable metric that wearables calculate.

IoT app development

In a study of a diverse set of people of different skin tones, fitness levels, weights, heights, and ages, Mattsson and his colleagues found that all wearables had a 20% or higher error rate for calorie burn calculations. Other studies have reported similar inaccuracies.

Mattsson believes the main problem lies in the proxy measurements and algorithms. The former gives the latter bad measurements to use in its calculations. But algorithms are already doomed from the start. “The biggest problem is that they’ve done the algorithms for a subset of people,” Mattsson says. “In most studies, you talk about white males in their 30s at an average fitness level. The farther away you get, the bigger the risk of a problem.”

The Catch-22 Facing Wearables

These problems of inaccurate readings put wearables in a perplexing conundrum. To improve their offerings, wearable developers need a diverse group of people purchasing and using their products. This, in turn, would give the developer more diverse data to train the algorithms with. But here’s the catch: People would have to use these devices, even if they’re not perfect.

IoT app development

Like before, the main takeaway from all of this is that wearable data must be taken in context. They can certainly inform and provide a person with a better picture of his or her health. Numerous people around the world have turned their lives around thanks to wearables. But, as with any new technology, fallibility is still present.

Tags: , , , , , , , , , , , , , , , , ,