Wearables at work: what's accurate today and what's next

Wearables have moved from fitness gadgets to serious data sources. But they're not medical devices, and accuracy varies by device and by measure. Here's an honest view.

What wearables measure well

Heart rate at rest. Optical sensors on the wrist or finger are generally reliable for resting heart rate, especially overnight when the body is still.

Total sleep time and timing. Consumer devices are reasonably good at estimating when someone fell asleep, when they woke and how long they slept.

Daily activity. Step counts and general activity levels are useful for trends, even if absolute numbers vary between devices.

HRV trends. Overnight HRV from good devices is useful for tracking change over time within the same person, which is how we use it.

Where they're weaker

Sleep stages. Classifying light, deep and REM sleep is much harder. Schyvens and colleagues (2024) compared consumer sleep trackers against polysomnography, the clinical standard, and found the devices performed better at distinguishing sleep from wake than at identifying individual sleep stages.

Readings during movement. Optical heart rate is less reliable during intense or irregular movement.

Comparisons between devices. Different devices use different algorithms. A reading from one brand isn't directly comparable with another.

How Everis handles this

We design around what wearables do well:

• Personal baselines. Each person is compared with their own data from their own device, so differences between brands matter far less.
• Trends, not single readings. We look for sustained change, which is robust to the noise in any one night.
• Data quality controls. Minimum wear-time thresholds and artefact filtering exclude unreliable data.
• Multiple signals. No single measure drives a flag. Physiology is combined with validated surveys.
• Professional review. Physiologists check flagged signals before any action.

Through our partner Penng, employees can connect a wearable they already own via Apple Health or Health Connect, or use a Penng band. Available measurements depend on the device.

What's next

Sensors keep improving, battery life is extending, and more devices are measuring signals such as blood oxygen and blood pressure. Validation research is growing. Over time, the gap between consumer and clinical devices for key measures is likely to narrow.

What this means for your organisation

Wearables are accurate enough today to give an early, continuous view of strain and recovery, as long as the data is read honestly and in context. That's the difference between a gadget and a workforce health tool.

Sources: Schyvens AM et al. (2024). Full references on our Science page.