Smartwatch accuracy warning
- Reporters say consumer smartwatches often give unreliable readings for steps, calories, heart rate, sleep, and recovery. (timesnownews.com) - The Times Now piece lists those specific metrics as commonly inaccurate across devices. (timesnownews.com) - The Journal adds wearables are reshaping why people move, so users should treat device numbers as directional rather than clinical. (thejournal.ie)
A smartwatch can spot useful trends, but its numbers for steps, calories, heart rate, sleep and “recovery” are often estimates, not measurements. (link.springer.com) That gap starts with how the devices work. Most watches use motion sensors for activity and a light-based wrist sensor called photoplethysmography, or PPG, to estimate pulse from changes in blood flow under the skin. (ahajournals.org) A 2024 umbrella review in *Sports Medicine* pulled together 24 systematic reviews, 249 validation studies and 430,465 participants. It found wearables showed about ±3% mean bias for heart rate, but step counts were often off by roughly -9% to 12%, and energy expenditure errors ranged far wider. (link.springer.com) Heart rate can drift furthest when the stakes rise. An American College of Cardiology summary of a 2024 *Journal of the American College of Cardiology* study said six commercial devices differed from electrocardiogram readings by 13.8 beats per minute at peak exercise in people with normal rhythm and 28.7 beats per minute in people with atrial fibrillation. (acc.org) Sleep data has a similar problem: the watch usually infers sleep from stillness, not brain activity. Johns Hopkins Medicine said exact sleep staging still requires a medical sleep study that measures brain waves, because consumer trackers are using inactivity as a stand-in. (hopkinsmedicine.org) “Recovery” scores add another layer of guesswork because they combine several indirect signals, often including sleep, resting heart rate and heart rate variability. If the inputs are estimates, the final readiness score is an estimate too. (ahajournals.org) (hopkinsmedicine.org) The evidence base is still thin relative to the number of products on wrists. The same 2024 review found only about 11% of commercially available wearables had been validated for at least one biometric outcome, and the published studies covered only 3.5% of the testing needed for a comprehensive check. (link.springer.com) Researchers and clinicians are not dismissing wearables outright. The American College of Cardiology said trend data over time may still be useful for some purposes, even when single readings are not precise enough for clinical monitoring during exercise. (acc.org) That is why the safest way to read a smartwatch is as a directional tool. If your resting heart rate rises for a week, your sleep time drops for three nights, or your step count falls sharply, the pattern may be worth noticing even if the exact number is not. (acc.org) (link.springer.com)
