Consumer Data vs Clinical Measurements

Direct measurements, estimates and scores

Consumer devices produce three kinds of output, and they deserve different levels of trust.

A direct reading can still be wrong, and an estimate or score has additional layers of assumption. No proprietary score has a universal clinical meaning, and there is no single ideal value of a score for everyone. See Algorithm-Derived Scores Explained.

"Medical grade" is a claim, not a standard you can check

Under the FDA's January 2026 general wellness guidance, a product is outside the wellness policy if its labeling or marketing claims things like clinical accuracy, "medical or clinical grade," or substitution for an FDA-authorized device. [1] The guidance also says that outputs which mimic values used clinically should be validated, for example through manufacturer testing or peer-reviewed clinical literature, to reflect those values. [1]

In practice, check the specific feature, the exact device model, the country and the official record. Clearance or authorization of one function does not make every other reading from the same device clinical. This is US information only; other regulators differ. See Wellness Optimization vs Medical Treatment.

Accuracy varies by measurement and model

Validation studies compare a device with a reference measure. A meta-analysis of 52 studies of recent Fitbit models found that the devices on average underestimated heart rate, energy expenditure and steps compared with criterion measures. [2] The pooled underestimate was about 3 beats per minute for heart rate and about 2.8 kcal per minute for energy expenditure, with wide limits of agreement. [2] Its authors concluded this can be acceptable on average for steps and heart rate, but energy expenditure may be inaccurate for some research purposes, and results varied by study quality, age, activity type and Fitbit model. [2]

Another study compared a consumer tracker with a research-grade actigraph in 25 people with insomnia. The devices generally did not differ significantly in sleep estimates, yet the authors still advised caution because consumer technology is "still in its infancy." [3]

Both findings apply to the devices, people and conditions studied. Do not carry them to other models or other groups. See Device vs Clinical Reference: How Validation Works.

Context changes the number

Clinical readings follow set conditions. For example, MedlinePlus says that to find a resting heart rate you must have been resting for at least 10 minutes. [4] A wearable may record at many times of day and activity. Sleep measures also vary from night to night in clinical settings; a review describes substantial night-to-night variability in sleep apnea severity as raising misdiagnosis concerns with single-night testing. [5]

Accurate does not mean beneficial

A device can measure a number well and still not help your health. In a randomized trial of 471 young adults, adding a wearable to a weight-loss program led to less weight loss at 24 months than self-monitoring on a website. [6]

What to do with consumer data

Treat it as a conversation starter, not a diagnosis. If a reading worries you, or you have symptoms, get a clinical assessment. See Discussing Wearable Data With a Clinician.