Smartwatches and health: what they can measure—and what they can’t prove

Watches and wristbands can collect data such as heart rate, activity and sleep, then present it in an app or in summaries of trends. That can make them useful tools for everyday tracking. It does not turn every figure into a medical assessment, nor does it mean every device measures the same things or uses an identical method.

The useful question is not just “What number does it show?” but which specific feature produced it, under what conditions it was recorded, and what the feature is described as doing. A reading on its own can be difficult to interpret without context: the time it was recorded, the activity taking place and how the device fits may all matter. It is worth distinguishing what the device records from what a person—or an app—infer from that record.

Some watches estimate heart rate at the wrist using optical sensors. Polar’s documentation for the Vantage M2 describes a measurement that combines optical reading with skin-contact detection, and recommends wearing the watch snugly and in a specified position for measurements during exercise or sleep. The technique and conditions of use are part of what the data means; they are not interchangeable details.

This does not allow us to conclude that every brand uses the same system or that a particular reading is accurate in every situation. The manufacturer describes a procedure for its specific product. So when checking a watch, look up the manual for the exact model and find how it defines the feature, when it records, and what placement instructions it gives. If you switch devices, do not assume that the same metric name implies the same method.

A device may offer features intended to support habits and wellness—for example, recording activity or showing a trend—without those features being intended to diagnose a disease. The distinction depends on the product’s stated purpose and claims about its use, not simply on whether it uses sensors or displays results that look precise.

The US FDA publishes guidance on low-risk general wellness products; the legal analyses included in the research describe a policy focused on those features and on distinguishing them from medical products. That framework applies to the United States: it should not be presented as a universal rule or automatically extrapolated to regulations in other countries. For someone buying or using a wearable, the practical takeaway is straightforward: check the official instructions for each feature, and do not treat a “health” label as proof that the device is authorized to diagnose or treat.

The manufacturer’s documentation is the starting point for finding out what a watch offers, but read the information for the relevant feature and model rather than relying on a marketing name. Apple’s manuals, for example, bring together Apple Watch usage information and explain health features. Google’s help pages describe heart-rate tracking on Pixel Watch or Fitbit devices. These sources are useful for learning the conditions and scope each provider states; they are not independent comparisons between brands.

Before relying on a measurement, look for answers to these questions in the manual, official help pages or feature description:

• What is measured or estimated: a direct reading, an estimate, or an indicator calculated from other data.
• When it is obtained: continuously, during a session, or only when a feature is activated.
• What conditions it requires: position, fit, rest, activity or settings.
• What limitations the provider states: uses not intended, circumstances that may affect the reading, or whether a health professional should be consulted.

A wrist measurement is not, by itself, equivalent to a clinical assessment. Polar’s documentation emphasizes correct placement of the Vantage M2 for its heart-rate measurements. This supports the conclusion that fit is relevant to the use intended by that manufacturer, but it does not quantify how much a reading may vary from person to person or in different circumstances.

The difference between a measurement and a derived indicator also matters. A figure that the watch presents as a summary or trend may depend on software calculations, earlier records or criteria specific to the product. Without validated information about the method and its performance in the relevant population and context, it is not prudent to turn that figure into a clinical conclusion. A trend can help you frame a question; it is not enough to answer it.

The most reasonable value of a wearable is usually in observing everyday patterns or supporting activity goals, not replacing healthcare. If a reading is worrying, keeps recurring or seems to conflict with how someone feels, the screen alone cannot determine which signal is correct. For a health decision, consult a professional and explain which device and feature produced the data.

A quick check before making decisions can prevent misunderstandings:

1. Identify the exact model and feature; do not generalize from another watch.
2. Consult the manufacturer’s manual and follow its instructions for fit and use.
3. Distinguish a measurement from an app calculation or summary.
4. Check whether the feature is described as wellness-related or as having a medical purpose, and consider the regulations that apply in your country.
5. Do not start, stop or change treatment based only on a wearable metric.