How each smartwatch ecosystem exposes data
The access route depends on the watch. There is no single smartwatch API, and some routes require your own mobile app while others are server to server.
Apple Watch
Data syncs to HealthKit on the paired iPhone. HealthKit is on-device only, with no Apple cloud API, so an iOS app must request per-type read permission and upload data to your backend.
Wear OS and Galaxy Watch
Data can be read on Android through Health Connect, which uses per-type permissions and requires a Play Console health permissions declaration. Samsung's health SDKs expose additional data and may require partner approval.
Garmin
The Garmin Health API delivers data from Garmin Connect to your server after the user authorizes access. It requires an approved business developer application.
Fitbit
The Fitbit Web API uses OAuth 2.0. Intraday data such as minute-level heart rate for server applications requires approval from Fitbit.
Data types and what to expect
Smartwatches report a broad set of measurements, but availability varies by model, region, firmware and user settings. Heart rate is near universal. Heart rate variability is reported differently across vendors: Apple reports SDNN, while other vendors may report RMSSD-based values or a proprietary score, so values are not directly comparable across devices.
SpO2 depends on model and region. ECG is available on some watches: on Apple Watch, recorded ECGs are readable through HealthKit as HKElectrocardiogram samples, including classification and voltage data, when the user grants permission. Irregular rhythm notifications, sleep stages and activity metrics are also available on several platforms, each with its own definitions.
- Heart rate: continuous, resting and workout
- HRV: metric type recorded alongside the value
- SpO2: where supported by device and region
- ECG: Apple Watch via HealthKit with user permission
- Irregular rhythm notification events
- Sleep stages and duration, steps, active energy, workouts
Clinical use of smartwatch data
Consumer smartwatch data is generally not a substitute for clinical-grade measurement. Some specific features, such as certain ECG and irregular rhythm functions, have their own regulatory clearances for defined uses, but the broader data stream, including continuous heart rate, SpO2 and sleep, is designed for wellness and trend tracking. Algorithms also change with software updates.
Any clinical decision based on smartwatch data requires appropriate validation for that use, and your clinical leadership should define where it fits. In practice, smartwatch data works well for engagement, longitudinal trends and research cohorts, and as context alongside validated measurements in remote patient monitoring programs. We label device type, model and source on every value so clinicians can judge it accordingly.
Normalizing smartwatch data
The same physiological concept arrives in different shapes from each watch. We map every source into a canonical model with UCUM units, timezone-aware timestamps, the original vendor identifiers and the exact metric definition, such as which HRV calculation was used.
Duplicate data is common. A Galaxy Watch user may sync to Samsung Health and Health Connect; a Fitbit user on iPhone may also write steps to HealthKit. We record each sample's originating app and device and apply documented source-priority rules so totals are not double counted. When data goes to an EHR, measurement types are mapped to LOINC and sent as FHIR Observations.
Permissions, privacy and background sync
Smartwatch platforms grant permissions per data type, and users can change them at any time. HealthKit does not tell an app whether read access was denied, so the app sees no data rather than an error; the integration has to handle that gracefully and avoid treating it as a gap in monitoring. Health Connect permissions can likewise be revoked from system settings.
On-device sync also depends on background execution limits on iOS and Android, which affects how quickly new data reaches your server. We design sync to use the platform's background delivery and change-tracking mechanisms and to catch up reliably when the app is next opened. Pipelines use HIPAA-aligned controls; compliance depends on your overall program, policies and BAAs. See HIPAA-aligned wearable integration.
Frequently asked questions
Can we read Apple Watch ECG recordings?
Yes, with the user's permission. ECG recordings taken on Apple Watch are available in HealthKit as HKElectrocardiogram samples, including the classification and voltage measurements, and an iOS app can read them and upload them to your backend. There is no Apple cloud API for this data.
How do we get Galaxy Watch data?
Galaxy Watch data can be read on the paired Android phone through Health Connect for supported data types, with per-type user permission and a Play Console health permissions declaration. Samsung's own health SDKs expose additional data, and some access may require Samsung partner approval.
Are HRV values comparable across smartwatches?
Not directly. Vendors use different HRV metrics, measurement windows and conditions, for example Apple reports SDNN while others use RMSSD-based values or proprietary scores. We record the metric definition with each value so analysis compares like with like.
Is smartwatch SpO2 suitable for clinical monitoring?
Generally no. Smartwatch SpO2 features are typically designed for wellness use, and readings can be affected by fit, motion and skin perfusion. Programs that need clinical oxygen saturation monitoring usually use an FDA-cleared pulse oximeter.
Do we need a mobile app for smartwatch data?
For Apple Watch, yes, since data must be read from HealthKit on the iPhone. For Wear OS and Galaxy Watch through Health Connect, also yes on Android. Garmin and Fitbit data can be retrieved server to server through their cloud APIs once the user authorizes your application.