Phone health data and wearables in one research pipeline

The source-independent ingestion pipeline is live: the versioned contract the mobile apps use (ingest v1), the collection plan, historical and incremental sync, validation and sync status per source. The metric catalog and data quality checks are in use as well.

The iOS and Android apps are code complete; store upload and real-device acceptance are pending. The apps will be announced on this site when they are published in the stores. Support for Chileaf wearables and hubs is in development and waits on physical devices and vendor answers. That is why this page is marked “in development”.

On iOS, data written to Apple Health by Apple Watch and other sources is read by the EnaResearch app with the participant’s operating system permissions. On Android the source is Health Connect; Samsung Health data is collected only if it is written to Health Connect. Huawei Health is out of scope.

The app collects data from the start of participation; if the study asks for it, up to 30 days of history before participation are also collected as a baseline. After that only new data is read. The phone may go offline, or the operating system may limit background work; sync resumes where it stopped and its status is reported.

Cumulative metrics such as steps and active energy are collected as 5-minute buckets through the platform’s de-duplicating aggregate query, so overlapping phone and watch samples are not counted twice.

The catalog holds about 197 device metrics today and is versioned. Researchers pick the metrics for a study; the permitted sources are derived automatically from the sources that can collect the selected metrics. Metrics a study does not select are not collected.

For mental health scales (such as GAD-7 and PHQ-9) only the total score and severity band are collected. Reproductive health and sexual activity metrics are flagged as sensitive: they are collected only if a study explicitly selects them, and selecting them shows a reminder about ethics approval and participant information.

Data from Chileaf devices is planned to arrive two ways. On the phone path, the device connects to the EnaResearch app through the vendor’s iOS and Android SDKs. On the hub path, a CL900 hub listens to CL837 devices worn by several participants; a bridge on the organization’s network forwards the data to the organization’s EnaResearch server over MQTT encrypted with TLS.

Because a hub serves several devices, the participant is never inferred from the hub identity; each device is linked to the right participant through a time-ranged assignment, and a device can be assigned to only one person at a time. History reading, background and multi-device behavior are being validated with the vendor and on physical devices; no capability that is not in the vendor’s documentation is assumed.

The mobile app sends data in batches. The server accepts a batch and returns a batch number; the app deletes its local copy only once the batch has been stored durably. We do not claim end-to-end “exactly once” delivery; instead, resends do not create duplicate records.

Each measurement is stored with its source: source type, device or app identifier, measurement time at the source, time of arrival at the server, the time zone offset where available and the transformation version. These details are the basis for data quality checks and research datasets.

EnaResearch is a research data collection and management platform. It is not a medical device, a clinical decision-support system or a real-time patient monitoring system; it raises no alarms and makes no claim about the clinical accuracy of measurements. Participants do not receive notifications. Whether device and platform data are fit for a study is assessed by the research team under the relevant protocol.