Passive sensing data from wearables provide exciting opportunities to quantify human behavior, e.g., sleep metrics, activity levels, and heart rate, in longitudinal research studies. While several options for wearables exist, there is a wide disparity in the quality of the data and the frequency at which researchers can access the data. Garmin smartwatches and fitness trackers stand out in this regard. They have high signal/data quality, and Garmin provides several options to access wearable data in different granularity, thus enabling researchers to integrate digital phenotyping in their existing and new studies. CTBH has partnered with Garmin to gain access to their Software Development Kits (SDKs) and Application Programming Interfaces (APIs) that enable passive data collection.
In this tutorial, two different ways to use Garmin wearables in research studies and to passively collect digital phenotyping data were discussed: First, the Garmin Health API, which allows researchers to obtain valuable all-day health metrics such as steps, heart rate, and sleep. Next, the Garmin Health Companion SDK, which enables real-time continuous data collection, e.g., accelerometer, beat-to-beat intervals, and respiration rate. The key difference is the resolution of data: while Health API provides aggregated metrics throughout the day, the Companion SDK provides raw, time-series data, thus supporting a variety of study types and research questions. Presenters discussed Garmin data from recent studies to understand the quality and resolution of the data. Finally, colleagues at CTBH discussed their experiences using Garmin in their studies and potential opportunities, challenges, and limitations.
Link to presentation slides.
Link to presentation recording.