1. Open mHealth and the Center of
Excellence for Mobile Sensor
Ida Sim, UCSF + Open mHealth
David Haddad, Open mHealth
Advancing biomedical discovery and improving health through mobile sensor big data
Cornell Tech ♦ Georgia Tech ♦ U. Memphis ♦ Northwestern ♦ Ohio State ♦ Open mHealth
Rice ♦ UCLA ♦ UC San Diego ♦ UC San Francisco ♦ UMass Amherst ♦ U. Michigan
2. Problem 1: more and more digital health tools are coming
to market but their data is trapped in silos.
3. Problem 2: your clinical visit is only a tiny slice of your
health pie. we know that it’s only through the combination
of multiple data streams that we can tell a person’s health
4. Solution: Open mHealth solves these problems by offering
developers and health systems a free and open API to
access digital health data with the right clinical context.
We’re working with a community of developers and clinical
experts to build an infrastructure for digital health.
5. Mission: Unlock the potential of digital health data so it can
be meaningful in the clinical environment.
● Penn Med
Symptom / pain management:
● UC Davis
● Global Alliance for
Genomes and Health
(Just a few examples!)
● Fitbit, Jawbone,
Building connections to:
● Google Fit
● Entra, Agamatrix
Clinically focused devices:
● Nutrify, FoodCare
● Mass Gen
● NY City
● 12 top clinical and
● UC Davis
● UCSF - Health
8. Linq is a way for people to collaborate with their doctor around
health tracking using the apps and devices they love
14. MD2K Mobile Data Science
• Generate software, tools, data, and science
• To enhance the ability to gather, analyze, and
interpret health-related mobile sensor data,
• Which facilitates the development of innovative
methods for early detection & prevention of complex
• Share data and tools via Open mHealth
• Establish an mHealth web portal and conduct
annual mHealth Training Institutes
18. Mobility-related Sensors
• Sensors used in MD2K
– 9-axis wrist sensors: in-house, and Microsoft Band?
– accelerometer on smartphones
• GPS/GIS algorithms from PALMS project
• Would like to be able to use commercial sensors
to increase generalizability
19. Mobility-related Features of Interest
• Sedentary minutes
• Minutes of any non-sedentary activity
• Minutes of moderate activity (3 - 6 METS )
• Minutes of vigorous activity (> 6 METS)
• 6 Minute Walk Test equivalent
• Flights of stairs climbed
• NYHA CHF class (i.e., dypsnea with exertion)
21. MD2K-Mobilize Data/Tool
• Open mHealth APIs to commercial devices
– what granularity of data?
– build jointly?
– feedback to companies on what to expose?
• Share algorithms as Open mHealth DPUs
– expose Mobilize algorithms as DPUs?
• Expose computed measures to BD2K and wider
community via Open mHealth schemas