mBTRC: Using mobile sensor big data to achieve spatio-temporal precision medicine

mHealth Biomarker & Technology Resource Center

Digital biomarkers obtained from mobile sensors provide unprecedented opportunities to bring about a new paradigm for maintaining wellness and managing disease: The delivery of behavioral and other digital therapeutic interventions —  anytime anywhere — at the optimal time and place for each individual.

Such spatio-temporal precision therapeutics is made possible with mHealth by tight feedback loops involving 1) sensed mHealth biomarkers, 2) data-driven predictive models for when, where, and how to intervene for a particular individual, 3) individualized interventions that are delivered via ubiquitous personal technologies such as smartphones, 4) identifying compliance with and the effects of the intervention, and 5) refining, over time, the predictive models for that individual.

Sensors are increasingly capable of measuring a rich variety of physical, biological, behavioral, social, and environmental factors relevant to health and disease, so the potential scope and impact of spatio-temporal precision therapeutics extends broadly across many clinical domains.

jitai graphicsIn order to achieve spatio-temporal precision medicine, these technical advancements and resources are needed: 1) scalable and efficient data collection of clinically meaningful biomarkers; 2) predictive intervention models, and 3) dynamic personalization
of the content and timing of mHealth therapeutics.

The NIH- funded Center of Excellence for Mobile Sensor Data-to-Knowledge has developed open source mobile sensor big data software for collecting large quantities of mobile sensor data for development and validation of digital biomarkers. This platform is now being used in ten NIH-supported field studies. MD2K provides the means for these studies to collect, store, transport and manage large quantities of raw sensor data (e.g., 4 GB/day per person) that become a digital biobank. MD2K allows these studies to obtain comprehensive assessment of behavioral and environmental risk factors by deriving a wide variety of newly discovered biomarkers presently unavailable in commercial sensors. But, most projects are unable to set up such big data infrastructure on their own and hence are largely limited to select few digital biomarkers offered in commercial wearable devices. Further, due to lack of predictive analytics methods, projects that are collecting digital biomarkers are unable to discover predictors to be used in intervention design.

The mHealth Biomarker & Technology Resource Center (mBTRC) will build upon MD2K’s unique infrastructure and intellectual resources to respond to the emerging mHealth needs of the biomedical research community.