We are pleased to announce the first set of wearables data available on Precision Health’s Analytics Platform and accessible via our self-serve tools.

The PROviding Mental health Precision Treatment (PROMPT) Precision Health Study has been gathering data from mobile fitness devices, FitBits, and HealthKits (Apple watch), including information about physical activity, weight, calories burned, heart rate, sleep patterns, and more.

Go to the Precision Health Analytics Platform documentation site to learn more about what is available. Questions? Email PHDataHelp@umich.edu.

The goal of the PROMPT study is to reduce the burden of depression by 1) increasing capacity in the mental health care system through expanding use of mobile-technology-delivered interventions, and 2) accelerating recovery from mental illness by better matching patients to pharmacological, psychological, and mobile-based treatments.

“Genetic analysis combined with data gathered from mobile technology and environmental data may have the capacity to inform mental health treatment decisions and allow for comprehensive precision treatment,” said study Co-PI Amy Bohnert, PhD, MHS.

Through a collaboration with Precision Health’s Data Analytics & IT Workgroup, study Co-PIs Bohnert and Srijan Sen, MD, PhD, have made PROMPT data accessible to researchers across campus. “We’re looking for people who want to run with it and look in the data and publish what they find…we don’t see it as just our study,” said Bohnert. These data are now available to researchers and instructors who choose to use real-world datasets in classroom projects.

“The data is already being used by multiple researchers across campus. For example, Dr. [Walter] Dempsey’s students in the Biostatistics 629 class are utilizing the data to investigate a broad set of questions,” said Sen.

Dempsey, an Assistant Professor of Biostatistics, explained how his students are applying the data:

Based on initial conversations with the PROMPT PIs, the students identified a scientific goal and developed their own analytic strategy. They then proceeded through the data analytic pipeline, from data cleaning/modification, to applying advanced data analytic techniques to the cleaned PROMPT data. For example, some students focused on using machine learning methods to identify key predictors of response to the mobile technology interventions from the wearable sensor and survey data.  At the end of the semester, students present their results in class … and receive feedback from faculty and PROMPT PIs.

“The data are fundamental to the course,” Dempsey said. They “allow students to practice a variety of statistical and machine learning techniques as well as real-world data skills, such as data modification, cleaning, and visualization.”

PROMPT data available via the Precision Health Analytics Platform include:

• Sensor data
  • activity level (steps, active minutes, sedentary minutes)
  • sleep data
  • heart rate
• Daily mood ratings
• Usage minutes of app-based mental health interventions Headspace (Mindfulness) and Silvercloud (Cognitive Behavioral Therapy)
• Genomic & environmental data
• A range of psychological assessment scales and measures

In addition to wearables and sensors, data are obtained from the electronic health record, DNA collection via saliva sample, and participant self-report surveys and mental health symptom assessments collected via the MyDataHelps app.

To date, PROMPT has enrolled 1450 participants (71% of whom are female), ranging in age from 18 to 80.

“The PROMPT study provides data that helps improve our ability to provide in-time mental health support to those who need it,” said Dempsey. “The data-driven treatment rules learned from this study can hopefully better identify and match patients to mobile-based treatments, helping to expand the reach of mental health care through mobile technology interventions.” He added, “The Precision Health team has been amazing in helping our class work with the data…. through everything from study background to data access to high-performance computation support.”

Questions? Email PHDataHelp@umich.edu.