NYU Langone Unlocks Personalized Healthcare Using IoT & Wearables

Business Situation

Diagnosing and treating sleep apnea is crucial to managing medical risks such as hypertension, heart disease, and memory and mood issues, but underserved populations are unlikely to be enrolled in sleep studies. This specific situation formed the basis for this project, but there’s a broader implication that the initiative addressed: medical providers such as NYU Langone face challenges in providing the comprehensive digital support for patients necessary to help achieve affordable, quality healthcare that’s accessible to all.

Solution

Maven Wave is working with Dr. Azizi Seixas, Ph.D., Assistant Professor in the Departments of Population Health and Psychiatry at NYU Langone Health, to create a secure machine learning and AI-powered platform that leverages data collected from wearables to provide real-time recommendations. For example, the platform can aggregate demographic data plus FitBit data outlining physical activity, sleep, and diet/nutrition trends to create custom profiles that send personalized notifications to patients.

Maven Wave is using Google Cloud Composer for workflow orchestration. Raw FitBit biometric data is pushed to Google Cloud Storage (GCS), where it is correlated and warehoused with patient profile data and historical data. This HIPAA sensitive data is housed in a secure environment, configured per best-practices for network and identity management, to protect customer information. The NYU Langone research team uses Data Studio and Datalab to perform data analysis that correlates metrics with population and geographic information such as census and Google Places API data. Patients can visualize their own progress and receive personalized messaging to encourage adherence through a web-based health and wellness portal.

Results

The framework that Maven Wave created with the NYU Langone enables the team there to mitigate the burden of chronic diseases. With this platform, physicians can advise patients remotely, providing automated health recommendations based on actual data. 

The technology also allows physicians to work more efficiently and provide specialized care to patients who may not otherwise get it, especially in underserved populations. Plus, once the architecture is fully developed and tested with the initial study population, it can easily be scaled up for wider applications in situations that don’t require in-patient monitoring.