Many countries including the United States (US) and Japan are facing a rapidly aging population.  One of the critical challenges is to improve the quality of healthcare and quality of life for senior citizens while managing and reducing the health and social care costs. The growth and accessibility of wearable devices such as smart phones and watches enable continuous monitoring of a person’s vital signs and other health indicators. These wearable data, combined with the medical records and other community and environment data, are particularly valuable for senior communities in order to identify new conditions or relapses and notify the caregivers and healthcare providers for early intervention before they become critical. Collectively, such data from different individuals and communities, and different populations in different countries, can be used to learn better predictive models, advance medical research, and improve population health at large. 

Our planning project aims to build a multidisciplinary team of researchers and develop a research agenda for tackling the unique challenges in building and deploying a health monitoring system for senior communities.  The project achieved several outcomes.  First, we established a team across the US and Japan with complementary expertise including computer science, sociology, sensing technology, nursing, and clinical health, as well as partnership with community stakeholders.  Second, we developed several novel data management and machine learning solutions in our pilot studies including: 1) a multimodal predictive framework for fusing and learning from the multimodal data collected by digital devices or sensors while capturing both inter-modality correlations and intra-modality temporal dependencies, 2) privacy enhancing technologies (PETs) based on federated learning to enable joint learning across user devices without directly sharing the data while respecting personalized and rigorous privacy preferences.  Third, we gained a better understanding of the social and privacy implications of the technology via online surveys and the social and cultural differences in US and Japan which will inform the design and adoption of the technology.  Finally, we identified several use cases and developed a research agenda that we are currently investigating with follow-up projects.  These include: 1) use digital monitoring devices to assess natural changes in disease progression of Amyotrophic Lateral Sclerosis (ALS) for seniors, 2) use multimodal sensors to assess the health effects of heat exposure for agriculture workers, and 3) use mobility data collected by GPS devices and social interactions data for hyperlocal infection risk monitoring during a pandemic. 

Last Modified: 01/24/2023
Modified by: Li Xiong