Learning Health Systems - overview
The Institute of Medicine in the US has defined a Learning Health(care) System as a system that is "designed to generate and apply the best evidence for the collaborative healthcare choices of each patient and provider; to drive the process of discovery as a natural outgrowth of patient care; and to ensure innovation, quality, safety, and value in health care" (National Academies Press, 2007).
With the widespread adoption and use of health IT, such as Electronic Health Records, consumer health wearables, and Internet of Things devices, the healthcare delivery system is generating a tremendous amount of electronic data that can be used to automatically evaluate outcomes of clinical care. However, the existing IT infrastructure largely lacks the regulatory, computational and evaluation methods needed to continually derive such knowledge and deploy it to providers, patients, families, administrators, payors and other stakeholders.
We seek to advance the development and capabilities of Learning Health Systems through three main research areas:
(1) Application of blockchain for capturing and sharing healthcare data and AI models in an immutable, trusted, federated and privacy-preserving manner;
(2) Development of AI methods that can perform real-time learning on dynamically generated healthcare data, both structured and unstructured, with human(s)-in-the-loop feedback; and
(3) Creation of new clinical research frameworks to assess the acceptability, explainability, actionability and outcomes of AI-based interventions in Learning Health Systems.
Director: Amar Das
Team Location: IBM Research Cambridge, Cambridge, Massachusetts