Created at the University of California, San Francisco (UCSF), Center for Digital Health Innovation (CDHI), BeeKeeperAI accelerates the development and deployment of artificial intelligence (AI) algorithms in healthcare. The BeeKeeperAI platform allows healthcare data stewards to keep their sensitive data in their secure cloud environment while safely providing access to the data for 3rd party development and deployment of AI solutions to improve outcomes and reduce the cost of healthcare.
USE CASEConfidential Computing
Problem of Timely and Secure Access to Highly Diverse, Real-World Data
- The primary challenge with clinical quality ML algorithms is the need to access high quality, diverse datasets that are representative of global patient populations. The standard of clinical generalizability applied by most regulatory agencies requires that algorithms must produce similarly accurate results regardless of the type of data acquisition equipment, the demographics of the patient population, clinical setting, or other social determinants. To meet the standard, an algorithm developer must have access to data representative of that which the model will face when it is deployed into diverse clinical environments.
- As stewards of protected health information, healthcare organizations have a legal and ethical obligation to prevent inappropriate data access resulting in privacy breaches. This obligation and the risk of financial and reputational consequences of a privacy breach has created an environment where data stewards are extremely hesitant to share or allow access to their sensitive patient data.