Co-sponsored by the Center for Biomedical Data Science at Yale and the Yale Institute for Network Science

“Donatello - A Machine Learning Platform for Clinical Real-World Data”

Speaker: Wade Schulz, MD, PhD
Assistant Professor; Director of Informatics, Laboratory Medicine
Center for Outcomes Research & Evaluation (CORE)

Talk summary: Real-World Data (RWD) offers the potential to improve the efficiency of biomedical research and increase the size and scope of studies. However, clinical data is extremely complex with few consistently-applied standards. In addition, there are regulatory and clinical validation requirements that are unique to the application of machine learning within the healthcare industry. To make data acquisition and the deployment of machine learning models less burdensome, we have begun to develop Donatello, an open-source predictive analytics platform that can be used to build and deploy predictive models within the healthcare setting. We will discuss the benefits of and barriers to the use of real-world data, the regulatory requirements for implementing clinical predictive models, a technical overview of the Donatello platform and other big data resources and Yale New Haven Health, and demonstrate a model being built within this platform.      

Speaker bio: Dr. Schulz is a physician scientist in the Department of Laboratory Medicine with a research focus in computational healthcare. His background includes work in basic science and both academic and industry experience in healthcare and software development. His group works on the development and validation of clinical predictive models and methods for computational phenotyping to accelerate precision medicine initiatives.  

.