Track Chair: May Wang
There has been exponential growth of real-time machine learning based risk prediction models in clinical medicine. Much of this growth, including in the sub-field of pediatric cardiac care, exists in the developmental and pre-clinical deployment stage. As the field evolves past development and validation, what are the biggest challenges to implementation and optimization? What is needed to effectively implement these risk prediction models in complex care systems? What lessons can be learned from clinical implementations of non-AI based near-real-time risk scores? How can we best study the impact of risk models in clinical care? This talk investigates this translational topic of AI-based clinical decision support, human factors, implementation science, and quality improvement.